[med-svn] [Git][med-team/relion][debian/master] 4232 commits: Initial commit with contributors
Roland Fehrenbacher
gitlab at salsa.debian.org
Tue Sep 29 20:31:43 BST 2020
Roland Fehrenbacher pushed to branch debian/master at Debian Med / relion
Commits:
445e1c8e by Bjoern Forsberg at 2015-02-09T14:39:31+01:00
Initial commit with contributors
- - - - -
946e5083 by Bjoern Forsberg at 2015-02-13T15:38:51+01:00
UPdate of branch readme to include plan
- - - - -
bcc01990 by Bjoern Forsberg at 2015-02-16T16:30:20+01:00
First draft: single trans test
- - - - -
6e9343a1 by Bjoern Forsberg at 2015-02-16T17:11:16+01:00
reproducing run.sh in test/single_trans
- - - - -
bb86ed31 by Bjoern Forsberg at 2015-02-17T12:47:45+01:00
Update of test single trans
- - - - -
cbad489a by Bjoern Forsberg at 2015-02-18T13:12:42+01:00
temp CMakeLists (draft)
- - - - -
00f03481 by Farhad Kimanos at 2015-02-19T09:59:39+01:00
introduction of cmake
- - - - -
e72bfe71 by Farhad Kimanos at 2015-02-19T15:20:22+01:00
cmake makes it to the linking
- - - - -
ec26366c by Bjoern Forsberg at 2015-02-19T15:37:20+01:00
Some changes
- - - - -
776c2980 by Bjoern Forsberg at 2015-02-19T16:40:50+01:00
CmakeLists check for system fltk/fftw/cuda
- - - - -
352a92c2 by Bjoern Forsberg at 2015-02-20T17:21:59+01:00
REL_REFINE binary links /w sys libs + external libs path
- - - - -
1834c023 by Bjoern Forsberg at 2015-02-23T14:05:36+01:00
cmake subdir w/ FindFFTW.cmake
- - - - -
010bc071 by Bjoern Forsberg at 2015-02-23T16:34:25+01:00
Compile object lib once for all binaries
- - - - -
17df01a3 by Bjoern Forsberg at 2015-02-23T17:42:21+01:00
update to alpha-5
- - - - -
415056d3 by Bjoern Forsberg at 2015-02-23T17:47:24+01:00
cuda_add_exec clause added
- - - - -
a1ee6fa8 by Bjoern Forsberg at 2015-02-23T18:05:20+01:00
merge GPU to master original alpha-5
- - - - -
46ea8f4e by Bjoern Forsberg at 2015-02-25T12:01:50+01:00
Introduced a test hardcoded to CMakeLists: (-DRELION_TESTING)
- - - - -
08502dca by Bjoern Forsberg at 2015-02-25T13:56:45+01:00
more tests
- - - - -
d5c2da77 by Bjoern Forsberg at 2015-02-25T18:38:46+01:00
moved test output to build dir
- - - - -
2be4bb6d by Farhad Kimanos at 2015-02-26T09:51:57+01:00
Minor changes in cmake files, mostly clean-ups
- - - - -
608ac3b6 by Farhad Kimanos at 2015-02-26T10:00:14+01:00
Main cmake file merged
- - - - -
7876d42b by Farhad Kimanos at 2015-02-26T14:00:29+01:00
Reduced cmake files to bare minimum
- - - - -
55bbf15a by Bjoern Forsberg at 2015-02-27T11:30:17+01:00
moved test flag and set relative path apps/CMakeLists
- - - - -
95b38278 by Farhad Kimanos at 2015-02-27T12:10:09+01:00
GPU utils
- - - - -
fda04da4 by Farhad Kimanos at 2015-02-27T12:12:18+01:00
Merge branch 'GPU-trans' of https://bitbucket.org/bforsbe/relion-1.4 into GPU-trans
- - - - -
bac69a2b by Farhad Kimanos at 2015-02-27T13:40:52+01:00
Cmake skips search for openGL, which is not required
- - - - -
f2845154 by Bjoern Forsberg at 2015-02-27T14:20:28+01:00
cuda header fisrt draft
- - - - -
e4fcc710 by Farhad Kimanos at 2015-02-27T14:21:46+01:00
Merge branch 'GPU-trans' of https://bitbucket.org/bforsbe/relion-1.4 into GPU-trans
- - - - -
56653374 by Bjoern Forsberg at 2015-02-27T14:24:24+01:00
first draft cuh
- - - - -
5be62381 by Farhad Kimanos at 2015-02-27T14:25:21+01:00
Merge branch 'GPU-trans' of https://bitbucket.org/bforsbe/relion-1.4 into GPU-trans
- - - - -
b13cb697 by Farhad Kimanos at 2015-02-27T16:17:06+01:00
Working GPU kernel
- - - - -
eb8beca3 by Bjoern Forsberg at 2015-03-02T16:50:44+01:00
Added diff2 test
- - - - -
18e909d4 by Farhad Kimanos at 2015-03-02T17:02:07+01:00
CUDA implementation of diff2 calculations.
- - - - -
95ed9ef0 by Farhad Kimanos at 2015-03-02T17:15:48+01:00
Merged
- - - - -
e548b2b0 by Bjoern Forsberg at 2015-03-02T18:19:54+01:00
Fixed gpu-diff2
- - - - -
8f20bd6f by Bjoern Forsberg at 2015-03-02T19:18:20+01:00
fix make build hangup by sep. lib for gpu-util
- - - - -
fb1d3459 by Bjoern Forsberg at 2015-03-03T10:41:57+01:00
--gpu flag + test deps
- - - - -
64246976 by Farhad Kimanos at 2015-03-04T17:24:42+01:00
New layout for massive acceleration of refine.
- - - - -
75ee1194 by Bjoern Forsberg at 2015-03-04T18:22:28+01:00
Draft SharedMem handle and diff-calc looped over partial images
- - - - -
2e782a15 by Farhad Kimanos at 2015-03-05T15:31:57+01:00
Non-working data managment for CUDA kernel calls.
- - - - -
fc6d73ed by Bjoern Forsberg at 2015-03-06T11:55:48+01:00
Fixed code for non-blocksized images
Also changed test to sun with non-blocksized image
- - - - -
8acd17c0 by Farhad Kimanos at 2015-03-06T12:01:59+01:00
New globally modified architecture of the data managment for an optimized CUDA implementation of optimizer.
- - - - -
e450563e by Farhad Kimanos at 2015-03-06T14:18:38+01:00
It compiles, YEAH! Though ignoring warnings like I just don't care.
- - - - -
9a76ee2a by Farhad Kimanos at 2015-03-06T14:19:41+01:00
Merge branch 'GPU-trans' of https://bitbucket.org/bforsbe/relion-1.4 into CUDA-Glob
- - - - -
bf62552f by Farhad Kimanos at 2015-03-06T17:20:19+01:00
Rearragements of loops in cuda optimiser.
- - - - -
899fc696 by Farhad Kimanos at 2015-03-09T10:36:43+01:00
Rearragements of loops in cuda optimiser.
- - - - -
86d36f4e by Farhad Kimanos at 2015-03-09T10:52:58+01:00
Merge within branch as a side effect of rewriting history that contained incorrect merging with other branches.
- - - - -
d15ea40e by Bjoern Forsberg at 2015-03-09T14:44:29+01:00
modularize host/device image diffs, add reduction alogrithm
- - - - -
20d00918 by Farhad Kimanos at 2015-03-09T16:29:16+01:00
Some dummy code to get it to run.
- - - - -
867662eb by Bjoern Forsberg at 2015-03-10T15:20:46+01:00
the most annoying bugfix yet
- - - - -
e677ce2a by Bjoern Forsberg at 2015-03-11T15:35:55+01:00
freeing cuda mem + expanded tests to diff2 set
- - - - -
75f3d391 by Farhad Kimanos at 2015-03-11T15:38:22+01:00
Working gigantic kernel call with trans x ref number of blocks.
- - - - -
3c36cc30 by Bjoern Forsberg at 2015-03-11T16:00:10+01:00
Merge branch 'GPU-trans' into CUDA-Glob
Conflicts:
src/gpu_utils/cuda_img_operations.cu
src/ml_optimiser.cpp
- - - - -
5b0cf563 by Farhad Kimanos at 2015-03-11T16:12:28+01:00
GPU reproduces identical values for first particle, when not ctf-corrected.
- - - - -
06634d64 by Farhad Kimanos at 2015-03-11T16:12:45+01:00
Merge branch 'CUDA-Glob' of https://bitbucket.org/bforsbe/relion-1.4 into CUDA-Glob
- - - - -
f84515a8 by Farhad Kimanos at 2015-03-12T11:13:01+01:00
Working implementation reproducing CPU results.
- - - - -
c985c256 by Farhad Kimanos at 2015-03-12T16:13:53+01:00
Working implementation reproducing CPU results.
- - - - -
edfbae44 by Farhad Kimanos at 2015-03-12T16:19:07+01:00
Merged
- - - - -
acddf747 by Farhad Kimanos at 2015-03-12T18:04:18+01:00
Copied storeWeightedSums into CUDA optimiser as is. Builds and works as previous commit.
- - - - -
c0b3e6e1 by Bjoern Forsberg at 2015-03-13T11:59:05+01:00
Modifed test to be relevant for massive cuda-kernel call
- - - - -
e366d004 by Bjoern Forsberg at 2015-03-13T13:44:31+01:00
GPU can handle ctf-correction
- - - - -
51efcc42 by Bjoern Forsberg at 2015-03-13T21:43:08+01:00
GPU can handle scale-correction
However accuracy when using scale_correction
may have been affected compared to CPU-based
computation. Need to validate.
- - - - -
595b0eab by Erato.bjoern at 2015-03-15T14:09:00+01:00
Update to 1.4-beta-1
- - - - -
d8ca8d65 by Bjoern Forsberg at 2015-03-16T11:25:20+01:00
Merge branch 'CUDA-Glob'
Conflicts:
src/ml_optimiser.cpp
- - - - -
eddaf491 by Farhad Kimanos at 2015-03-16T18:05:14+01:00
Non-working first cuda-implementation of storeWeightedSums.
- - - - -
7707af0d by Farhad Kimanos at 2015-03-16T18:06:27+01:00
Merge branch 'CUDA-Glob' of https://bitbucket.org/bforsbe/relion-1.4 into CUDA-Glob
- - - - -
b89622d3 by Bjoern Forsberg at 2015-03-17T10:22:11+01:00
Faster diff2 kernel algorithm
also some clarifying annotations in weight calc
- - - - -
fd3101c5 by Bjoern Forsberg at 2015-03-18T17:45:48+01:00
CUDA Diff-kernel avoids insignificant diffs
from being calculated. Some additional minor fixes.
- - - - -
2e0e7286 by Bjoern Forsberg at 2015-03-19T09:45:53+01:00
Amended testing to remove old output before tests
- - - - -
cf67c32c by Bjoern Forsberg at 2015-03-19T10:50:32+01:00
diff2-kernel can handle #projs>65535
by decomposing a larger number into 2 components
of size ceil(sqrt(number)), and submitting a 3D
blockgrid instead of a 2D blockgrid.
- - - - -
7cbf88e1 by Farhad Kimanos at 2015-03-19T17:37:46+01:00
(Unverified) Working implementation of storeWeightedSums.
- - - - -
539c408b by Farhad Kimanos at 2015-03-19T17:54:15+01:00
Merge and small modification of cuda optimiser block sizes.
- - - - -
3a100cf7 by Bjoern Forsberg at 2015-03-20T16:05:45+01:00
diff2-kernel based on paired tasks
as opposed to intiating a block for each possible
task and exiting those which are insignificant.
Now this assesment is cpu-based and the minimum
(padded) number of blocks is invoked with the
kernel, utilizing references for all blocks for
read/write.
- - - - -
bc8f9209 by Farhad Kimanos at 2015-03-23T12:12:50+01:00
#
- - - - -
44e10af4 by Bjoern Forsberg at 2015-03-26T12:05:37+01:00
rot/transidx memleak fix (no apparent GPU memleak)
- - - - -
565fa242 by Bjoern Forsberg at 2015-03-30T10:37:22+02:00
~2x reduction in GPU wavg-kernel
(and alot of memory debug code)
- - - - -
2b43cdd9 by Bjoern Forsberg at 2015-03-30T10:37:41+02:00
new layout of GPU wavg-kernel
+ fewer global memory access requests
+ more scalable blocksize (more threads =
faster execution, up to a point)
- - - - -
5ea110b1 by Bjoern Forsberg at 2015-03-30T14:12:21+02:00
GPU wavg-kernel can handle #projs>65535
- - - - -
91a68a87 by Farhad Kimanos at 2015-03-31T13:57:08+02:00
Working CUDA storeWeightedSums
- - - - -
c1b245b4 by Farhad Kimanos at 2015-03-31T15:10:33+02:00
Merge
- - - - -
7f9c9e20 by Farhad Kimanos at 2015-04-01T20:49:53+02:00
Added some utils for memory managment and data flows.
- - - - -
dfb11824 by Erato.bjoern at 2015-04-03T08:35:20+02:00
Revised tests
- - - - -
6cbf66f1 by Erato.bjoern at 2015-04-03T08:37:39+02:00
Merge branch 'CUDA-Glob' of https://bitbucket.org/bforsbe/relion-1.4 into CUDA-Glob
- - - - -
cc5cfadd by Bjoern Forsberg at 2015-04-03T10:45:17+02:00
Cmake handles compile on K80-machines
- - - - -
7a2faa82 by Bjoern Forsberg at 2015-04-03T10:46:15+02:00
Expanded expectation()-tests
It is now evident that the ctf-correction via the
--ctf flag does not work on GPU resources.
- - - - -
43eb00ec by Bjoern Forsberg at 2015-04-04T17:46:08+02:00
GPU can ctf-correct with oversampling
- - - - -
4da5f81b by Erato.bjoern at 2015-04-04T20:28:17+02:00
Draft GPU cc_diff-kernel (untested)
- - - - -
6f3dfd4b by Dari Kimanius at 2015-04-07T12:46:36+02:00
CMAKE checks whether X11 and FLTK is found before compiling GUI targets
- - - - -
1d2af9f8 by Dari Kimanius at 2015-04-08T18:09:22+02:00
Extended cmake instructions for the non-GUI case and simple debugging-options.
- - - - -
a0b570c9 by Dari Kimanius at 2015-04-08T18:30:02+02:00
Working ctf correction for storeWeightedSums.
- - - - -
eb2c731a by Bjoern Forsberg at 2015-04-08T21:18:16+02:00
more tests
- - - - -
491555f8 by Bjoern Forsberg at 2015-04-08T21:30:13+02:00
Merge branch 'temp' into CUDA-Glob
- - - - -
de7e360d by Dari Kimanius at 2015-04-09T14:20:52+02:00
Added precompiler macros for CPU and GPU benchmarking.
- - - - -
96f24e2c by Dari Kimanius at 2015-04-09T14:21:10+02:00
Merge branch 'CUDA-Glob' of https://bitbucket.org/bforsbe/relion-1.4 into CUDA-Glob
- - - - -
e386a888 by Bjoern Forsberg at 2015-04-09T16:28:40+02:00
Tests on relion native output
../tests/ref_Fweights.txt
- - - - -
10bec6d1 by Bjoern Forsberg at 2015-04-09T16:53:56+02:00
Merge branch 'CUDA-Glob' of https://bitbucket.org/bforsbe/relion-1.4 into CUDA-Glob
- - - - -
49bb1ef9 by Dari Kimanius at 2015-04-10T10:42:53+02:00
Added benchmarking tool in rest of the cuda ml optimiser code. Did some modifications to names and syntax.
- - - - -
6bbdf296 by Dari Kimanius at 2015-04-10T10:43:08+02:00
Merge branch 'CUDA-Glob' of https://bitbucket.org/bforsbe/relion-1.4 into CUDA-Glob
- - - - -
17588afc by Dari Kimanius at 2015-04-10T12:11:51+02:00
Better debug warnings.
- - - - -
74ce65dc by Dari Kimanius at 2015-04-10T14:56:15+02:00
Fixed memory leakage caused by a drunk and code incident.
- - - - -
b3c4e4b3 by Dari Kimanius at 2015-04-10T15:36:21+02:00
Kernel precision can now be selected at precompile time.
- - - - -
3fda9c31 by Dari Kimanius at 2015-04-10T16:58:06+02:00
Fixed another dunk-n-code incident and converted relevant indices to long, including pair indices sent to the kernel.
- - - - -
c067740f by Lus at 2015-04-10T23:02:25+02:00
Added some documentation.
- - - - -
382d1a84 by Lus at 2015-04-10T23:05:26+02:00
Fixed error in Fweights and wavgs.
- - - - -
6d30d3df by Dari Kimanius at 2015-04-13T15:17:36+02:00
Updated testing reference-set
- - - - -
0bf0f533 by Dari Kimanius at 2015-04-16T11:21:24+02:00
Moved the gpu-cpu-fork in the code further up in the iteration-tree. Also introduced two classes holding all the optimisation and sampling parameters to improve syntax readability.
- - - - -
2e0e89bd by Bjoern Forsberg at 2015-04-16T11:43:10+02:00
Draft GPU-proj-kernel
- - - - -
7bb45123 by Bjoern Forsberg at 2015-04-16T11:59:57+02:00
Merge branch 'CUDA-Glob' of https://bitbucket.org/bforsbe/relion-1.4 into CUDA-Glob
Conflicts:
src/gpu_utils/cuda_ml_optimiser.cu
- - - - -
bf9ad69e by Dari Kimanius at 2015-04-16T14:33:39+02:00
Fixed time unit for benchmark output and added some more benchmarking checkpoints.
- - - - -
cbe7eff8 by Dari Kimanius at 2015-04-16T14:34:46+02:00
Merge branch 'CUDA-Glob' of https://bitbucket.org/bforsbe/relion-1.4 into CUDA-Glob
- - - - -
cc58e5cd by Dari Kimanius at 2015-04-16T16:35:31+02:00
Working CUDA-implementation of reference projection.
- - - - -
7d206594 by Bjoern Forsberg at 2015-04-17T10:24:32+02:00
Using build type + moved flags into debug config
- - - - -
baa5dfb3 by Dari Kimanius at 2015-04-17T11:13:01+02:00
CUDA-acceleration of reference projections for both 'diff2' and 'store'
- - - - -
9f077af3 by Dari Kimanius at 2015-04-17T11:13:53+02:00
Merge branch 'CUDA-Glob' of https://bitbucket.org/bforsbe/relion-1.4 into CUDA-Glob
- - - - -
b6e6365f by Bjoern Forsberg at 2015-04-17T16:09:42+02:00
cuda (CPU) version of Euler->RotMatrix
Added since the cuda-version needs to be able to
make euler matrices with dynamic FLOAT precision,
which should be indep. of all-cpu code.
- - - - -
3171c302 by Bjoern Forsberg at 2015-04-19T14:09:08+02:00
Small memory fix based on profiling + suggest
Further changes to wavg-kernel based on profiling
are necessary, as may they be for diff2-kernel.
- - - - -
350e38a0 by Bjoern Forsberg at 2015-04-19T14:11:27+02:00
Added build-type PROFILING
However native flags are not caught, i.e. we
are not appending to current build-type flags.
- - - - -
cb5645d3 by Bjoern Forsberg at 2015-04-20T09:54:05+02:00
GPU shared_mem wavg kernel 1.0
Showing some improvement of memory limtation,
but no significant performance gain (no diff
at all, basically)
- - - - -
657e3129 by Bjoern Forsberg at 2015-04-21T20:53:58+02:00
GPU shared_mem wavg kernel 2.0
Introduces using groups of references covered
by a single block in order to reuse image data
loaded grom global memory, but at the cost of
consuming quite a large amount of shared mem.
Attains ~2x and makes wavg less time-consuming
than finely resolved diff2.
ALSO: Switches from Complex objects of given
floating point accuracy to separte objects
for real and imaginary components, this to
coalesce memory access.
- - - - -
99466b1a by Bjoern Forsberg at 2015-04-22T09:34:56+02:00
2x better GPU diff-kernel (avoids double-cast)
- - - - -
871bf72b by Bjoern Forsberg at 2015-04-23T14:31:49+02:00
GPU ConvertDiffsToWeights-kernel 1.0
estimated ~55ms timesave every call, but may be increased by
the time it takes to do one host-to-device copy, since diffs
are calculated on GPU.
- - - - -
b5bef8a1 by Bjoern Forsberg at 2015-04-23T18:04:07+02:00
Build type flags are appended, as intended
- - - - -
a068cd49 by Bjoern Forsberg at 2015-04-23T18:04:47+02:00
Thrust::sort on release build, in wavg-kernel
Speeds up sort on -O3 by ~10x
- - - - -
b66296a3 by Bjoern Forsberg at 2015-04-23T21:11:08+02:00
Spring cleaning
- - - - -
54b6178d by Bjoern Forsberg at 2015-04-26T22:06:21+02:00
Prepared for moving euler-matrix up
When finsihed, the boolean implementation will use the full set of
orientation (euler-) matrices for all classes, making the euler-
matrices shared across all classes, thus they need only be constructed
once.
Also modified code in storeWSums() - collect_data stage, to expose
parallelism and indicate future change.
- - - - -
d3eecd0f by Bjoern Forsberg at 2015-04-28T18:57:33+02:00
Projection using texture memory
The projection is performed by one of three kernels;
double-precision:
global memory access - explicit interpolation
single-precision:
texture memory access - explicit interpolation
--||-- - implicit (texture-read) interpolation
- - - - -
959ceb1a by Bjoern Forsberg at 2015-04-29T12:56:19+02:00
Working reorder of storeWS_collect_2 loop
Also draft (inactive) storeWS_collect2-kernel
- - - - -
dc9a0565 by Dari Kimanius at 2015-04-29T15:28:11+02:00
CUDA acceleration of the backprojector.
- - - - -
23af3c1d by Dari Kimanius at 2015-04-29T17:51:25+02:00
Merge branch 'CUDA-Glob' of https://bitbucket.org/bforsbe/relion-1.4 into CUDA-Glob
Conflicts:
src/gpu_utils/cuda_ml_optimiser.cu
- - - - -
ead91893 by Bjoern Forsberg at 2015-04-29T20:26:27+02:00
Partial solution GPU storeWS_collect_2
- - - - -
390dff2b by Dari Kimanius at 2015-04-30T14:07:16+02:00
Faster backprojection using an simpler approach.
- - - - -
e2d32476 by Bjoern Forsberg at 2015-04-30T15:11:27+02:00
Working GPU storeWS_collect_2
- - - - -
52b6309a by Bjoern Forsberg at 2015-04-30T15:12:04+02:00
Merge branch 'CUDA-Glob' of https://bitbucket.org/bforsbe/relion-1.4 into CUDA-Glob
- - - - -
9f1b4176 by Dari Kimanius at 2015-04-30T18:10:59+02:00
New approach for backprojection.
- - - - -
c2ab71c1 by Bjoern Forsberg at 2015-05-03T09:27:18+02:00
Merge branch 'CUDA-Glob' of https://bitbucket.org/bforsbe/relion-1.4 into CUDA-Glob
- - - - -
5a95dfc7 by Dari Kimanius at 2015-05-06T18:04:56+02:00
Optimized backprojection kernel.
- - - - -
27d79e96 by Dari Kimanius at 2015-05-11T18:01:25+02:00
Code put into functions to clean the code flow up a bit, re-ordering of projection settup to get coalesced rotational angles.
- - - - -
09bd3655 by Dari Kimanius at 2015-05-12T17:18:21+02:00
Further simplification of the code and slight optimisation of the BP-kernel.
- - - - -
c99780b2 by Bjoern Forsberg at 2015-05-20T17:08:33+02:00
1.37x diff2-kernel / 1.75x wavg-kernel
This from using the __ldg() intrinsic for some of the global loads;
it routes the load through the texture cache, meaning one may score
some hits. NOTE: Not all global loads benefitted from __ldg(), and
were omitted from this commit, BUT THE DEPENDENCY OF THE PERFORMANCE
INCREASE WAS NOT TESTED ON OTHER HARDWARE. The use of __ldg() herein
may thus not be the optimal across other hardware.
ALSO: __ldg() requires compute sm_35 or higher, thus the flag was
added to the nvcc build flags. Fallback should perhaps be made.
- - - - -
e2e8d45a by Bjoern Forsberg at 2015-05-21T11:38:07+02:00
GPU Reduce_wdiff2s-kernel 1.0
This makes for a ~50x-70x speedup of the image reduction, which
before took ~50ms on my hardware (Profiling build). However, it
could be improved by increasing the number of image-loads per
block, as this would decrease loads and writes per compute.
The kernel-call should also be able to handle more blocks than
the maximum 65535 in a single block-grid dimension, which would
be partially fixed by the above-mentioned improvement. Otherwise
a simple multidim-blockgrid solution is fine.
- - - - -
bb4095f6 by Bjoern Forsberg at 2015-05-21T15:21:54+02:00
Data_collect_1 rationalized, Rel Build-type added
(Post diff2-kernel mapping)
Memcpy:s that can be avoided if the diff2-kernal can do this
mapping directly:
line 1742 D-to-H diff2s
line 2012 H-to-D Mweights
Performance increase may not be evident over an entire run on
any build type other than RELEASE.
- - - - -
8dba2cea by Bjoern Forsberg at 2015-05-22T16:29:07+02:00
Index methods, reduce_wdiff_kernel, cuda-streaming
1. ProjectAllViews uses __ldg() on eulers, and some formalism changes.
2. Release build has thread-based cuda-stream parallelism (compile flag)
NOTE: -j 2 causes different output result (accuracy is compromised non-deterministically)
(see issue #14)
3. Reduce_wdiff2s-kernel can handle >65535 images to reduce
4. New Class "Indices" that has method to cast "ihidden_over" to partial indices and vice versa.
NOTE: this is a duplicate of a method in the class HealpixSampling, but has a bit clearer
index naming. It's better, because I wrote it. So there. Should be able to make use of
it to improve pair-list generation, but that might go away soon anyway. Has the following
methods:
- fineIndexToFineIndices(SamplingParameters sp)
- fineIndicesToFineIndex(SamplingParameters sp) FIXME Untested
- coarseIndexToCoarseIndices(SamplingParameters sp) FIXME Untested
- coarseIndicesToCoarseIndex(SamplingParameters sp) FIXME Untested
- - - - -
cd14e87a by Bjoern Forsberg at 2015-05-25T09:49:45+02:00
Diffkernel functionalized to same level as Proj
- - - - -
217c9ffc by Bjoern Forsberg at 2015-05-25T13:41:01+02:00
Release build compiles and runs with wrong output
- - - - -
dbaa4c07 by Dari Kimanius at 2015-05-25T15:52:39+02:00
Included an implementation of atomic add for cuda device code.
- - - - -
d5625a9b by Dari Kimanius at 2015-05-25T16:23:24+02:00
Merge branch 'CUDA-Glob' of https://bitbucket.org/bforsbe/relion-1.4 into CUDA-Glob
Conflicts:
src/gpu_utils/cuda_ml_optimiser.cu
- - - - -
02527698 by Dari Kimanius at 2015-05-25T16:25:41+02:00
Simplified the generation of the eulers matrices and added generation of 2D rotation matrices.
- - - - -
9a792fce by Dari Kimanius at 2015-05-25T16:29:39+02:00
Switching to int type for spacefill array in backprojection, due to tiny kernel speed-up.
- - - - -
b4d22fb3 by Dari Kimanius at 2015-05-25T18:25:17+02:00
Fixed bug only effecting the results in case of multiple iterations over ipart.
- The submission to global mem in wavg-kernels are modified to additions
- Memeory for wavg output is initialized to zero before kernel-call
- - - - -
cd324d37 by Bjoern Forsberg at 2015-05-25T19:29:01+02:00
ProjInDiff-kernel 1.0 works with single prec.
under control of bool do_combineProjAndDiff
- - - - -
2fc509cf by Bjoern Forsberg at 2015-05-25T19:30:25+02:00
Merge branch 'CUDA-Glob' of https://bitbucket.org/bforsbe/relion-1.4 into CUDA-Glob
- - - - -
75c462f0 by Bjoern Forsberg at 2015-05-25T19:38:40+02:00
Merge branch 'ProjInDiff' into CUDA-Glob
- - - - -
a72c3b34 by Bjoern Forsberg at 2015-05-26T10:47:54+02:00
wavg-kernel functionalized
This in preparation for ProjAndWavg kernel.
Also device mem usage at diff2- & wavg-time.
- - - - -
0bac1ef0 by Bjoern Forsberg at 2015-05-26T13:29:55+02:00
ProjAndWavg kernel 1.0
Works with single prec.
- - - - -
d30af4ad by Dari Kimanius at 2015-05-26T15:53:02+02:00
Cleaned up the code around collect_2.
Removed extra iclass loop
decoupled block_dim from wavg-kernel
- - - - -
5410b533 by Bjoern Forsberg at 2015-05-26T16:55:07+02:00
Fixedd previous issue with ProjAndWavg
- - - - -
97ef6f9c by Bjoern Forsberg at 2015-05-26T19:47:05+02:00
Reorganizing functions into separate files
- - - - -
a926a0bc by Bjoern Forsberg at 2015-05-27T10:15:56+02:00
Bindless (object-) textures
- - - - -
9ed1f948 by Bjoern Forsberg at 2015-05-27T16:15:24+02:00
Reorganizing texture-functions into separate files
Also, som preparation for diffs-grouping of images
- - - - -
ecde77a4 by Bjoern Forsberg at 2015-05-28T15:04:00+02:00
New block-control of PAV_Txx_D2 kernels
Currently only works with PROJDIFF_CHUNK_SIZE=1,
which reproduces previous behaviour. Higher values
will presumably reduce execution time by decreasing
the global load frequency and resuing computed
ref-values (up to) proportionally to chunk size.
NOTE: not all _syncthreads() calls may be necessary
in the PAV_TI_D2 kernel
- - - - -
f5c81541 by Bjoern Forsberg at 2015-05-28T19:24:50+02:00
Working block-control of PAV_Txx_D2
- - - - -
3f2a4005 by Bjoern Forsberg at 2015-05-29T16:10:44+02:00
Minor performance boosts PAV_Txx_D2
- - - - -
85883ad2 by Dari Kimanius at 2015-05-30T13:34:44+02:00
Minor clean up
Removed auto athor claims
Swithed cuda.h to cuda_runtime.h
- - - - -
1335c991 by Dari Kimanius at 2015-05-30T13:44:01+02:00
Filetree clean-up. All kernels in one dir.
- - - - -
b32c95fa by Bjoern Forsberg at 2015-05-31T10:37:14+02:00
Padding_factor into ProjAnd-X kernels
ALSO minor boost to reduction in PAV_TTI_D2
- - - - -
bc4bf4c9 by Bjoern Forsberg at 2015-05-31T10:54:44+02:00
Benchmarking works again
- - - - -
e4f219a7 by Bjoern Forsberg at 2015-05-31T14:28:32+02:00
Map & pair_list improved ~2x
- - - - -
c9c46a04 by Dari Kimanius at 2015-05-31T17:25:14+02:00
The projection set is divided up into subsets to spare memory during maximization.
There are 8 subsets currently. This number can be set in cuda_ml_optimiser.cu:2458
- - - - -
c4eb215a by Dari Kimanius at 2015-05-31T18:35:06+02:00
Merge branch 'CUDA-Glob' of https://bitbucket.org/bforsbe/relion-1.4 into CUDA-Glob
Conflicts:
src/gpu_utils/cuda_ml_optimiser.cu
- - - - -
85a7f2d7 by Dari Kimanius at 2015-06-01T11:59:04+02:00
Cleaned up some bad code in rotation setup where psi-values are grouped.
- - - - -
14419e46 by Dari Kimanius at 2015-06-01T12:41:32+02:00
Further bugfixes for data with sparce fine angle search.
- - - - -
cb0c4730 by Dari Kimanius at 2015-06-01T15:49:16+02:00
Smarter projection dividing based on significant number of orientation rather than the maximum number.
- - - - -
a7d90752 by Dari Kimanius at 2015-06-05T15:37:19+02:00
Benchmark related improvments and some minor clean-ups.
the gpu-toc does no longer take an arg
better nameing and clean-ups in cuda_benchmark_utils
- - - - -
a35da616 by Bjoern Forsberg at 2015-06-06T20:39:14+02:00
PAV_D2 kernels and -usage is double capa-bloue
Also added preprocessor control of exponentiation
limits (which limit accuracy loss according to
comment on line 4129 in ml_optimiser.cpp) in
sumweight instances;
double - limit is 700 exp(700) < 1.8E308
single - limit is 88 exp(88) < 3.4E38
- - - - -
6c56e12c by Bjoern Forsberg at 2015-06-07T11:10:03+02:00
Sumweight kenrel does both coarse and fine sampled
passes. The previous kernel used 32 threads to access
oversampled points, which are adjacent in global mem
and therefore very fast to load. In order for the kernel
to generalize for coarse points, the access is now made
with each thread fetching an (over)sampled orientation,
which are in multiples of 32 (so as to keep all threads
occupied), but not adjacent in global memory. The mem-
ory access should therefore be far less efficient, but
the code works equally well for all healpix- and over-
sampling, provided healpix_order>1.
- - - - -
05885bad by Erato.bjoern at 2015-06-08T09:21:12+02:00
Functionalised block-dim split
- - - - -
e33ec514 by Bjoern Forsberg at 2015-06-09T19:52:26+02:00
PAV_TTE and PAV_TGE fixed, double build works
They had two separate issues;
1 - One input int type was unsigned instead of long,
which became an issue because it actually needed to
be signed. thus -33 -> HUGE NUMBER
2 - The padding_factor amendments were only imple-
mented on the PAV_TTI and PAV-combo kernels. This
meant that when double precision forced a split of
PAV and WAVG, it forced usage of PAV_TGE, which was
unaware that it should correct the euler matrix by
padding factor.
- - - - -
2745f670 by Dari Kimanius at 2015-06-10T09:53:23+02:00
Fixed a minor issue with padding factor and also an adjustment to the projection divition algo
- - - - -
6f884058 by Bjoern Forsberg at 2015-06-10T10:23:03+02:00
Functionalised texture initialisation
- - - - -
f9bbb246 by Bjoern Forsberg at 2015-06-10T10:23:41+02:00
Merge branch 'CUDA-Glob' of https://bitbucket.org/bforsbe/relion-1.4 into CUDA-Glob
- - - - -
b3443a43 by Bjoern Forsberg at 2015-06-10T15:12:43+02:00
doCombineProjAndDiff=false works
a.k.a. the D2 kernel is up-to-date
with the block-control of commit f5c8154
- - - - -
91c55207 by Bjoern Forsberg at 2015-06-12T12:29:12+02:00
Indexing issue in higher order coarse pass fixed
Previous code relied on the assumption that all coarse orientations
should be difference-calculated, regardless of being the first or any
later refinement-iteration. This is apparently not true, as evident by
the if-statement in generateProjectionSetup which is dependent on
pdf_orientation being finite.
The subsequent errors with/without --ctf flag and for fewer particles,
is then likely a result of the varying probability that there should
be some overlap between the orientational priors (weights) and the
erraneously mapped differences. If it happened to be zero overlap,
the result would be sumweights=0 and exit=failure, otherwise the
algorithm would continue and exit with success, but have completely
nonsensical output data, since weights and priors are then utterly
mismatched.
- - - - -
12b85761 by Dari Kimanius at 2015-06-12T15:09:11+02:00
Minior fix, throws signal on cuda error.
- - - - -
220ca771 by Dari Kimanius at 2015-06-12T15:09:39+02:00
Merge branch 'CUDA-Glob' of https://bitbucket.org/bforsbe/relion-1.4 into CUDA-Glob
- - - - -
c8f70664 by Dari Kimanius at 2015-06-14T00:03:24+02:00
Diff2s in the coarse pass are taken care of by a much simpler approach, rendering a factor 2 speed up.
Program debug exits are replaced by "signal-rasing", for easier debugging.
- - - - -
c83d659d by Bjoern Forsberg at 2015-06-15T12:01:14+02:00
Fix on splitCudaBlocks()
- - - - -
e8f8fba1 by Bjoern Forsberg at 2015-06-15T15:57:22+02:00
Reengineered index-array use in GetAllSq (fine)
Also functionalised use of thrust min-val search
- - - - -
bbbbf832 by Bjoern Forsberg at 2015-06-16T12:08:56+02:00
GetAllSqDiffs() is coarse/fine-parted
Some small fixes and strategic comments
thrust::reduce on partial sumweight results
- - - - -
9237f7e2 by Bjoern Forsberg at 2015-06-16T19:06:58+02:00
Partial removal of dependence on sparse Mweight
Current code relies on the implicit indexing of diffs and
weights by mapping the into an array "Mweights", which in
the gpu-based code belongs to the op class. For somewhat
converged data this tends to be a VERY sparesly populated
array by way of the low significance of many of the asso-
ciated orientations of an image. Combined with the large
size of the array with fine sampling, the cost of
a) transferring this object to and from the gpu
b) keeping a largely empty array on the gpu and
cycling through comparisons of insignificant
elements
becomes too large compared to the actual computations able
to complete. The solution is also twofold;
A) move the object up in scope and keep it on the gpu
B) coalesce data my keeping an explicit indexing, rather
than implicit mapping
This commits adresses both for the fine pass of difference-
calculation, while still mapping to the Mweight array at
the end of converting differences to weights. This because
the storeWeightedSums()-function has not yet been adapted
to handle explicit indexing rather than implicit mapping.
- - - - -
e485bb34 by Dari Kimanius at 2015-06-17T16:38:09+02:00
Major revison of projector setup.
Projector model is loaded into GPU once at rank level.
A class for maintenance of projectors and device code for projections is added.
A section presumably faulty code was outcommented in src/metadata_container.cpp.
This action did not have any effect of the data-output-file for two iterations
of both tutorial and ryr data.
- - - - -
056de775 by Dari Kimanius at 2015-06-17T16:45:19+02:00
Minor clean-up in code non-related to GPU-acceleration.
- - - - -
640c9e80 by Dari Kimanius at 2015-06-23T14:36:42+02:00
Supermerge branch 'CUDA-Glob' of https://bitbucket.org/bforsbe/relion-1.4 into CUDA-Glob
Conflicts:
src/gpu_utils/cuda_kernels/proj_diff.cu
src/gpu_utils/cuda_ml_optimiser.cu
Major problems emerged during merge that required rewitting of large parts of the code.
Memory issue caused by varying class definition depending on compiler (gcc or nvcc)
However, code is not fully functional since further issues remain. These are fixed in next commit.
- - - - -
e551069f by Dari Kimanius at 2015-06-23T15:12:51+02:00
Minor fix of the index managment causing issues in fine pass. Now code is fully functional.
- - - - -
358d9cbf by Erato.bjoern at 2015-06-24T15:14:36+02:00
collectjobs kernel
- - - - -
9e5a4e14 by Dari Kimanius at 2015-06-24T16:17:32+02:00
Backproject is moved up to rank level.
Some minor improvements on benchmarking.
- - - - -
41dfa9b0 by Dari Kimanius at 2015-06-24T17:03:47+02:00
Merge remote-tracking branch 'origin/collectjobs' into CUDA-Glob
Conflicts:
src/gpu_utils/cuda_ml_optimiser.cu
- - - - -
2fa1abca by Dari Kimanius at 2015-06-24T20:31:41+02:00
Added a centralized class taking care of projections on the device named Cuda3DProjectorKernel.
This class holds all the parameters and the projection function, which tremendously reducing code complexity.
The performance impact have been considered and to my great surprise there was a slight improvment
on all the three related kernels except for cuda_kernel_wavg that remained the same.
- - - - -
11fd7d2d by Dari Kimanius at 2015-06-25T14:26:25+02:00
Project built in double precision using the cmake option -DCudaDouble=ON
- - - - -
9cbe515c by Dari Kimanius at 2015-06-25T15:47:18+02:00
Fixed projector for double precison.
- - - - -
2628c50f by Dari Kimanius at 2015-06-25T16:28:42+02:00
Major reorganization of code to move out as much as possible from the central file cuda_ml_optimiser.cu
- - - - -
ad4db909 by Dari Kimanius at 2015-06-26T14:41:28+02:00
Fixed error in weights, caused by exclution of ctf correction in the wavg kernel.
- - - - -
d0034001 by Erato.bjoern at 2015-06-26T15:57:00+02:00
Fix to collectjobs indexing
and some excessive index-juggling in the kernel
- - - - -
72bcbd3a by Bjoern Forsberg at 2015-06-30T19:01:00+02:00
Additional index fix
- - - - -
de076ec9 by Bjoern Forsberg at 2015-06-30T19:03:05+02:00
Merge branch 'collectjobs' into CUDA-Glob
- - - - -
7eecf42a by Bjoern Forsberg at 2015-06-30T23:43:26+02:00
Removal of Mweight-based collect2-kernel
Collect2 is from now on based on a dense
indexed array, which makes it faster, but
more importantly also removes the need for
copying the largely zeroed Mweight-array
to the device for each particle.
- - - - -
bfb24a49 by Dari Kimanius at 2015-07-01T11:03:54+02:00
Better benchmarking output and higher resolved benchmarking entries in convert and store.
- - - - -
24e7d1dc by Bjoern Forsberg at 2015-07-01T12:11:20+02:00
add thrust max-reduce, removed Mweight-dep MaxVal
- - - - -
39f60a22 by Bjoern Forsberg at 2015-07-01T14:37:25+02:00
Pass op by ref, more exhaustive benchmarks
- - - - -
3bc5c228 by Bjoern Forsberg at 2015-07-01T17:38:16+02:00
Getrow & nonZero-picking removed from fine pass
These apparently only constitute a significant part
of the CPU-bound code in the fine pass. That taken
together with the fact that the Mweight array is
used for much additional implicitly indexed checks
(like setting coarse significance for the fine pass)
makes it easier to (at the moment) let the two methods
(sparse Mweight and dense weights) coexist and simply
use the latter in the fine pass where it removes the
need to fetch a large section of an even larger
Mweight-array isolate non-zero weights.
- - - - -
f7e2c31c by Bjoern Forsberg at 2015-07-03T10:01:37+02:00
Fine pass Mweight-dep removed
NOTE: cross-correlation will fail with explanatory
error message. Since there is some doubt as for the
functionality of the CC-based kernels we see this
as a reasonable sacrifice at the moment.
- - - - -
6af7e693 by Bjoern Forsberg at 2015-07-03T16:49:36+02:00
New class IndexedArray to care for dense data
- - - - -
a659f22b by Bjoern Forsberg at 2015-07-03T20:51:55+02:00
New class ProjectionParams holding projection data
NOTE: Some methods added to the IndexedDataArray
are not working, commented out and "labeled" as
"NOT WORKING" in files
cuda_utils.cuh
cuda_helper_functions.cu
These methods were not identifiable in eclipse if
inlined, and producing segfaults in unrelated parts
of the code if not inlined.
- - - - -
1498c27c by Bjoern Forsberg at 2015-07-05T18:03:45+02:00
ProjectionSetup in fine pass made global
clarification: Since the coarse pass results in a
different set of orientations and translations, this
is still local in scope, held separate fro the fine
pass.
- - - - -
9501eab3 by Bjoern Forsberg at 2015-07-06T11:53:34+02:00
ProjectionParams handles classes (untested)
Only when IndexedDataArray can handle classes as well, can we
test and debug the multiple-class-capability of this commit
- - - - -
68ed4086 by Bjoern Forsberg at 2015-07-06T20:28:50+02:00
Draft proj_div-capable ProjParams (not working)
- - - - -
58d370a2 by Dari Kimanius at 2015-07-07T13:08:26+02:00
Fix for memory issues cauesed by an invalid delete and a incorrect boolean expression. Also fixed group-id in fine pass.
- - - - -
08cac241 by Dari Kimanius at 2015-07-07T13:08:45+02:00
Merge branch 'CUDA-Glob' of https://bitbucket.org/bforsbe/relion-1.4 into CUDA-Glob
- - - - -
093ff452 by Bjoern Forsberg at 2015-07-07T15:40:11+02:00
Fix to proj_div under new ProjParms
At som point the wavg-kenrel might benefit
from direct use of the dense weight data,
rather than map to a partial sparse object
which reduces all translations implicitly
for each significant orientation.
- - - - -
45c50917 by Bjoern Forsberg at 2015-07-07T15:43:13+02:00
Merge branch 'CUDA-Glob' of https://bitbucket.org/bforsbe/relion-1.4 into CUDA-Glob
- - - - -
d947c83a by Bjoern Forsberg at 2015-07-07T19:56:30+02:00
Minor fixes which do not affect single-class runs
- - - - -
9248ce51 by Bjoern Forsberg at 2015-07-09T18:16:57+02:00
ArrayMasks for classes (multi-class capab. prep)
We use dense arrays with explicit indexes for the sparse pass,
where oversampling could make the very few significnatly con-
tributing expensive to iterate through with nothing to do. At
some stages, we need to perform reductions of all classes. This
offers a solution;
Make an explicitly indexed array, carrying weights and indices
for all classes of this image. That way any class-wide reduction
has access to coalesced data with direct indexing. When we need
data for a specific class, we use a mask which when fed into a
constructor for a sub-indexedArray, returns a handle for the data
belonging to that class. The mask also holds arrays which deter-
mine the job-partion necessary for some kernels to exploit re-
use of some data.
The "global" index array for all classes of an image is (in
some sense) appended by the data of each new class as signficance
of each finely sampled gridpoint is determined, and the mask of
each class is simultaneoulsy made, to be available in the conv-
ersion of weights and the backprojection of them. The job parti-
tion of the mask is independent of the indexedArray, and can be
re-defined for the purposes of any kernel designed to use them.
- - - - -
08ce24f6 by Bjoern Forsberg at 2015-07-09T19:02:58+02:00
Job-splitting in function, consistently named
- - - - -
b3dd36c7 by Bjoern Forsberg at 2015-07-10T19:23:36+02:00
2 classes works for first iteration
The only apparent discrepancies are in the usual suspects; norm-corr
and log-likelihood essitmations. Class assingment and sumweight are
correct within the allowed precision.
- - - - -
b38dba71 by Bjoern Forsberg at 2015-07-10T20:45:14+02:00
2 classes pass second iter, w/ errs in metadata
- - - - -
e3434fd0 by Bjoern Forsberg at 2015-07-15T14:24:30+02:00
2 classes pass second iter, less errs in metadata
A shift in usage compared to previous commits
to enable multi-class refinement;
before:
only the classes which had significant orinetations to be examined were
prepared, by appending any new classes and weights to existing objects.
The result of this was a difficult-to manage class indexing, since loops
are over all classes, rather than all classes which are implicity deemed
relevant.
after:
Make space for all classes in class-arranging masks and ProjectionParam-
arrays, while keeping the weights-array coalesced. This way all classes
can be accessed by explicit indices in the masks and ProjParams-arrays,
and simply skipped when nothing is deemed significant in this class
to fix:
Some stuff is being done without purpose; breaks and continues should be
injected where nothing is to be done, so as make the run proceed quicker ..
Some general cleanup might be good as well, and the error with increasing
particle number persists.
- - - - -
29e2aa5c by Bjoern Forsberg at 2015-07-17T13:32:08+02:00
2 classes pass second iter for many particles
This commit fixes an error not apparent for low numbers of particles,
where the first iteration produces probability densities which are so
sparse that the multi-class passes essentially only ever tries the
same class which they were randomly assigned in the first iteration.
The error was that the weights which were determined for several classes
into a coalesced array, were never actually split in storeWS, but acc-
identally just used in its entirety. As long as the sparse probability
density limited the number of classes for all particles to a single one,
this was exactly equivalent to what was inteded. With more particles,
not splitting the array with weights for all classes caused undefined
behaviour, and eventually segfaults.
fixes #28
- - - - -
6108cdc1 by Bjoern Forsberg at 2015-07-17T18:28:25+02:00
Cleanup of cpu & gpu-code, no functional changes
- - - - -
4519a106 by Bjoern Forsberg at 2015-07-19T12:11:58+02:00
no comparison tests, add ryr tests /w local files
- - - - -
bf520de6 by Bjoern Forsberg at 2015-07-20T10:45:58+02:00
minor cmake cleanup
fixes #29 by changing debug-flags from normal vars to cache-vars
- - - - -
4ab55ba0 by Dari Kimanius at 2015-07-20T17:25:09+02:00
Moved projection-setup to rank-level.
Introduced a new class called "projector plan" calculating and holding projection-setup.
Added a vector of these as memebers of ml_optimiser.
- - - - -
6b46f561 by Dari Kimanius at 2015-07-21T12:38:28+02:00
Removed unnecessary writes ans therad-syncs to global mem in diff-coarse-kernel.
- - - - -
153cead1 by Dari Kimanius at 2015-07-21T15:42:45+02:00
Fiexed precalculation of projection-plan for multiple classes.
- - - - -
8f6de4f7 by Dari Kimanius at 2015-07-21T15:44:06+02:00
Fixed cmake build types to include definitions.
fix #32
- - - - -
def702b2 by Bjoern Forsberg at 2015-07-22T10:10:07+02:00
Removed old test file output references
Files this big should probably not be in the repo anyway,
but on an associated remote resource referenced by the
repo tests.
- - - - -
4d800f10 by Bjoern Forsberg at 2015-07-22T10:11:41+02:00
Merge branch 'ctest' into CUDA-Glob
- - - - -
d3207dd8 by Dari Kimanius at 2015-07-22T14:40:08+02:00
Minor clean-up
Removed cudaComplex and some thread locks.
- - - - -
5ce45b0d by Bjoern Forsberg at 2015-07-22T16:32:20+02:00
cmake sets cufft-libs/includes when specified
cmake -DCUFFT=ON turns on inclusion of cufft libs and headers
INSTEAD of regular fftw-libs and headers. NOTE: this commit only
adds this swap-functuionality, the cuFFT usage is untested.
- - - - -
8acc756b by Dari Kimanius at 2015-07-23T00:01:56+02:00
Dedicated cuda stream for each backprojector.
Also model-data reduced by half hence freeing up more memory.
- - - - -
b776a196 by Dari Kimanius at 2015-07-23T10:03:26+02:00
Merge branch 'CUDA-Glob' of https://bitbucket.org/bforsbe/relion-1.4 into CUDA-Glob
- - - - -
9360714b by Dari Kimanius at 2015-07-23T11:08:19+02:00
Moved maximization to the top in store to put as much CPU-time between it and the next particle-iteration.
- - - - -
3ca8b21a by Dari Kimanius at 2015-07-23T11:19:15+02:00
Added some new shortcut functions for cuda mem-copies.
- - - - -
880b181c by Dari Kimanius at 2015-07-23T11:55:39+02:00
Removed extra call to generateEulerMatrices before backprojection and instead using previously generated matrices.
And some minor clean-ups.
- - - - -
2fa6277b by Bjoern Forsberg at 2015-07-23T14:11:36+02:00
make-warnings cleanup
- - - - -
b22a8845 by Dari Kimanius at 2015-07-23T17:16:10+02:00
First implementation of a global device memory managment for effective asynchronous kernel calls and memcpys.
New class added, named CudaDevicePtr. A light device pointer handler that can be included in none nvcc context and reuse allocated space.
All device pointers in maximization (wavg + BP) are replaced with more efficent reusable allocations.
However, the pointers are not reused for more than one particle, hence no significant perfomance gain yet.
- - - - -
a3b933de by Dari Kimanius at 2015-07-23T17:46:27+02:00
Merge branch 'CUDA-Glob' of https://bitbucket.org/bforsbe/relion-1.4 into CUDA-Glob
Fixed minor error impaling the results of BP and hence output data-file of iter 2, introduced in commit 880b181c.
Conflicts:
src/gpu_utils/cuda_projector_plan.cu
- - - - -
1fd6da6f by Bjoern Forsberg at 2015-07-25T18:13:21+02:00
re-implemented relion-native profiling
...for those parts relevant for the CPU vs. GPU comparisons
we are looking to make: Diff 1&2 + convert 1&2 + storeWS
(with sub-timing on BP)
NOTE: NOT TESTED WITH MULTITHREAD
Also: adjusted ctest refs to nethome-spcific files, so that
one can now run ctests 5 & 6 on tcblab machines with no
extra fuzz.
- - - - -
e701f3eb by Bjoern Forsberg at 2015-07-26T15:20:46+02:00
Forked cuda3Dprojector-method project to 2D/3D
In current state, 3D is assumed and used, so 2Dclassification
is still not functional. This is prep for 2D-capability.
see ticket #30
- - - - -
ec377f9b by Bjoern Forsberg at 2015-07-27T11:33:29+02:00
Forked Backprojector-method to 2D/3D
The documentations states that a malloc3Darray with a zero-depth
extent will just allocate a 2D array. We are then betting that the
cudaMemcpy3DParms Struct (doc: cuda 7.0 runtime API sect 4.8) can
also handle an implicit 2D cudaArray, and that the existing method
CudaProjector::setMdlData will just work wih mdlZ=0.
Also generalized the 3Dprojector object to Projector
See ticket #30
- - - - -
a01dda08 by Bjoern Forsberg at 2015-07-28T12:32:36+02:00
added apparently missing cuda-header
- - - - -
bf662d22 by Bjoern Forsberg at 2015-07-28T14:50:56+02:00
Template difference kernels based on 2D/3D-proj
This will keep a single diffCoarse-kernel in the source-code, but
expand the compiled code to include separate kernels for 2D- and
3D-projections, respectively. This to avoid an if-statement being
evaluated once for every pixel treated in the difference-kernels.
The above is also implemented for the diffFine-kernel.
There is probably some additional streamlining of the source code
to be done, for instance the calls to the templated functions are
runtime-dependent and stated explicitly for either case, both out
of necessity (the comiler should make both) and simplicity. It's
not too pretty though ( see cuda_helper_functions.cu:518:539, for
instance )
- - - - -
9e9b85bd by Bjoern Forsberg at 2015-07-29T10:38:03+02:00
Template wavg kernel based on 2D/3D-proj
see previous commit (bf662d2) for more info.
see ticket #30
- - - - -
2df5c6fc by Bjoern Forsberg at 2015-07-29T13:17:13+02:00
Reorder iclass > ipart loop in diff2Fine
This in preparation for movie-frame support, which will use
ipart as primary index for all associated objects;
-PassWeights
-ClassMasks
-ProjectionData
For this to work smoothly the ipart-loop (which iterates over
movie-frames in differenceCalculation-associated functions)
should be at the highest level to pick the object according
to those indices at (function-)global scope.
- - - - -
cc5c82e0 by Bjoern Forsberg at 2015-07-29T15:33:49+02:00
Reorder iclass > ipart loop in storeWS, step 1
The safest way to reorder the index hierarchy when the second
index is divided into two codeblocks 1 and 3, is to simply make
each block exhaustive, effectively finishing everything in
block 1 before ever reaching block 3. This comes at the expense
of doing block 0 as many times as we force exhaustive loops, in
this case twice:
-- original version -- iclass above ipart in partial blocks
for iclass
codeblock0
for ipart
codeblock1
cobelock2
for ipart
codeblock3
-- new version -- ipart above iclass in exhaustive blocks
for ipart
codeblock0
for iclass
codeblock1
codeblock2
codeblock0
for ipart
for iclass
codeblock3
NOTE! : THIS COMMIT ONLY FORCES EXHAUSTIVE,
AND DOES NOT REARRANGE THE INDEX HIERARCHY.
- - - - -
62218be2 by Bjoern Forsberg at 2015-07-29T16:06:04+02:00
Reorder iclass > ipart loop in storeWS, step 2
This commit does 2 fixes in one action. The action is
to switch the index hierarchy of the first storeWS loop-
block as follows;
-- old version --
for iclass
for proj_div
for ipart
-- new version --
for ipart
for iclass
for proj_div
the fixes are:
1. The old hierarcy implicitly assumes that the same
projdivision would work for all iparts, which is not
true, thus it could never have worked for more than
one ipart; it needs to be the lowest index
2. Allows the completion of all classes for each ipart,
which is preferrable since it is building an indexed
Array with coalesced access over all classes for each
ipart. To avoid injections and reallocations of alot of
data for each ipart, this is preferrable.
Also: One associated if(do_x && do_y)-block is replaced
by a if(!do_x || !do_y)-continue case.
- - - - -
19b3a67a by Bjoern Forsberg at 2015-07-29T17:41:22+02:00
Vectorization of Fine-pass weight objects
This commit represents the capability of the GPU-
based algorithm to handle sub-images of what in the
code is called and ori_particle, which is essen-
tially a particle which was picked in a cryo-EM
micrograph. Collecting movies instead om images
produces precisely such sub-images of an obesrved
particle, and thus this commit add movie-capability
to the current GPU-code.
NOTE: THIS IS AN UNTESTED FEATURE, SUPPORT IS
CURRENTLY THEORETICAL, AS ONLY SINGLE-IMAGE PART-
ICLE TEST HAVE BEEN DONE.
untested fix #38
- - - - -
af55b35f by Dari Kimanius at 2015-08-03T16:21:39+02:00
Moved rank-level initialization to MLOCuda-constructor, which now is called once initially at each optimization iteration.
Also more cudaMallocs and cudaFrees are moved to rank-level.
Memory issues regarding the vanilla code are inconsistently casuing problems to pop-up here and there, someone should look more deeply into this (later on).
- - - - -
9a783bfd by Dari Kimanius at 2015-08-04T12:32:51+02:00
The functionality for two benchmark output files is moved to a new build type, namely "Benchmarking".
Now the profiling build instead creates NVProf markers at the same points in code as the becnhmarking tool.
- - - - -
8783f651 by Dari Kimanius at 2015-08-04T16:22:19+02:00
Some more prepartions for better utilization of cuda streams.
Each CudaDevicePtr can be assigned to a stream, which is used in all copy operations.
- - - - -
9c788809 by bj.o.forsberg at 2015-08-05T06:54:52-04:00
Derived build types inherit props
- - - - -
9870fb40 by Bjoern Forsberg at 2015-08-06T15:18:18+02:00
Merge branch 'MovieSupport' into CUDA-Glob
Conflicts:
src/gpu_utils/cuda_ml_optimiser.cu
Some 90/180/360 offsets in metadata angles observed,
using non-movie data. No movie data has yet proven
that movie-support is functional, but planned modi-
fications require a common codebase.
Also, perhaps slighlty larger normcorr errors than usual.
This will be investigated
- - - - -
a12a7f9c by Bjoern Forsberg at 2015-08-07T11:03:43+02:00
Fixed mistakes in merge regarding benchmark timing
Only affects benchmark build, and only timings.
- - - - -
3eabcc25 by Bjoern Forsberg at 2015-08-10T10:08:26+02:00
Precompiler ifdef for cufft-compatibility
This will only allow compilation of the code with -DCUFFT=ON,
those binaries still suffer from NULL fftw_plans being returned
(see ticket #34 for details).
- - - - -
7d64512f by Dari Kimanius at 2015-08-10T14:12:00+02:00
Some syntax related changes in cuda-projector that makes the compilation more robust.
cuda_runtime.h is now included in pure GCC-context and void-types are removed.
- - - - -
290e7228 by Bjoern Forsberg at 2015-08-10T17:23:47+02:00
cmake is MPI-aware
mpicxx-includes are tcbsys-dependent, this is a FIXME
FLOAT macro is reserved in openmpi, S&R to XFLOAT in cuda-files
- - - - -
f9e9f394 by Bjoern Forsberg at 2015-08-10T17:34:36+02:00
Merge branch 'CUDA-Glob' into MultiDevice
- - - - -
5bc5b57a by Dari Kimanius at 2015-08-11T11:38:43+02:00
Some fixups for the backprojector to prevent read outside array boundary.
This however does not fix the memory errors.
- - - - -
fd55847d by Bjoern Forsberg at 2015-08-11T11:56:02+02:00
CudaMlOptimiser has device_id member
Device_id is default 0 and only set to anything different in the mpi-
compatible refine-code.
NOTE:
code compiles and works for
- non-MPI runs with non-MPI refine binary
- MPI-runs with one master and one slave
code does not run for
- multiple slaves
this is probably due to either initalization on devices != 0, or
cudaSetDevice not being MPI-threadprivate.
- - - - -
fc94a2d8 by Bjoern Forsberg at 2015-08-11T13:51:26+02:00
MPI-parallel device mngmt functional
(N+1) MPI-ranks will run N GPU-enabled threads, provided
there are M devices, where M>=N.
If N>M, the run will fail with a telling error message.
Possible improvements:
- if N>M, device_id could be set modulo M.
- if N>M, ranks>M could run gpu-disabled.
Mdulo will work out of the box, but take up uneccessary ammounts of
memory without increasing effficiency beyond latency hiding, at
this stage of development.
- - - - -
39bec21b by Dari Kimanius at 2015-08-11T17:36:32+02:00
Fixed issue related to projection division in "storeWeightedSums".
Also moved the the handling of the size of FinePassWeights to a new variable called newDataSize.
- - - - -
a19f5709 by Dari Kimanius at 2015-08-12T10:45:19+02:00
Fixes many of the valgrind warnings within the project.
A few still remain, however this commit fixes issue #39.
- - - - -
69aa2a38 by Bjoern Forsberg at 2015-08-12T11:19:51+02:00
Merge branch 'MultiDevice' into CUDA-Glob
Conflicts:
src/gpu_utils/cuda_backprojector.cu
src/gpu_utils/cuda_helper_functions.cu
src/gpu_utils/cuda_ml_optimiser.cu
- - - - -
65f3ef6a by Dari Kimanius at 2015-08-12T13:22:32+02:00
Backprojector setup for 2D calssification.
The CPU implementation of the 2D backprojector is used for now. Data is fetched from GPU and passed to the existing algorithm.
A member variable is added to cudaMLO, named modelIs3D, that is true during 3D and false during 2D classification.
- - - - -
5ad4beab by Bjoern Forsberg at 2015-08-13T14:46:14+02:00
Forcing 2D runs to abide to psi-sammplings mod 32
Some kernels in the convert-stage of the coarse pass currently
relies on the number of orietnations to be divisible by 32, since
this is an efficient thread-wise division of work. During 2D-
classification the diretional sampling provided by the Healpix
mesh is not factored in, yielding very low orientational numbers,
the default being every 10degrees => 36 sampling points. This will
unfortunately not be compatible with current kernels.
This commit fixes this in the simplest and fastest way possible,
by forcing psi_step to increments which produce 0 mod 32 orienation
numbers. The default is then implicitly modified as
10degrees => 36 sampling points , ROUND(/32) -> 32 sampling points,
so psi_step is modified to 360 / 32 = 11.25
NOTE: this commit has issus with texture interpolation, so to see this
edit targeting issue #40, double precision build is required.
- - - - -
7a2529ba by Dari Kimanius at 2015-08-13T16:25:10+02:00
Added a custom cuda allocator and reverted memory holders from CudaDevicePtr back to CudaGlobalPtr.
The allocator is not used yet and the pointers are allocated using the default cuda allocator.
Removed CudaDevicePtr
Removed cuda_helper_functions.cuh since it was never used.
Removed "--default-stream per-thread" from the NVCC flags since we are moving towards custom stream managment.
- - - - -
d2443078 by Dari Kimanius at 2015-08-13T16:27:27+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-1.4
- - - - -
d1a63eb6 by Dari Kimanius at 2015-08-13T17:25:25+02:00
The allocator is now working as expected. *pops champagne*
- - - - -
5f619001 by Dari Kimanius at 2015-08-13T19:00:50+02:00
The allocator is distributed out to the pointers in the code. It is however not activated since this cased an unknown issue.
Some helper functions had there parameters modified to make them more generic considering input pointers.
- - - - -
dd1efc33 by Dari Kimanius at 2015-08-14T10:19:39+02:00
The last bugs in the allocator was found and removed.
Hence also it was activated for the entire cudaMlOptimiser.
- - - - -
f06e249f by Dari Kimanius at 2015-08-14T13:45:25+02:00
Moved translation and scaling (ipart dependent) in maximization to only be done once for every ipart.
Added getFreeSpace-function to the custom allocator.
- - - - -
12711197 by Bjoern Forsberg at 2015-08-14T14:11:28+02:00
2D classification proj uses textures correctly
The layout for object texture creation follows very different methods,
opposite to remarks made in commit ec377f9. The hope was that the
existing method would be able to create a 2D-texture capable of using
bilinear interpolation, as per the CUDA documentation. This commit
resorts to the 'default' method of allocating a pitched (aligned) cuda-
array and uses bilinear interpolation intrinsics during 2D classfication.
fixes issue #40
- - - - -
97f4ca41 by Bjoern Forsberg at 2015-08-14T14:17:08+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-1.4
- - - - -
cb102d0b by Bjoern Forsberg at 2015-08-14T16:35:06+02:00
Objects shared by classes moved out of class-loop
Addresses ticket #43
- - - - -
ee297eae by Dari Kimanius at 2015-08-17T16:23:26+02:00
Fixed Issue #42 (2D texture object segfault on free). Also added some new stream synchronizations where needed.
The program runs through two iterations, however the output file is not correct.
- - - - -
d25a1c5d by Bjoern Forsberg at 2015-08-17T17:01:25+02:00
Index fix in 2D priors & psi-domain control
An array reduction was wrongfully using the iorient index to read from a
reference an array, whereas it should be indexed by a pure iteration var.
The reulting reduction was used as prior in iterations>=2 during 2Dclass-
ification, so no 3D test up until now have been sensitive to this error.
Also, the first iteration produced correct meta-data, apart from the psi-
angle which was set to the [-180,180[ domain for the CPU-only code, where
the GPU code gave output in th [0,360[ domain. The reason is not fully
clear but an explicit control has been added regardless of using GPU hard
ware or not.
NOTE : the psi-angle domain control should be tested on iteration >2 when
possble, to rule out any possible issues with the extra domain check in
both 2D- and 3D-classification.
this commit resolves ticket #43, but there are further errors persiting,
which are described in ticket #46
- - - - -
b6630a59 by Bjoern Forsberg at 2015-08-17T18:32:18+02:00
2D classif. runs 2 iter w/o metadata errors
In order to swiftly implement 2D BackProjection, this was placed on the CPU
hardware. Due to this, the copying back of an unused device BackProjector
caused the CPU-Backprojector to be overwritten, resulting in an essentially
NULL comparison between images, since they were compared to all-zero models.
The only difference between weights was then due to the error associated
with interpolating pixel-values of the rotated images.
NOTE: there are still accuracy errors on single precision, the "w/o errors"
declared above refers to the obvious erros in the method implementation.
fixes #46
- - - - -
2656ecf5 by Dari Kimanius at 2015-08-18T18:54:46+02:00
Added prioritized streams to run backprojects on a low priority stream in parallel with next particle iteration.
Moving to compute capability 5.2
Added a out of memory handler to free up backprojection datat in case of memory shortage.
- - - - -
40ef0319 by Bjoern Forsberg at 2015-08-18T20:01:08+02:00
Moved two thrust-wrapped reducts out of class-loop
It was noticed that there was a large overhead associated with
launching thrust-kernels. In the 3D-case this overhead is negligble
compared to the time it takes to run the reductions, however, in
the 2D case the problem size is so much reduced that theis is no
longer the case. This commit represent edits to the code which
benefits the 2D-classification by grouping reductions into a single
call over all classes. It will benefit the 3D-classification to
the same absolute extent, however that might not even be noticable.
The changes are made to thrust reductions
- min(diff) @ end of diff2coarse
- max_element(weight) @ end of storeWS
- - - - -
9d111135 by Bjoern Forsberg at 2015-08-18T21:53:55+02:00
Sumweight thrust-reduction moved out of class loop
Targets 2D refinement efficiency @ low block count
NOTE: only tested on 2D refinement
- - - - -
af8d3811 by Bjoern Forsberg at 2015-08-19T10:18:15+02:00
Reorder ipart/iclass loops in diff2Coarse
In preparation for global thrust-min reduction of diffs
- - - - -
671a341d by Dari Kimanius at 2015-08-19T10:28:20+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-1.4
Conflicts:
src/gpu_utils/cuda_ml_optimiser.cu
- - - - -
12e34805 by Bjoern Forsberg at 2015-08-19T10:42:36+02:00
Thrust-min moved out of class-loop in diff2Coarse
This is to reduce the number of thrust calls, effectively removing
their dependence on the number of classes. This does come with add-
itional memory requirements on the device, but nothing dramatic in
the 2D case;
estimations of typical settings:
2D : 134'400 diff-vals / ipart
32 orients
21 translations
200 classes
3D : 967'680 diff-vals / ipart
192 * 24 orients
21 translations
10 classes
- - - - -
f6187f83 by Dari Kimanius at 2015-08-19T11:02:19+02:00
Added the the CUB headers to the project.
- - - - -
3f99edbc by Dari Kimanius at 2015-08-19T11:03:05+02:00
Replaced thrust with CUB for the max-search in the metadata-section.
- - - - -
7eef6ca8 by Bjoern Forsberg at 2015-08-19T11:52:26+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-1.4
- - - - -
5e0ec00b by Bjoern Forsberg at 2015-08-19T17:29:58+02:00
Some device<->host cpys are coalesced
Mostly using temporary device pointer children as kernel input
- - - - -
1194e0f8 by Dari Kimanius at 2015-08-19T18:29:52+02:00
Replaced thrust with CUB everywhere in the project.
- - - - -
08150654 by Bjoern Forsberg at 2015-08-19T18:43:50+02:00
diff2coarse avoids repeating memcpys
There is a lingering if()-case which precludes copying the calculated
diffs straight into op.Mweight, which would avoid a mapping which for
a 2D run of 10 classs takes as long ass all 10 kernels.
- - - - -
5a0d3bdb by Dari Kimanius at 2015-08-19T18:44:33+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-1.4
Conflicts:
src/gpu_utils/cuda_ml_optimiser.cu
- - - - -
b9d57d45 by Bjoern Forsberg at 2015-08-19T19:05:23+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-1.4
Conflicts:
src/gpu_utils/cuda_ml_optimiser.cu
- - - - -
6ccd6155 by Bjoern Forsberg at 2015-08-19T19:22:59+02:00
Minor fix to merge & one sync removed
- - - - -
933a7533 by Bjoern Forsberg at 2015-08-20T15:37:14+02:00
Sumweight priors are prepared for all classes
This is to
- short-term: eliminate memcpy-ops between kernels calls
- long-term: allow the sumweight-kernels to be grouped into
one massive kenrel call for all-classes at once
ALSO: fix to indexing in diff2coarse mapping between op.Mweight
and 'AllWeights'
- - - - -
696d91c6 by Dari Kimanius at 2015-08-20T17:11:09+02:00
"COLLECT 2 AND SET METADATA" is moved back before of "MAXIMIZATION" in execution order.
MlOptimiserCuda is plugged in as outOfMemoryHandler.
Custom allocated space is increased to 90% of remaining device space after.
New memory analysis tools added to allocator.
- - - - -
d7c9fec6 by Dari Kimanius at 2015-08-20T17:11:26+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-1.4
- - - - -
1ae1b1b9 by Bjoern Forsberg at 2015-08-20T18:48:28+02:00
coalesced some memcpy-ops of collect2
Some, which are harder to coalesce in the present state, were moved
outside the kernel-calling loop, so as to enable coalescing of the
collect2jobs-kernel in the long term.
- - - - -
29305420 by Bjoern Forsberg at 2015-08-20T20:19:39+02:00
last memcpys between collect2-calls removed
Just like some recent commits, this is to
- short-term: eliminate memcpy-ops between kernels calls
- long-term: allow the sumweight-kernels to be grouped into
one massive kenrel call for all-classes at once
NOTE that his changes the affected part of storeWS from
for class
for ipart
preprocess_the_stuff()
send_the_stuff()
calculate_the_stuff()
fetch_the_stuff()
post_process_the_stuff()
to
for ipart
for iclass
preprocess_the_stuff()
send_the_stuff()
for iclass
calculate_the_stuff()
fetch_the_stuff()
for iclass
post_process_the_stuff()
- - - - -
3f289b17 by Dari Kimanius at 2015-08-23T23:43:26+02:00
Added some error handling wrapping the cub function calls.
- - - - -
0ee0fcc4 by Dari Kimanius at 2015-08-24T17:48:32+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-1.4
- - - - -
8bf6186b by Dari Kimanius at 2015-08-24T18:23:10+02:00
Custom allocator can be switched off with the cmake option "CustomAlloc=OFF" (default is ON).
Reverted device arch back to sm_32, since parallel kernel execution is omitted.
- - - - -
d829cd15 by Dari Kimanius at 2015-08-24T20:05:18+02:00
More debug-warning handling added to CUB wrapper functions.
- - - - -
66c0f22e by Dari Kimanius at 2015-08-26T11:15:05+02:00
Fixed error in free-type of temporary arrays in backprojection data extraction.
Also added a new constructor for CudaGlobalPtr for assigning sub regions of pointers.
- - - - -
6787d4ef by Dari Kimanius at 2015-08-26T11:22:35+02:00
Moved the following classes from cuda_utils.cuh to cuda_ml_optimiser.h: IndexedDataArrayMask, IndexedDataArray and ProjectionParams
- - - - -
8f7347ea by Bjoern Forsberg at 2015-08-26T15:27:23+02:00
failsafe-throwback : STL-based array ops
Also fix to indexing in convert (coarse pass) inherited dev-ptr
- - - - -
e42ef018 by Dari Kimanius at 2015-08-26T16:33:54+02:00
Separated cuda_utils into one with CUB and one with thrust.
Atomic add device-function definition was moved to a separate file called cuda_device_utils.cuh
Inclutions of cuda_utils in every file was removed, since none use these functions.
- - - - -
fc7a78e4 by Dari Kimanius at 2015-08-26T17:20:12+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-1.4
Inserted thrust instead of stdlib for sorting, sum, min and max search.
Conflicts:
src/gpu_utils/cuda_backprojector.cu
src/gpu_utils/cuda_benchmark_utils.cuh
src/gpu_utils/cuda_helper_functions.cu
src/gpu_utils/cuda_kernels/diff2.cuh
src/gpu_utils/cuda_kernels/helper.cuh
src/gpu_utils/cuda_kernels/wavg.cuh
src/gpu_utils/cuda_ml_optimiser.cu
src/gpu_utils/cuda_projector.cu
src/gpu_utils/cuda_projector.cuh
src/gpu_utils/cuda_projector_plan.cu
- - - - -
597610c6 by Bjoern Forsberg at 2015-08-27T10:54:18+02:00
rm:ed excessive memcpy and CPU-based reduction
in wdiff2-reduction (storeWS) which occurs over all classes.
- - - - -
9068a3b7 by Bjoern Forsberg at 2015-08-27T11:01:11+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-1.4
- - - - -
af1520cc by Dari Kimanius at 2015-08-27T11:51:06+02:00
Rearranged memory allocation for the backprojector data bundle to the end of the custom managed device memory.
Data bundles are now arranged in a stack rather than a vector.
- - - - -
8d240eff by Dari Kimanius at 2015-08-27T11:58:29+02:00
Removed legacy if-statement in 'SET METADATA'.
- - - - -
77efaa30 by Bjoern Forsberg at 2015-08-27T14:49:20+02:00
fixes to STL-based utils (fallback utils)
- - - - -
0fb65041 by Bjoern Forsberg at 2015-08-27T14:50:20+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-1.4
- - - - -
1a075715 by Dari Kimanius at 2015-08-27T15:25:27+02:00
Minor changes in BP kernel, yielding roughly 20% speed-up in kernel execution time.
- - - - -
946384ab by Bjoern Forsberg at 2015-08-27T16:19:18+02:00
Coalesced some diff2Fine-memcpys
prep for coalesced kernel call in 2D (and possibly 3D)
- - - - -
ab096149 by Dari Kimanius at 2015-09-01T16:01:11+02:00
2D backprojector ported to GPU.
A scatter approach was selected due to the homogenious write load to output.
- - - - -
64c023bb by Dari Kimanius at 2015-09-01T16:03:09+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-1.4
- - - - -
75593cea by Bjoern Forsberg at 2015-09-01T17:35:29+02:00
new class cudaStager to coalesce _small_ mem-ops
Given a number of cudaGlobalPtr:s wit hthe same type, they can be
handed off to a stager which will move their memory positions into
a coalesced memory space and transfer them simultaneously.
A number of improvements are possible:
- optional use of pinned memory as staging space
this relies on whiching to management of a copy of the supplied
cudaGLobalPointer, rather than reassigning the only existing copy
- allocate space only once total size is known
this relies on the development of an appendable list
of cudaGlobalPtr:s to reassign host- and device-pointers of, at stage-
time.
- - - - -
88670bdc by Bjoern Forsberg at 2015-09-01T17:43:45+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-1.4
- - - - -
57912701 by Bjoern Forsberg at 2015-09-02T17:01:23+02:00
Memleak fixed
Fixed issue #51
- - - - -
7ee19b31 by Bjoern Forsberg at 2015-09-03T12:39:59+02:00
Third iteration priors set correctly
A for-loop was missing an incremental offset to its max.
Fixed issue #52
- - - - -
6e54bc8c by Bjoern Forsberg at 2015-09-03T15:24:10+02:00
Additional psi domain fix with GPU-code.
Peripherially adresses issue #53
- - - - -
269cf899 by Dari Kimanius at 2015-09-03T18:32:30+02:00
3D backproject is optimized with a factor ~200 (yes, it's freckin' 200 times faster now).
I'm very very sad that I didn't try this 5 month ago (if only I knew how efficient atomic adds are).
- - - - -
9c13effd by Dari Kimanius at 2015-09-03T18:41:33+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-1.4
Conflicts:
src/gpu_utils/cuda_ml_optimiser.cu
- - - - -
b1efabea by Dari Kimanius at 2015-09-04T11:28:45+02:00
Actually activated the new 3D BP.
- - - - -
f2a6217a by Bjoern Forsberg at 2015-09-05T11:41:18+02:00
Added missing loop for norm-corr factor
This was somehow glanced over and manifested as a small error which
could be attributable to larger particle groups being affected by a
different order of summation on the GPU comapred to the CPU.
NOTE 1:
The errors in METADATA which this commit adresses only appear when
multiple micrographs are used.
NOTE 2:
There is still some (reduced) deviation from CPU normcorr-values
when using double prec. GPU code. This could be additional factors
which are unaccounted for.
NOTE 3:
Using a .star-file which conatins 100s of particles from a single
micrograph and a single particle from another micrograph, the error
of the norm-correction of the single particle compared to CPU-code
is much higher than that of other particles. This (possible issue
is entered as issue #54
Fixed issue #53
- - - - -
6790a40d by Bjoern Forsberg at 2015-09-05T12:02:07+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-1.4
- - - - -
ea6b30b8 by Bjoern Forsberg at 2015-09-08T13:48:43+02:00
Fixed multi-class reduction issues
Issues fixed:
- One legacy zero-initialization kept resetting a reduced variable.
- One legacy if-clause (which was unneccessary) was given a wrong
index to fetch a boolean.
fixed issue #55
- - - - -
b23e1eaf by Bjoern Forsberg at 2015-09-08T15:20:07+02:00
Fix to wavg-kernel scale correction
The scale correction was included in the ctf-object in order to avoid
excessive multiplications. However, the ctf-correction was dependent
on a an if-clause inside the kenrel, and consequently the scale-corr
was only applied to the references if the ctf was also applied.
this fixes issue #54
- - - - -
4428618f by Bjoern Forsberg at 2015-09-08T17:43:17+02:00
moved sync out of class loop
Result of bad implementation in commit f2a6217
NOTE: extreme degradation of gpu efficiency without this edit:
this might be worth studying closer.
fixes issue #56 (documentation issue)
- - - - -
ff08d28b by Dari Kimanius at 2015-09-09T14:50:23+02:00
Ported translations to CUDA. However, data is sent back end forth between GPU and RAM so no speed-up yet.
Also fixed an error in backproject initialization step that occured with double build.
- - - - -
6c2a43a3 by Bjoern Forsberg at 2015-09-10T10:21:39+02:00
CC-support in diff-kernels (draft)
May be severely limited by shared mem, but this is mostly because
sjors indicated that --do_firstiter_cc might be needed, and the first
iteration is hardly a priority at this stage
adresses issue #11
- - - - -
b1a49228 by Dari Kimanius at 2015-09-10T11:13:42+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-1.4
Fixed error in translation caused by misuse of scales in SWS.
- - - - -
b909997a by Bjoern Forsberg at 2015-09-10T11:19:12+02:00
CC-support in convert (draft) + fix compile errs
Setting of weights on device handled by two async memset-calls,
the latter of which handles a single value, which might be inefficient
also fixed compilation errors from parent commit
- - - - -
383966ad by Bjoern Forsberg at 2015-09-10T12:44:21+02:00
Added missing is-statement in partial class-loop
The major class-loop in SWS has been partitioned into pre-process,
memory transfer, kernel-calls, and post process loops. In the final
such loop, an if-statement was missing which may have caused some
segfault under certain circumstances. It is unclear wheter this is
the fix to, but it was certainly found because of, an error similar
to that described in issue #57
- - - - -
f1468275 by Bjoern Forsberg at 2015-09-11T12:05:34+02:00
Cross-correlation works (CC)
The kernels are NOT optimised for --always_cc, but rather just
to be compatible with the need for --firstiter_cc, whereunder we
simply assume that a stupid implementation will not affect the over-
all perfomrnace of an entire classification/refinement run.
NOTE: A third factor in the GPU-based image-correction applied to
the difference metric is introduced, which should cause some concern;
the accuracy might in some cases suffer. This commit leaves an
alternate availiable; however commented out. op.local_sqrtXi2 can be
applied either within the buildCorrImage, or at the end of CC-based
difference kernels. The present approach is the former.
fixes issue #11, creates optimisation issue #58
- - - - -
467e482a by Bjoern Forsberg at 2015-09-11T13:27:57+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-1.4
Conflicts:
src/gpu_utils/cuda_helper_functions.cu
src/gpu_utils/cuda_ml_optimiser.cu
- - - - -
6df1feb9 by Dari Kimanius at 2015-09-11T13:53:49+02:00
Added the translation-class and the plan nested class for handling of transaltion-data and kernel calles.
Some clean-up of old code in BP.
- - - - -
f7f1706e by Dari Kimanius at 2015-09-11T14:52:50+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-1.4
Conflicts:
src/gpu_utils/cuda_ml_optimiser.cu
- - - - -
1ab45d84 by Dari Kimanius at 2015-09-11T14:53:42+02:00
Data prefetch to GPU memory implemented in GSD-fine and SWS.
- - - - -
ed9083db by Dari Kimanius at 2015-09-11T16:06:51+02:00
Minor clean-up of old translation leftovers.
- - - - -
98d5e100 by Dari Kimanius at 2015-09-14T16:50:08+02:00
Removed xdiff reduce by introducing atomic adds to wavg.
This have rendered a slight speed-up but is mainly a preperation for a kernel call for a classes at once.
- - - - -
5ed88444 by Dari Kimanius at 2015-09-14T17:15:12+02:00
Removed proj_diff and the second euler generation between wavg and BP by reusing the first one.
- - - - -
71b195f2 by Dari Kimanius at 2015-09-14T17:41:01+02:00
Some minor clean-up in wavg-kernel.
- - - - -
82c2e63d by Bjoern Forsberg at 2015-09-15T09:49:37+02:00
Separate control of doff2-coarse block size
- - - - -
ac848b06 by Bjoern Forsberg at 2015-09-15T09:51:01+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-1.4
- - - - -
ea586e6b by Dari Kimanius at 2015-09-15T12:07:35+02:00
Cuda-devices are now associated with threads instead of MPI-rank.
To run on 4 GPUs use "--j 4 --gpu".
- - - - -
f854a06f by Bjoern Forsberg at 2015-09-15T13:17:21+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-1.4
- - - - -
5edaca78 by Bjoern Forsberg at 2015-09-15T14:04:06+02:00
removed unused collect2, enable wavg block_size
- - - - -
79995a47 by Bjoern Forsberg at 2015-09-15T15:03:28+02:00
sumweight coarse update, wavg sync
- - - - -
7cec93d6 by Bjoern Forsberg at 2015-09-15T15:44:40+02:00
streamline SWC-kernel
- - - - -
712e05f6 by Dari Kimanius at 2015-09-15T16:08:46+02:00
Better opbject integrity for CudaGlobalPtr.
Added a setter for device and host memory pointer.
Added setter and getter for size.
Added operator () for fetching device-pointer, like the corresponding [] for the host pointer.
- - - - -
c9035928 by Dari Kimanius at 2015-09-15T16:37:09+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-1.4
Conflicts:
src/gpu_utils/cuda_ml_optimiser.cu
- - - - -
13d6ba2f by Bjoern Forsberg at 2015-09-15T17:21:26+02:00
killed GPU_T*Cs + 2D test on nethome/empiar
- - - - -
ac4211d0 by Bjoern Forsberg at 2015-09-15T17:22:06+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-1.4
- - - - -
a296be02 by Bjoern Forsberg at 2015-09-15T17:50:22+02:00
added 2d test & forceSTL compile option
- - - - -
c2b002d4 by Bjoern Forsberg at 2015-09-16T11:23:27+02:00
--gpu flag can take string of gpu_ids
--j N --gpu
where N is an integer, results in default behaviour;
trying to set the dev_id to the (--j) thread id
--j N --gpu 00...00
where 00...00 is N zeros, results in putting all threads on device 0.
--j N --gpu 01...(N-1)N
is analogous to default behaviour.
example; want 2 threads on each of 4 available devices;
--j 8 --gpu 00112233
or
--j 8 --gpu 01230123
or
...
Possible reasons for errors:
NOTE: device ids start with 0 !
Trying to specify more threads than devices without specifying id's
will result in explicit errors. This will also happen if an id larger than
the available device number is specified. For example, with 4 available
devices:
--j 5 --gpu => error, since this will attempt dev_id=thread_id>avail_devs
--j 4 --gpu 1234 => error, since dev_id=4 <=> device no. 5
- - - - -
051fbadf by Bjoern Forsberg at 2015-09-16T17:30:03+02:00
cuda-owed getFour...AndCtfs copy & timings
- - - - -
76b40983 by Dari Kimanius at 2015-09-17T13:45:13+02:00
Added some new funtions to CudaGlobalPtr
- - - - -
58f0efc4 by Dari Kimanius at 2015-09-17T13:46:55+02:00
BP data is transfered from GPU to RAM thread safe.
- - - - -
6f2e316e by Dari Kimanius at 2015-09-17T14:21:17+02:00
First fourier transform in getFourierTransformsAndCtfs is gpu-accelerated. 2x speed-up of this subrutine.
- - - - -
2000c5b7 by Bjoern Forsberg at 2015-09-17T16:33:03+02:00
SoftMaskOutside on GPU
until tested by time, a bool remains to switch on native (CPU) function.
ALOT of optimization to be done, at the moment a ~2-3x speedup of
SoftMaskOutside-block and wrapping ZeroMask-block.
This could be reduced by half; A kenrel which prepares a [0,1]-domain
filter for all particles could devote SoftMaskOutside() to figuring out
just the scaling and apply a pix-by-pixel filter without the need to
check all norms and range-conditionals within the kernel.
- - - - -
eed5a432 by Bjoern Forsberg at 2015-09-17T16:34:02+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-1.4
- - - - -
38983937 by Bjoern Forsberg at 2015-09-17T18:42:14+02:00
SumW-kernels overlap, using cudaMLO class-streams
In order to keep the number of streams down, the classstreams
were made a member of the cudaMLO-class.
NOTE : at the moment we use DeviceSyncs after the only class loop
which uses this class-streaming. This might be bad, so we should
think about some way of syncing just these streams before continuing
along the CPU-code.
- - - - -
6b494a2a by Bjoern Forsberg at 2015-09-17T19:19:07+02:00
class-streams for diff2-kenrels
NOTE: using devicesyncs
- - - - -
6210b71e by Dari Kimanius at 2015-09-18T16:12:31+02:00
Added some currently unused functions for future reference.
- - - - -
acb5b573 by Dari Kimanius at 2015-09-18T16:43:02+02:00
"onthefly_shifts" is automatically set to true when do_gpu is true.
Also do_gpu now works in "continue" runs.
- - - - -
00bada0f by Dari Kimanius at 2015-09-18T16:44:59+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-1.4
- - - - -
f9d251f5 by Dari Kimanius at 2015-09-21T15:55:51+02:00
Removed the use of void pointers in MetaDataContainer to fix some memory leaks reported by valgrind.
- - - - -
ca0f4bda by Dari Kimanius at 2015-09-21T18:48:26+02:00
Disabled cufft (commented out) and fixed some other bits of code causing valgrind reports.
- - - - -
cb1e6f98 by Dari Kimanius at 2015-09-21T18:59:29+02:00
The gpu memory (in GBs) reserved for each thread can now be set by the new flag "--gpu_memory_per_thread".
- - - - -
eb265230 by Dari Kimanius at 2015-09-24T11:03:28+02:00
"HANDLE_ERROR" is now active in release build, while the old debug error handler is renamed to "DEBUG_HANDLE_ERROR".
- - - - -
8f9fc313 by Dari Kimanius at 2015-09-24T11:07:20+02:00
Prepartion for the new setup where GPU resources (streams, textures and allocations) are reused over all the iterations.
- - - - -
852f3a65 by Bjoern Forsberg at 2015-09-24T15:05:13+02:00
Rounding psi_sampling up instead of to nearest
NOTE: still forcing nr_psi%32==0, just being careful rather than
approximately the same
- - - - -
2721a691 by Dari Kimanius at 2015-09-28T18:01:13+02:00
Non-working first commit
- - - - -
3d9c4751 by Dari Kimanius at 2015-09-29T13:09:23+02:00
Some minor changes.
- - - - -
c28d94b0 by Dari Kimanius at 2015-09-29T13:25:01+02:00
Merge branch 'reused-gpu-resources'
- - - - -
b87c4345 by Dari Kimanius at 2015-09-29T18:23:25+02:00
Fixed error occurring due to onthefly_shifts being switched off during auto-refine.
The translators are by-passed in case the onthefly_shifts-flag is not true. Instead the translated images are used.
However, auto-refine still dosn't work due to some other error ("trying to host-alloc a stager with size=0")
- - - - -
1144ac61 by Bjoern Forsberg at 2015-09-30T12:39:15+02:00
binaries prepended by 'relion_' as originally
- - - - -
01f08d68 by Bjoern Forsberg at 2015-09-30T15:07:45+02:00
CMakeLists-files updated
- - - - -
d094ea3a by Dari Kimanius at 2015-09-30T20:04:39+02:00
The memory assigned to custome allocator is set to 70% of available device memory if gpu-memory-flag is not set.
- - - - -
3ca9600b by Dari Kimanius at 2015-09-30T20:06:28+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-1.4
- - - - -
24ae8be6 by Dari Kimanius at 2015-10-05T11:46:59+02:00
Texture can be turned on and off in single precision compilation using the cmake option CudaTexture.
However selecting double will overwrite this option to non-texture compilation.
- - - - -
c7a07b75 by Bjoern Forsberg at 2015-10-05T15:47:05+02:00
GUI bool option in cmake
- - - - -
ba03303b by Bjoern Forsberg at 2015-10-05T15:47:41+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-1.4
- - - - -
13c6de34 by Dari Kimanius at 2015-10-06T17:59:58+02:00
Devices are automatically mapped to threads in a smarter fashion.
If nothing provided (even if followed by another flag [previous bug]) the devices are evenly distributed amongs threads.
- - - - -
cfeb544d by Bjoern Forsberg at 2015-10-07T09:59:26+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-1.4
- - - - -
b324b9d9 by Bjoern Forsberg at 2015-10-07T14:48:08+02:00
Minor changes to get FourierAndCtfs
no changes to functionality
- - - - -
bf9f47b8 by Bjoern Forsberg at 2015-10-07T16:57:05+02:00
turn off gpu-tic/tocs & ml_optimiser-bugfix
- - - - -
c70ce37d by Bjoern Forsberg at 2015-10-08T14:10:28+02:00
local searches tagcase works w/o custom allocator
- - - - -
236c78b7 by Dari Kimanius at 2015-10-08T15:04:50+02:00
Custom allocator assignes aligned memory, dynamically adjusted to device specifications.
- - - - -
3c9a74be by Bjoern Forsberg at 2015-10-08T15:55:30+02:00
Updated warns and non-0 mod 32 compatibility
Previously thew orientational sampling was implicitly assumed to be
0 mod 32 in the the coasre pass due to the healpix sampling. When
doing 2D classification the psi_step dictates the number of increments,
the default being 10degrees => 36 steps. The quick and dirty fix was to
force 0 mod 32 by adjusting psi_step to fulfil psi_step = 360/(N*32)
for some N.
During local searches and/or auto-refine 32-multiple coarse psi-sampling
might not be fulfilled or even desirable for effciciency. To circumvent
this we have now allowed any psi-sampling, at the cost of wasted cycles
on the GPU. A cleaner fix from the point of view of functionality, but
still not the computationally sleakest fix.
ALSO some cleanup of unused class_ids in IndexedDataArray
- - - - -
36166982 by Bjoern Forsberg at 2015-10-08T16:04:29+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-1.4
- - - - -
11b20c65 by Dari Kimanius at 2015-10-08T16:54:37+02:00
Translationa are done on the fly inside diff2coarse-kernel. This yields a slowdown, but removes L2 bottleneck.
- - - - -
a4e3e2d0 by Dari Kimanius at 2015-10-09T18:07:24+02:00
Multiple projection per block yeilding a factor-2 speed up.
- - - - -
ad94cde4 by Bjoern Forsberg at 2015-10-12T14:31:11+02:00
double rplaced by RFLOAT
locations where Sjors preserved double have been similarly preserved.
- - - - -
db2228e1 by Dari Kimanius at 2015-10-12T14:49:02+02:00
Fixed output error caused by synchronization issue.
- - - - -
12a2cd6c by Bjoern Forsberg at 2015-10-12T15:04:03+02:00
sjors dev-code @a7f194b (src)
- - - - -
ffb71977 by Bjoern Forsberg at 2015-10-12T15:57:49+02:00
Merge branch 'sjors-dev' into sjorsmerge
Conflicts:
src/Healpix_2.15a/cxxutils.h
src/apps/CMakeLists.txt
src/apps/image_handler.cpp
src/apps/project.cpp
src/apps/stack_create.cpp
src/assembly.cpp
src/autopicker.cpp
src/autopicker.h
src/backprojector.cpp
src/backprojector.h
src/complex.h
src/ctf.h
src/ctffind_runner.cpp
src/displayer.cpp
src/euler.cpp
src/exp_model.cpp
src/exp_model.h
src/fftw.cpp
src/fftw.h
src/filename.cpp
src/funcs.cpp
src/funcs.h
src/gcc_version.h
src/gpu_utils/cuda_backprojector.cu
src/gpu_utils/cuda_benchmark_utils.cuh
src/gpu_utils/cuda_kernels/diff2.cuh
src/gpu_utils/cuda_kernels/helper.cu
src/gpu_utils/cuda_mem_utils.h
src/gpu_utils/cuda_ml_optimiser.cu
src/gpu_utils/cuda_ml_optimiser.h
src/gpu_utils/cuda_projector.cu
src/gpu_utils/cuda_projector.cuh
src/gpu_utils/cuda_projector.h
src/gui_entries.h
src/gui_jobwindow.cpp
src/gui_jobwindow.h
src/healpix_sampling.cpp
src/healpix_sampling.h
src/image.cpp
src/image.h
src/macros.h
src/matrix2d.h
src/metadata_container.cpp
src/metadata_container.h
src/metadata_label.h
src/metadata_table.cpp
src/ml_optimiser.cpp
src/ml_optimiser.h
src/ml_optimiser_mpi.cpp
src/ml_optimiser_mpi.h
src/mpi.cpp
src/mpi.h
src/multidim_array.h
src/numerical_recipes.cpp
src/numerical_recipes.h
src/particle_polisher.cpp
src/particle_polisher.h
src/particle_polisher_mpi.cpp
src/particle_sorter_mpi.cpp
src/preprocessing.cpp
src/preprocessing.h
src/projector.cpp
src/rwIMAGIC.h
src/rwMRC.h
src/rwSPIDER.h
src/strings.cpp
src/strings.h
src/tabfuncs.cpp
src/time.cpp
src/transformations.h
- - - - -
f9735ee7 by Bjoern Forsberg at 2015-10-12T16:04:50+02:00
Cmake fixes
- - - - -
95e90613 by Dari Kimanius at 2015-10-12T18:35:31+02:00
More control over data flow in diff2 kernel.
- - - - -
2b2ff00a by Bjoern Forsberg at 2015-10-13T09:30:00+02:00
Removed double device set
healpix formalism corrected
- - - - -
8524c6a8 by Bjoern Forsberg at 2015-10-13T15:50:56+02:00
relocated gpu device-set and added timings
- - - - -
e83c4519 by Bjoern Forsberg at 2015-10-13T15:51:35+02:00
printf specifier
- - - - -
4cd6de0b by Bjoern Forsberg at 2015-10-13T15:51:54+02:00
strcpy overflow
- - - - -
9c6276ba by Bjoern Forsberg at 2015-10-13T16:07:52+02:00
return objs from sys-calls added
to remove compile-time warnings
- - - - -
0a3b4b28 by Dari Kimanius at 2015-10-13T19:09:02+02:00
Better control of statics properties of the diff2-coarse-kernel.
Optimal properties set, should be tested for multiple systems.
- - - - -
dd3a1c93 by Bjoern Forsberg at 2015-10-14T13:22:25+02:00
cub appears to work, replaces thrust
The identified and amended issue was two-fold, and roughly independent.
The custom allocator was not correctly aligning memory addresses, resulting
in wrong and/or misaligned adressing.
Switching to the default allocator, the temporary storage size got set to
zero for small sorts, which resulted in a NULL temporary storage ptr, and
subsequent sort again tried to estimate the needed size, rather than
perform an actual sort.
this resolves issue #67
- - - - -
08bc0719 by Dari Kimanius at 2015-10-14T17:59:57+02:00
Added warnings for cases where orientation nr is not evenly divisible by 16 and 4, for 3D and 2D respectivly.
- - - - -
f00ba01b by Dari Kimanius at 2015-10-14T18:15:54+02:00
Added support for orientation nr other than evenly divisible by 16 (3D) and 4 (2D) in calls to diff2-kernel.
- - - - -
58ad10b4 by Dari Kimanius at 2015-10-15T17:19:11+02:00
Merge branch 'experimentalDiff2Coarse'
- - - - -
98749205 by Dari Kimanius at 2015-10-16T17:59:45+02:00
Removed an unnecessary stream synchronization after BP and added stream management to cub-utils.
- - - - -
60d79ce0 by Bjoern Forsberg at 2015-10-19T12:01:35+02:00
Merge branch 'sjorsmerge'
- - - - -
d43cf71d by Dari Kimanius at 2015-10-19T15:11:10+02:00
Moved the mapping of AllWeights to Mweights to the gpu, hence removing back and forth copy.
Yielding 5-7% speed-up in 3D classification with one class.
- - - - -
02c1609d by Dari Kimanius at 2015-10-19T15:57:12+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-1.4
Conflicts:
src/gpu_utils/cuda_ml_optimiser.cu
- - - - -
76274f98 by Dari Kimanius at 2015-10-20T14:05:02+02:00
Fixed issue 71, caused by translation data not being sent to the coarseDiff2_CC-kernel.
Clean-up in convertAllSquaredDifferencesToWeights.
- - - - -
8976ffa6 by Sjors Scheres at 2015-10-21T11:13:58+01:00
added --average, flipZ and ampl_corr to image handler
- - - - -
a17ad36c by Bjoern Forsberg at 2015-10-21T16:58:40+02:00
device ids are set consistently mpi/non-mpi
- - - - -
93aa3b21 by Bjoern Forsberg at 2015-10-22T13:40:22+02:00
mapAllWeightsToMweights uses class-streams
also passes streams, rahter than cudaMLO, to helper functions.
- - - - -
7ccfdaf0 by Dari Kimanius at 2015-10-22T13:45:13+02:00
Fixed two valgrind warnings cased by cuda stream creates writing outside the data size of cudaStream_t.
- - - - -
fd013c93 by Dari Kimanius at 2015-10-22T14:30:05+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
Conflicts:
src/gpu_utils/cuda_helper_functions.cu
- - - - -
75127a2d by Dari Kimanius at 2015-10-22T17:10:32+02:00
Fix for "DEBUG_ERROR: Non-zero allocation count encountered in custom allocator between iterations."
Caused by backprojector data bundles remaining in custom allocator.
- - - - -
92894265 by Dari Kimanius at 2015-10-22T17:38:57+02:00
Fixed "DEBUG_WARNING: Device double allocation."
Caused by double allocation of object in projection-plan-class.
- - - - -
eda3042b by Sjors Scheres at 2015-10-22T17:04:23+01:00
second attempt to commit changes in GUI for wrapper to Gctf
- - - - -
16458363 by Sjors Scheres at 2015-10-22T17:06:51+01:00
second attempt to commit changes in temp hack to allow using 2 gpus through mpi
- - - - -
3d0bb2b8 by Sjors Scheres at 2015-10-22T17:09:10+01:00
second attempt to commit changes to get rid of annoying print statements that keep coming back
- - - - -
24aad353 by Bjoern Forsberg at 2015-10-23T13:56:10+02:00
wavg is class-streamed
- - - - -
43b46cc5 by Dari Kimanius at 2015-10-23T17:23:16+02:00
Rearranged data flow in convertPostKernel during coarse pass.
All calculations are done on gpu, yeiling a factor 7 speed-up of convertAllSquaredDifferencesToWeights in the coarse pass.
- - - - -
e56c1b1c by Bjoern Forsberg at 2015-10-26T09:01:24+01:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
6d7fcc54 by Dari Kimanius at 2015-10-26T15:44:13+01:00
Summation of the weights are moved from the sumweight-kernel to instead use the last value in the cumulative sum of significant weights.
- - - - -
c556d853 by Dari Kimanius at 2015-10-26T15:46:22+01:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
0eed68b7 by Dari Kimanius at 2015-10-26T16:14:21+01:00
Clean-up in generate-projection-setup. Prior look-up is embedded in Mcoarse_significant and not needed to be fetched again.
- - - - -
24e33e77 by Bjoern Forsberg at 2015-10-26T16:21:57+01:00
timing fix
- - - - -
b6c22af2 by Bjoern Forsberg at 2015-10-26T16:22:41+01:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
d83ea8f8 by Bjoern Forsberg at 2015-10-27T10:35:02+01:00
fixes to debug warnings
(was checking against old stuff)
- - - - -
2f6265ec by Dari Kimanius at 2015-10-27T10:56:28+01:00
Also convert-fine uses the cumulative-sum-approach to fined the significant weight and sum of all weights.
- - - - -
c0afe4d0 by Sjors Scheres at 2015-10-27T10:51:56+00:00
changed warnings about norm corrections and made initialization of GPUs sequential on MPI ranks
- - - - -
2abb3f00 by Bjoern Forsberg at 2015-10-27T13:00:30+01:00
2 passes on GPU to avoid --oversampling 0 issues
This due to the difference in datastructure between coarse and fine pass
in the GPU-code. This MAY clash with some local-searches or similar. Be
on the lookout.
fixes issue #73
- - - - -
6067d7f1 by Dari Kimanius at 2015-10-27T15:37:51+01:00
Fixed the value of the metadata field my_nr_significant_coarse_samples.
Also fixed debug and double precision build issues.
- - - - -
c9ec3451 by Dari Kimanius at 2015-10-27T15:39:21+01:00
Replaced cub with thurst in the filtering function.
Also setup the custom allocator to be used by thrust. Hence no cudaMallos or cudaFrees are called by thrust and memory allocation on top of the reserved amount is not required.
- - - - -
8197b070 by Dari Kimanius at 2015-10-27T15:48:39+01:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
Conflicts:
src/gpu_utils/cuda_ml_optimiser.cu
- - - - -
b54bc28d by Dari Kimanius at 2015-10-27T19:13:59+01:00
Fixed valgrind error in storeWeightedSum.
- - - - -
6d94012e by Sjors Scheres at 2015-10-28T09:15:25+00:00
Merged in tcblab/relion-devel-tcblab (pull request #1)
Master
- - - - -
e028c28b by Sjors Scheres at 2015-10-28T11:42:38+00:00
implemented option to apply solvent-mask correction every iteration of auto-refine
- - - - -
b1963a2e by Sjors Scheres at 2015-10-28T11:48:40+00:00
trying to resolve conflicts in merge with relion-devel
- - - - -
16a71a2c by Sjors Scheres at 2015-10-28T11:52:54+00:00
trying to resolve conflicts in merge with relion-devel, part 2
- - - - -
bd8517b5 by Sjors Scheres at 2015-10-28T11:53:14+00:00
Merged tcblab/relion-devel into master
- - - - -
2e9fec64 by Sjors Scheres at 2015-10-28T12:03:41+00:00
put back my changes in ml_optimiser_mpi.cpp and removal of persistent printf statements in src/gpu_utils/cuda_ml_optimiser.cu
- - - - -
54ba25a6 by Sjors Scheres at 2015-10-28T12:07:43+00:00
Merged in scheres/relion-devel-lmb (pull request #2)
Master
- - - - -
15fd471e by Bjoern Forsberg at 2015-10-28T13:14:13+01:00
Merged tcblab/relion-devel into master
- - - - -
ee95b446 by Sjors Scheres at 2015-10-28T12:17:28+00:00
put back ontheflyshifts=true in initialise of mpi
- - - - -
4394ceab by Dari Kimanius at 2015-10-28T15:08:57+01:00
Fixed an error caused by omitted handling of the excessive blocks created by the obsolete block-split-function.
- - - - -
79a7132c by Sjors Scheres at 2015-10-28T15:37:36+00:00
Merged in tcblab/relion-devel-tcblab (pull request #4)
Bug fix
- - - - -
5bd32c8e by Sjors Scheres at 2015-10-28T15:37:48+00:00
Merged tcblab/relion-devel into master
- - - - -
d516a1f7 by Bjoern Forsberg at 2015-10-28T17:52:59+01:00
Merged tcblab/relion-devel into master
- - - - -
ba3aed03 by Bjoern Forsberg at 2015-10-30T17:08:32+01:00
Fixing a bug when there is a single significant weight
Also some added debug-verbosity by adding cudaPeekAtLastError()
under DEBUG_CUDA precompiler def, which should catch some more
errors closer to their true origin.
- - - - -
02feb2ea by Dari Kimanius at 2015-10-30T17:37:25+01:00
Huge revison on the custom allocator, where allocation can be marked with a cuda event which have to expire before deallocation.
Backproject data bundle is no longer required.
Out-of-memory-handlers are no longer required.
- - - - -
c5cc9a78 by Dari Kimanius at 2015-10-30T17:45:42+01:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
de4f43a0 by Bjoern Forsberg at 2015-10-30T18:03:12+01:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
586f54c2 by Dari Kimanius at 2015-11-05T09:48:39+01:00
Removed a memory related debug print out.
- - - - -
c28e0d5b by Dari Kimanius at 2015-11-05T09:49:28+01:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
cbafca7b by Sjors Scheres at 2015-11-10T14:28:46+00:00
add preread_images to cuda code
- - - - -
cdd2b8a6 by Sjors Scheres at 2015-11-12T14:49:16+00:00
implemented new sigma_rot for dna-origami+p53 refinement
- - - - -
4a9c59fe by Sjors Scheres at 2015-11-12T17:00:23+00:00
big merge with helical and multibody processing
- - - - -
550b6de3 by Bjoern Forsberg at 2015-11-17T18:25:38+01:00
Forcing some intrinsics in diff2
also cleaning 2D/3D projection code, which seems to have
minor improvements (2D diffcoarse ~7%, 3D diffcoarse ~1%)
due to reduced thread divergence.
- - - - -
0c1da926 by Bjoern Forsberg at 2015-11-18T15:42:27+01:00
cmake cleanup and warns about double when single
- - - - -
1c77efea by Bjoern Forsberg at 2015-11-18T18:51:59+01:00
Fixed compilation errors
Also a double-prec computation in single mode in translator.
- - - - -
41890eb7 by Bjoern Forsberg at 2015-11-19T12:26:37+01:00
MlDeviceBundle introduced but not taken effect
- - - - -
fe764181 by Bjoern Forsberg at 2015-11-19T19:07:02+01:00
Draft devicebundle which seems to work.
Some ugly stuff, like cuda_ml_optimiser.cu:653-656
- - - - -
e300f20c by Bjoern Forsberg at 2015-11-20T14:09:02+01:00
Simplifying without changing
- - - - -
fde6a604 by Sjors Scheres at 2015-11-20T16:49:08+00:00
first moves towards a pipeline-version of the GUI
- - - - -
533c1957 by Bjoern Forsberg at 2015-11-24T11:21:44+01:00
pool-flag --pool
does NOT modify block_size of pooled particles
see issue #74
- - - - -
d20778eb by Dari Kimanius at 2015-11-24T11:38:53+01:00
Added RelWithDebInfo as a valid build type.
Currently CUDA_DEBUG is activated when this is set, the performance impact of this is however not evaluated.
- - - - -
ee0f4913 by Bjoern Forsberg at 2015-11-25T18:19:23+01:00
single/single prec on CPU/GPU is default
Some changes to names of cmake vars to clarify precision choices.
Also some changes to how the fouriertransform plans are reused
on the GPU.
- - - - -
2013caae by Bjoern Forsberg at 2015-11-26T10:39:04+01:00
More efficient fftw-array allocation
- - - - -
91470a18 by Bjoern Forsberg at 2015-11-27T12:00:59+01:00
Include single+double prec fftw libs default
This so that we can choose prec after cmake invocation.
- - - - -
963abe6b by Bjoern Forsberg at 2015-11-27T15:00:00+01:00
Choose shared or static libs in cmake build proc.
- - - - -
25e3466c by Sjors Scheres at 2015-11-27T17:24:15+00:00
progressing with pipeline, up until process deletion
- - - - -
975f8a20 by Bjoern Forsberg at 2015-11-30T15:25:23+01:00
fixes to relion_mpi transformer clearing
also fftw cmake stuff refined slightly
- - - - -
b46c77ea by Bjoern Forsberg at 2015-11-30T16:30:33+01:00
cmake fix to cuda version check
- - - - -
34a9f40e by Dari Kimanius at 2015-11-30T17:02:11+01:00
fftw package finder checks the require flag, required by the cmake standard.
Also if one of the paths are not found fftw_found is not set to true.
Only the required library are included in the lib-path instead of all three (single, double and thread). This is better in case one does not have all of them installed. Besides relion doesn't uses the thread enabled version anymore.
- - - - -
49b15958 by Dari Kimanius at 2015-11-30T17:05:36+01:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
d099446a by Dari Kimanius at 2015-11-30T18:18:48+01:00
Build type is checked and only valid values are accepted.
Release is the default value.
Some minor clean-ups.
- - - - -
fb81f38d by Bjoern Forsberg at 2015-12-01T15:20:43+01:00
fftw is external project which fetches fftw-3.3.4
Looks for system libs, but builds local fftw-3.3.4 libs and headers
which are automatically included in the build, if system libs are
not found.
If built in the local directory, any new build will find these and
refrain from rebuilding the loacla fftw.
The proeccess is able to modify the CPU-precision through ccmake, although
new libraries will be built if they do not allready exist in the system or
local directories.
- - - - -
7ac67f67 by Bjoern Forsberg at 2015-12-01T15:42:03+01:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
Included in merge is the fetaure to specify -DOWN_FFT=ON/OFF on cmake cmd-line
Conflicts:
CMakeLists.txt
cmake/FindFFTW.cmake
- - - - -
886e7a03 by Bjoern Forsberg at 2015-12-02T13:58:45+01:00
fltk is external project which fetches fltk-1.3.3
also, fltk and fftw are only build as local libs if the user
specifies -DOWN_FFTW=ON and -DOWN_FLTK=ON in the cmake cmd-line.
If no system libs are found while these flags are NOT passed on
the cmd-line, warnings with suggestions are displayed.
- - - - -
30769848 by Bjoern Forsberg at 2015-12-02T17:21:37+01:00
Default build type set to release
Also further fixes to the cmake procedure
- - - - -
97f2e84e by Sjors Scheres at 2015-12-03T10:22:56+00:00
now can add scheduled jobs and execute them without loosing track of history
- - - - -
05195096 by Bjoern Forsberg at 2015-12-03T11:50:45+01:00
nvcc avoids float128 by ugly hack
gcc supports a type float128, which the wrapper nvcc does not support.
fftw has enabled support for this under the gcc compiler, but not for
the intel compiler. So in order to compile fftw using nvcc as a gcc-
wrapper, we can fool nvcc to think it is an intel compiler during fftw
compilation/inclusion.
The better solution would be for
a) cuda to enable float128
b) fftw to sense nvcc, rather than intel
Option bi) is possible through minor modification of the fftw3.h-header,
as described here:
https://github.com/FFTW/fftw3/issues/18
This has for some reason not made it into the past or current release
of ftw, however, so option a) is forced to provide compatibility or
RELION with past and present fftw-verisons without modification. This
is achieved by adding an nvcc-flag which provides an intel compiler
identifier. ( se ../cmake/BuildTypes.cmake)
- - - - -
def34930 by Dari Kimanius at 2015-12-03T13:26:22+01:00
Added some error reporting in coarse-convert to catch invalid cases of zero number of weights.
- - - - -
de04ebbe by Dari Kimanius at 2015-12-03T13:27:59+01:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
1905ec02 by Bjoern Forsberg at 2015-12-03T17:45:23+01:00
Fixes to cmake fftw-search in sys-libs
- - - - -
4eebaf69 by Sjors Scheres at 2015-12-03T17:28:33+00:00
added import and motioncorr jobwindows, gctf wrapper in ctffund_runner is pending
- - - - -
8136a3a0 by Sjors Scheres at 2015-12-04T09:30:30+00:00
Merged in tcblab/relion-devel-tcblab (pull request #5)
Mainly build environment updates and memory optimizations
- - - - -
10d86820 by Sjors Scheres at 2015-12-04T09:50:15+00:00
trying to resolve conflicts with merge from Dari and Bjoern
- - - - -
ab070a8f by Sjors Scheres at 2015-12-04T09:50:37+00:00
Merged tcblab/relion-devel into master
- - - - -
2b5ffdd7 by Sjors Scheres at 2015-12-04T09:58:51+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
9eb5224d by Bjoern Forsberg at 2015-12-04T13:29:59+01:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
27e54550 by Bjoern Forsberg at 2015-12-04T13:36:11+01:00
Merged tcblab/relion-devel into master
- - - - -
510b7336 by Dari Kimanius at 2015-12-04T14:01:53+01:00
Fixed compilation error and added external to gitignore
- - - - -
ceb012b2 by Dari Kimanius at 2015-12-04T14:02:05+01:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
12055f48 by Dari Kimanius at 2015-12-04T15:36:18+01:00
Script files are copied into the bin-dir at build-time.
- - - - -
953d98cc by Bjoern Forsberg at 2015-12-04T15:45:41+01:00
Fixed float overflow compile-time warning
- - - - -
134e9e5b by Bjoern Forsberg at 2015-12-04T15:46:29+01:00
Preparation (non-mpi compliant) gpu-autopicker
- - - - -
6320e588 by Bjoern Forsberg at 2015-12-04T15:47:32+01:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
fdf1eafd by Dari Kimanius at 2015-12-04T17:36:45+01:00
Intall target setup in cmake.
Rpath activated for the shared-libs.
Renamed shared lib gpu_util.so to relion_gpu_util.so
Scripts are not installed yet.
- - - - -
e24a670c by Dari Kimanius at 2015-12-04T17:39:06+01:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
94d6df57 by Sjors Scheres at 2015-12-04T17:21:49+00:00
now gctf wrapper works
- - - - -
ad00c84d by Sjors Scheres at 2015-12-04T17:22:09+00:00
small debugs of update from Bjoern and dari
- - - - -
d7d5cc45 by Bjoern Forsberg at 2015-12-07T14:34:32+01:00
Restructure helper_functions + autopick stuff
- - - - -
f260dc79 by Dari Kimanius at 2015-12-07T15:13:42+01:00
cmake-install now works for static builds.
Install targers of the shared libray is omitted in this case.
resolves issue #75
- - - - -
d5978d71 by Dari Kimanius at 2015-12-07T15:14:25+01:00
Added cmake install targets for the script files.
- - - - -
44d89546 by Bjoern Forsberg at 2015-12-07T16:13:50+01:00
RelWithDebINfo fix and autopicking stuff
- - - - -
e4f6132e by Bjoern Forsberg at 2015-12-07T16:15:20+01:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
Conflicts:
src/apps/CMakeLists.txt
- - - - -
72d16f51 by Bjoern Forsberg at 2015-12-08T10:20:52+01:00
Autopick centerFFT w/ mulitple blocks
reduced precompiler debug output in gpu code for dev legibility
- - - - -
b33085b4 by Bjoern Forsberg at 2015-12-08T13:34:10+01:00
ProbRatio kernel works
- - - - -
73987551 by Dari Kimanius at 2015-12-08T15:17:08+01:00
Fixed Issue #77 by adding some booleans to account for the status of the fft plans.
- - - - -
8df19c8b by Dari Kimanius at 2015-12-08T15:18:25+01:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
40f3ae3c by Dari Kimanius at 2015-12-08T18:55:51+01:00
Fixed issue #76. Changed default behavior to what is requested in the issue.
- - - - -
52e4fa7f by Dari Kimanius at 2015-12-08T19:09:59+01:00
Fixed issue with cached library-paths for fftw.
Fixed linking between external libraries and internal shared libraries.
- - - - -
a8fc09ca by Bjoern Forsberg at 2015-12-09T10:09:00+01:00
AutoP-texture comp. and run w/ incorrect results
- - - - -
576d2bec by Bjoern Forsberg at 2015-12-09T11:16:43+01:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
e1c7445f by Bjoern Forsberg at 2015-12-09T15:19:15+01:00
Working texture autopicker
- - - - -
18fd7ee7 by Dari Kimanius at 2015-12-09T15:29:23+01:00
A bit better error output.
- - - - -
a3d1bded by Dari Kimanius at 2015-12-09T15:55:32+01:00
Fixed CPU double precision compilation issue caused by unhandled type missmatch.
- - - - -
5c4c788f by Dari Kimanius at 2015-12-09T15:56:49+01:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
e00032b2 by Bjoern Forsberg at 2015-12-09T16:54:16+01:00
autopick convolution is also on GPU
- - - - -
13608132 by Bjoern Forsberg at 2015-12-09T16:54:45+01:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
6dba8449 by Dari Kimanius at 2015-12-09T17:20:48+01:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
c3c739eb by Sjors Scheres at 2015-12-09T17:22:51+00:00
now manual pick, extract and sort work inside pipeline
- - - - -
926d973c by Bjoern Forsberg at 2015-12-10T09:55:35+01:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
17cf55e8 by Bjoern Forsberg at 2015-12-10T12:03:03+01:00
float2-arrays are used for complex numbers
This is to maintain the interleaved real & complex components in prep
for device-bound fourier transforms. I did NOT check performance.
- - - - -
73e41da4 by Dari Kimanius at 2015-12-10T12:15:14+01:00
Allocator is made thread-safe and shared between threads.
Device allocations made by CudaGlobalPtr are not freed in the destructor but marked by a cuda event record to be freed later.
- - - - -
19e309e5 by Dari Kimanius at 2015-12-10T12:19:35+01:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
da7e4696 by Dari Kimanius at 2015-12-10T12:25:13+01:00
Changed the name of the gpu-memory-flag to better suite its new purpose.
- - - - -
e6ec2635 by Dari Kimanius at 2015-12-10T18:22:55+01:00
Cleaned-up some (non-significant) errors.
- - - - -
61fefa86 by Dari Kimanius at 2015-12-10T18:23:15+01:00
Added the CudaFFT class that handles plans and pointers for R2C (only) transforms.
Ported the largest cpu transform to gpu in autopicker.
- - - - -
b367e1f6 by Bjoern Forsberg at 2015-12-10T18:38:39+01:00
Working complex texture
Tested only sparsely for autopicker (only 2D-textures).
- - - - -
23b6a8c1 by Bjoern Forsberg at 2015-12-10T18:42:04+01:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
Conflicts:
src/gpu_utils/cuda_helper_functions.cuh
- - - - -
a09d7550 by Bjoern Forsberg at 2015-12-11T15:16:32+01:00
Enabling a single complex texture is now possible
There is some doubt as to the benfit, however. One would expect that
3D projection kernels would benefit through decreased artimetic count
during texture-intrinsic interpolation. This is to some extent observed,
but the entire beenfit does not pervail due to a slowdown in other parts
of the code, for reasons unknown.
The projection-instructions are maintained such that a call separates
real and imaginary components. If (as in autopicker) the desired out-
put is a complex array, then these components are either joined after
projection or written separately.
- - - - -
ca186d73 by Dari Kimanius at 2015-12-11T16:55:03+01:00
Minor build related changes.
- - - - -
0e5177f7 by Dari Kimanius at 2015-12-11T16:55:23+01:00
windowFourierTransform is now implemented on the GPU both for intertwined complex data and separate real and imag.
It is inserted into autopicker which primarily removes some back and forth memory copies.
- - - - -
de7687ce by Dari Kimanius at 2015-12-11T17:02:27+01:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
6e8712cd by Bjoern Forsberg at 2015-12-14T09:29:58+01:00
Some autopicker memcpys rm:ed
- - - - -
743b69fe by Dari Kimanius at 2015-12-14T14:16:50+01:00
Disabled functions to fix broken standared build.
- - - - -
800fb774 by Sjors Scheres at 2015-12-14T16:40:39+00:00
Merged in tcblab/relion-devel-tcblab (pull request #6)
New autopicker code (not activated), bug-fixes, and better memory management
- - - - -
ada60a2c by Sjors Scheres at 2015-12-15T14:37:32+00:00
now pipeline with io_node_display working
- - - - -
09e1cee6 by Dari Kimanius at 2015-12-16T11:59:11+01:00
Better error reporting.
- - - - -
d8156085 by Sjors Scheres at 2015-12-16T11:24:34+00:00
trying to resolve conflict with Dari and Bjorn
- - - - -
487d0012 by Sjors Scheres at 2015-12-16T11:25:48+00:00
trying to resolve conflict with Dari and Bjorn
- - - - -
d8b3c3d3 by Sjors Scheres at 2015-12-16T11:28:17+00:00
trying to resolve conflict with Dari and Bjorn
- - - - -
7b19fd61 by Sjors Scheres at 2015-12-16T11:29:52+00:00
trying to resolve conflict with Dari and Bjorn
- - - - -
8d27a84b by Sjors Scheres at 2015-12-16T11:30:08+00:00
Merged tcblab/relion-devel into master
- - - - -
ce0d5310 by Dari Kimanius at 2015-12-16T12:33:25+01:00
Merged tcblab/relion-devel into master
- - - - -
5c14f39d by Dari Kimanius at 2015-12-17T14:26:17+01:00
Only devices that are used by the optimiser is allocated upon not all the devices like before.
- - - - -
5beb1a3c by Bjoern Forsberg at 2015-12-17T18:32:05+01:00
Most autopicker memcpys rm:ed
also some kernel-related fixes
- - - - -
1066c644 by Bjoern Forsberg at 2015-12-17T18:49:00+01:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
Conflicts:
src/gpu_utils/cuda_kernels/helper.cu
- - - - -
22543ae3 by Bjoern Forsberg at 2015-12-18T10:24:39+01:00
cuda_autopicker wors for SS and DS prec. combos
- - - - -
518af0fc by Bjoern Forsberg at 2015-12-18T10:35:56+01:00
autopicker works for all prec. combos - no memcpys
- - - - -
8bdd537f by Bjoern Forsberg at 2015-12-18T11:48:26+01:00
Minor improvement to ProbRatio-kernel
- - - - -
58a7e8d5 by Bjoern Forsberg at 2015-12-18T13:55:57+01:00
buffer centerFFT instead of D2D-memcpy
- - - - -
c0819b4e by Bjoern Forsberg at 2015-12-18T14:30:39+01:00
cuda-own calculateStddevAndMeanUnderMask
- - - - -
a19bc2e6 by Dari Kimanius at 2015-12-18T15:09:08+01:00
Multiple fixes:
1) Improved numerical accuracy of offset prior by uniformly shifting values towards unity with a factor that is remove in the normalization.
2) Fixed concurrency issue due two threads simultaneously setting shared projector plan in auto refine.
3) Moved projector plan into coarse-diff function, since it's only used there.
4) Removed weird conditional raising signal.
- - - - -
9617e46b by Dari Kimanius at 2015-12-18T15:16:28+01:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
07dc26b0 by Dari Kimanius at 2015-12-21T12:36:11+01:00
computeStats is calculated in double-precision regradless of precompilation settings.
- - - - -
b860007c by Bjoern Forsberg at 2015-12-21T14:27:05+01:00
sutopick pre-calc largely on gpu
- - - - -
2b287a05 by Bjoern Forsberg at 2015-12-21T14:27:49+01:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
943ad357 by Dari Kimanius at 2015-12-21T14:55:33+01:00
Added stream-synchronizations after all device to host copies where data is used on the CPU immediately after the copy.
- - - - -
144c1a03 by Bjoern Forsberg at 2015-12-21T17:06:11+01:00
Preparation batch execution (autopicker)
- - - - -
46e6e1ef by Bjoern Forsberg at 2015-12-21T17:08:31+01:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
Conflicts:
src/gpu_utils/cuda_autopicker.cu
- - - - -
2f683fc5 by Bjoern Forsberg at 2015-12-21T17:35:21+01:00
Autopick makes batched projection call
- - - - -
f73bc4ec by Bjoern Forsberg at 2015-12-21T18:05:23+01:00
Autopicker uses batch-plans of size 1 (prep)
- - - - -
8897996c by Bjoern Forsberg at 2015-12-21T18:36:16+01:00
Prep for fourier-batches
separated first-psi and remaining transformers
prepped runCenterFFT for multiple image input
- - - - -
535991d9 by Bjoern Forsberg at 2015-12-22T12:45:16+01:00
Prep for fourier-batches
separated first-psi and remaining transformers
prepped runCenterFFT for multiple image input
- - - - -
f4e5994b by Bjoern Forsberg at 2015-12-22T12:54:30+01:00
First-psi calcs moved out of psi-loop
- - - - -
1b913a19 by Bjoern Forsberg at 2015-12-22T12:59:14+01:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
Conflicts:
src/gpu_utils/cuda_autopicker.cu
- - - - -
ea24eb49 by Bjoern Forsberg at 2015-12-22T17:06:35+01:00
Batched psis in autopicking, but w/ incorr results
This commit should not be ussed, it is a waypoint.
- - - - -
f756baa6 by Bjoern Forsberg at 2015-12-28T15:54:20+01:00
Fix to WindowFourierTransform2
Also best_psi image is written correctly
Also compatibility added for cuda-Double prec. in planMany-stage
- - - - -
0bed3b4b by Dari Kimanius at 2015-12-28T16:19:16+01:00
Clean-up and standardization of cuda-autopicker.
- - - - -
979feb4f by Dari Kimanius at 2015-12-28T18:55:09+01:00
Ported micrograph preparation to GPU.
- - - - -
cf67089e by Dari Kimanius at 2015-12-29T14:17:58+01:00
Removed back and forth memory copys surrounding calculateStddevAndMeanUnderMask2.
- - - - -
cae7db47 by Bjoern Forsberg at 2015-12-29T17:07:47+01:00
CPU-based code can shrink micrograph
the real value of this is evident only in the GPU-code, however it is easier
to implement on the CPU and safer to test there as well.
- - - - -
2d3a6dc6 by Bjoern Forsberg at 2015-12-30T14:57:00+01:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
032d536a by Bjoern Forsberg at 2015-12-30T15:32:40+01:00
The output of CPU-autopicker is micrograph size
(even if a reduced micrograph is used for calcs)
- - - - -
26733213 by Bjoern Forsberg at 2015-12-30T18:46:32+01:00
GPU also shrinks the micrograph during calcs
NOTE:
This may be due to the difference in FFTW and CUFFT, but it may also
be the fact that the GPU has limited ammounts of memory. I think it
is the combinations them, but anyway:
Some micrograph sizes cause issues, which appear to be due to the
accidnetal or occational occurence of large prime factors in the
fourier-representation image sizes. LArge primes means alot of extra
(temp) storage during FFT calcs.
Why was this not an issue on the CPU?
a) CPU is serial -> onlye ever 1 FFT at a time
b) FFTW != CUFFT, algorithmically
c) y=1950 => x=976, so largest prime is 61. Big, but ok.
For, instance, trying y=1940 on the GPU with 5deg orient-step:
360/5 = 72 FFTs at once (in memory at once)
y=1940 => x=971 (which is prime), so largest prime is 971 (!)
So even though 1940 < 1950, the estimated space needed to perform
the FFT is more than 8 times bigger. combine that with performing 72
FFT at once, and the space requirement is a whopping 576 times larger.
This is not so much a problem as a hurdle, which is solved by two
considerations/modifications:
1 - Reduce 1950 much beyond 1940. The smaller the number, the smaller
the possibility of large prime, in addition to smaller objects and faster
calculations.
2 - Enforce numbers which doesn't contain large prime factors. GROMACS
allready does this, but a table lookup is not that complicated if we feel
like doing our own. The formal requirement is to force a basic dimension
x such that
max(largest_prime_factor(x),largest_prime_factor((x-1)*2)) < M
where M is a cutoff which we specify as some ok-sized prime, like 61 or so.
FOR NOW, THIS IS AN ISSUE WITH THE GPU-CODE WHICH NEEDS TO BE FIXED
- - - - -
c61904b6 by Bjoern Forsberg at 2016-01-04T11:28:11+01:00
Micrograph size during calcs uses --shrink flag
default value for --shrink is 1.0
--shrink 0<val<=1 results in a size = val*mic_size
--shrink val>1 results in a size = val
size is allways adjusted down to nearest int which is 0 mod 2,
since both size and (size/2+1) should be int.
size < downsize_mic gives a non-fatal warning, since this limits
resolution beyond the (requested or default) resolution.
TODO:
Mpsi array should also be up-sized when calcs are finished.
Large prime factors in size should be avoided by allowing dynamical
adjustment of size according to allocation estiamtes, see
cufftEstimateMany()
- - - - -
f8bdbd73 by Bjoern Forsberg at 2016-01-04T11:48:35+01:00
Psi-image is also micrograph-sized in output
- - - - -
8ecba9e7 by Sjors Scheres at 2016-01-04T13:49:25+00:00
now with polsihing, maskcreate and resmap pipeline working
- - - - -
caf5d350 by Sjors Scheres at 2016-01-04T15:09:43+00:00
nicer lokking buttons on maingui
- - - - -
bc7b3bf8 by Bjoern Forsberg at 2016-01-05T11:23:19+01:00
Rearrangements & speedups for autopicker
also fix to warning-condition regarding --shrink option
- - - - -
15817836 by Sjors Scheres at 2016-01-05T14:12:33+00:00
now picking with refined coordinates and even nicer-looking GUI
- - - - -
539def18 by Bjoern Forsberg at 2016-01-06T10:42:50+01:00
finalize autopicker features for PullReq
- - - - -
140656e4 by Sjors Scheres at 2016-01-06T14:35:22+00:00
with gctf bfactors etc, alias of jobs, and regrouping/centering in ParticleSelect
- - - - -
2b02a620 by Dari Kimanius at 2016-01-06T16:43:49+01:00
PDF-offset is normalized with the mean value rather than the the minimum value.
This is done with the same factor for all classes, which is theoretically more correct.
- - - - -
f65b0b08 by Dari Kimanius at 2016-01-06T16:53:01+01:00
PDF-orientation is also normalized with the mean over all classes to further improve numerical accuracy.
- - - - -
c47c2584 by Sjors Scheres at 2016-01-07T09:08:26+00:00
some debugging, also aliases in GUI
- - - - -
e656c1a1 by Sjors Scheres at 2016-01-07T09:17:44+00:00
try to add cuda autopicker and merge new-GUI code with bitbucker from Old Year
- - - - -
3bdf474f by Sjors Scheres at 2016-01-07T09:20:36+00:00
get rid of persistent test.cpp
- - - - -
ceab0e23 by Sjors Scheres at 2016-01-07T09:27:09+00:00
for now no background image in GUI, and repaired merge error in autopicker.cpp
- - - - -
3aa91d02 by Sjors Scheres at 2016-01-07T09:39:57+00:00
changed order of particle selection in left-hand GUI menu
- - - - -
5fdb28bd by Sjors Scheres at 2016-01-07T09:48:50+00:00
called subtract job Particle subtraction
- - - - -
1b6e8af8 by Sjors Scheres at 2016-01-07T11:35:13+00:00
already add join star fiel process (not funciotnal yet, but to prevent errors in pipeline files
- - - - -
35c3046b by Dari Kimanius at 2016-01-07T13:15:02+01:00
Merge https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
aa44856a by Sjors Scheres at 2016-01-07T13:09:54+00:00
repaired bug in import job
- - - - -
12d65b51 by Dari Kimanius at 2016-01-07T14:20:46+01:00
Fixed some suboptimal device synchronizations.
- - - - -
d0f1930c by Sjors Scheres at 2016-01-07T19:51:46+00:00
Merged in tcblab/relion-devel-tcblab (pull request #7)
Improved accuracy in refine and faster autopicker.
- - - - -
85bce6e5 by Dari Kimanius at 2016-01-08T17:01:31+01:00
Fixed debug-build false errors.
- - - - -
c57edd02 by Dari Kimanius at 2016-01-08T17:01:54+01:00
Inserted new code from cpu-autopicker for helical-segments; this code is commented out.
Also changed the metadata file output path and field order to match that of the cpu code.
- - - - -
63006e76 by Sjors Scheres at 2016-01-08T17:20:01+00:00
getting many bugs out, and also make aliases for node browsers
- - - - -
235f8a44 by Dari Kimanius at 2016-01-08T20:04:43+01:00
Fixed peculiar error with incorrect ccf-best for uneven ref-indices in gpu-autopick.
- - - - -
9a9846cf by Bjoern Forsberg at 2016-01-11T10:02:32+01:00
Fixed some compilation warnings
- - - - -
88e5288b by Sjors Scheres at 2016-01-11T14:23:39+00:00
dont have model.star as outputNode for auto3D runs
- - - - -
573aee1b by Sjors Scheres at 2016-01-11T15:41:15+00:00
removed some commented old code
- - - - -
1739e935 by Sjors Scheres at 2016-01-11T15:42:18+00:00
remove _movie from input filename if it already contains it
- - - - -
c15636aa by Sjors Scheres at 2016-01-11T15:48:53+00:00
repaired bug making output directory extract_mpi
- - - - -
d0dff4d1 by Sjors Scheres at 2016-01-11T17:09:04+00:00
trying to make movie-rpocessing work
- - - - -
8bb0d473 by Sjors Scheres at 2016-01-11T17:23:32+00:00
program to join STAR files
- - - - -
1be92a08 by Sjors Scheres at 2016-01-11T17:28:03+00:00
removed cerr statement
- - - - -
b1716538 by Sjors Scheres at 2016-01-11T18:37:49+00:00
removed cerr statement
- - - - -
a27c2718 by Sjors Scheres at 2016-01-12T09:34:00+00:00
allow to invert micrograph selection
- - - - -
80eadfa3 by Sjors Scheres at 2016-01-12T15:10:59+00:00
now with display of stdout and stderr on the GUI
- - - - -
e669bd61 by Sjors Scheres at 2016-01-12T16:10:50+00:00
merge with Shaodas stable helical processing
- - - - -
830949bd by Sjors Scheres at 2016-01-12T16:31:54+00:00
repaired bug in building of local FLTK
- - - - -
58867f4c by Sjors Scheres at 2016-01-12T17:25:02+00:00
adapted some continuation settings
- - - - -
a92bb3c5 by Bjoern Forsberg at 2016-01-13T13:32:12+01:00
Scale particles picked from shrunken mic
Given that the calculations are performed on a shrunken micrograph,
this commit realized the idea of perfroming particle picking on this
image, then scaling coordinates to the original mircograph size. This
rather than making the image bigger simply to perform particle picking
on it.
The question remains, however; Does this affect the picking quality?
In test, shrinking affects the picked results at ~55% of original
dimesions, i.e. ~30% of original image size. This sounds reasonable,
but the resolution limit for this run (lowpass=20), is as low as ~35%
of image dimensions, i.e. ~13% of image size.
It should be noted that shrinking the micrograph only very slightly,
can improve the picking (in my subjective assesment, by vis. inspect.)
- - - - -
8a61c078 by Sjors Scheres at 2016-01-13T17:13:33+00:00
repaired bug in moving output files away for continuation jobs in maingui
- - - - -
ac79bc5f by Sjors Scheres at 2016-01-14T10:18:44+00:00
further removal of minor bugs
- - - - -
9ace17cd by Sjors Scheres at 2016-01-14T15:47:20+00:00
replaced allmicnames by vector TODO: file i/o in GCTF
- - - - -
992a9aa9 by Sjors Scheres at 2016-01-14T17:06:28+00:00
further debugging ctffind runner with Gctf
- - - - -
39bc9ca9 by Dari Kimanius at 2016-01-14T18:14:13+01:00
Fixed some class related indexing errors in coarse-diff. And removed bad code.
- - - - -
db3e0b3e by Dari Kimanius at 2016-01-14T18:15:09+01:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
763d406c by Sjors Scheres at 2016-01-14T17:15:48+00:00
further debugging ctffind runner with Gctf
- - - - -
27ee6724 by Sjors Scheres at 2016-01-14T17:19:36+00:00
further debugging ctffind runner with Gctf
- - - - -
bfc76124 by Sjors Scheres at 2016-01-15T11:26:50+00:00
Merged in tcblab/relion-devel-tcblab (pull request #8)
Minor changes fixing error in refine and also some fixes in autopick
- - - - -
c3f38cfe by Sjors Scheres at 2016-01-15T12:02:57+00:00
report errors to stderr, not stdout
- - - - -
3ce940af by Sjors Scheres at 2016-01-15T12:03:20+00:00
run gctf on 20 mics at a time
- - - - -
e8e77888 by Sjors Scheres at 2016-01-15T12:03:52+00:00
repaired bug in motioncorr logfile placement
- - - - -
806d61f4 by Sjors Scheres at 2016-01-15T12:06:14+00:00
allow selection of micrographs
- - - - -
cba8208a by Sjors Scheres at 2016-01-15T12:13:07+00:00
repaired bug joinSTARs for empty STAR files
- - - - -
44108878 by Sjors Scheres at 2016-01-15T12:16:44+00:00
dont write out to stdout which ones already exist
- - - - -
dfadae1d by Sjors Scheres at 2016-01-15T14:33:08+00:00
added functionality to store notes
- - - - -
c1ae9022 by Sjors Scheres at 2016-01-15T14:37:22+00:00
reorganized some buttons
- - - - -
8c4424a6 by Sjors Scheres at 2016-01-15T17:30:27+00:00
repairing minor bugs
- - - - -
580a99e0 by Sjors Scheres at 2016-01-18T09:48:32+00:00
make backup copy of pipeline.star
- - - - -
69c8cb45 by Sjors Scheres at 2016-01-18T12:05:18+00:00
fuurther debugging of minor things in GUI, eg saving hidden jobfiles, better aliases and cleaning those
- - - - -
1d75d16c by Sjors Scheres at 2016-01-18T17:14:07+00:00
start of mpi support with multiple gpus
- - - - -
0db6f095 by Sjors Scheres at 2016-01-18T17:14:26+00:00
repaired bug in finding logfiles of Gctf and ctffind
- - - - -
3218a13e by Dari Kimanius at 2016-01-20T13:52:46+01:00
GPU accelerated first FFT transform and centerFFT in getFourierTransformsAndCtfs.
- - - - -
4f964034 by Dari Kimanius at 2016-01-20T13:55:23+01:00
GPU accelerated second transform and centerFFT in getFourierTransformsAndCtfs.
- - - - -
a11e5f82 by Dari Kimanius at 2016-01-21T13:32:30+01:00
Merged tcblab/relion-devel into master
- - - - -
5fa91dba by Sjors Scheres at 2016-01-21T12:57:40+00:00
yet another bug in ctfind runner
- - - - -
809086b7 by Sjors Scheres at 2016-01-21T12:58:31+00:00
change default outputnames refine jobs with scheduled jobs
- - - - -
60ba35dc by Sjors Scheres at 2016-01-21T12:59:27+00:00
schedule jobs now running in a first version
- - - - -
190d3c05 by Sjors Scheres at 2016-01-21T14:21:51+00:00
merge again with Shaodas helices
- - - - -
9ba87715 by Sjors Scheres at 2016-01-21T14:44:20+00:00
bug in continuation relion_refine
- - - - -
921a0434 by Sjors Scheres at 2016-01-21T14:50:45+00:00
ask for repeats and min_wait in running schedfuled jobs
- - - - -
a1f84de7 by Sjors Scheres at 2016-01-21T15:01:22+00:00
bug non-defined current_job
- - - - -
86ccda59 by Bjoern Forsberg at 2016-01-21T16:50:11+01:00
Autopicking batches ffts if neccessary (mem-limit)
Due to the scheme used, the batches are M in number, where the first
M-1 are of size X, and the last has size Y, where (X-M+1)<Y<=X. As long
as the number of batches remains relatively small compared to M, this
ensures good efficiency (no spilling)
The Autopicker CAN still run out of memory since this batching, so far,
only concerns the fft-associated arrays (and temp storage) which are
proportional in number to the orientations.
- - - - -
01d3a268 by Sjors Scheres at 2016-01-21T16:14:34+00:00
trying to resolve conflicts for merge with relion-devel
- - - - -
01672172 by Sjors Scheres at 2016-01-21T16:21:10+00:00
trying to resolve conflicts for merge with relion-devel
- - - - -
a2832a04 by Bjoern Forsberg at 2016-01-21T17:30:54+01:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
3b4e59f6 by Sjors Scheres at 2016-01-21T16:31:04+00:00
trying to resolve conflicts
- - - - -
c1094c45 by Sjors Scheres at 2016-01-21T16:31:36+00:00
Merged in scheres/relion-devel-lmb (pull request #3)
put back ontheflyshifts=true in initialise of mpi
- - - - -
3ffe14b2 by Sjors Scheres at 2016-01-21T16:33:20+00:00
repairing what was broken in merge
- - - - -
c1440d2b by Sjors Scheres at 2016-01-21T16:34:19+00:00
Merged in scheres/relion-devel-lmb (pull request #9)
repairing what was broken in merge (had to do some weird stuff in src/gui_jobwindow.cpp and src/gui_mainwindow.cpp because of silly merge conflicts)
- - - - -
f17b88e4 by Sjors Scheres at 2016-01-21T16:37:13+00:00
Merged tcblab/relion-devel into master
- - - - -
052fbcb4 by Sjors Scheres at 2016-01-21T17:09:24+00:00
repaired crash when marking as finsihed refine jobs that hadnt even started and copy pipeline.star upon calling cb_run_i
- - - - -
f0f4fd88 by Dari Kimanius at 2016-01-21T18:10:49+01:00
Merge https://bitbucket.org/tcblab/relion-devel
- - - - -
f9562eec by Sjors Scheres at 2016-01-22T13:50:29+00:00
repaired bug empty MDin from fn_data
- - - - -
57e49d67 by Dari Kimanius at 2016-01-22T15:15:50+01:00
Fixed FLTK compilation error due to bad include-directives in the cmake instructions.
- - - - -
621049f3 by Sjors Scheres at 2016-01-22T14:16:08+00:00
repaired bug that job files weren't saved in output dirs
- - - - -
02f38cbc by Sjors Scheres at 2016-01-22T14:35:49+00:00
Merged in tcblab/relion-devel-tcblab (pull request #10)
Accelerated getFourierTransformsAndCtfs and batched FFT-calls in autopicker
- - - - -
e28d46dc by Sjors Scheres at 2016-01-22T14:38:16+00:00
Merged tcblab/relion-devel into master
- - - - -
e67dcfe7 by Sjors Scheres at 2016-01-22T14:39:17+00:00
allow visualisation of re-extracted coordinates thru the gui
- - - - -
352adb95 by Sjors Scheres at 2016-01-22T14:39:44+00:00
added unistd.h to includes to get sleep function
- - - - -
bbd5250b by Bjoern Forsberg at 2016-01-22T15:41:05+01:00
Merged tcblab/relion-devel into master
- - - - -
82cb0ea1 by Sjors Scheres at 2016-01-22T15:07:43+00:00
also allow visualisation of helical segments
- - - - -
dfb5a12c by Sjors Scheres at 2016-01-22T16:19:04+00:00
Merged in scheres/relion-devel-lmb (pull request #11)
Master
- - - - -
131065ed by Dari Kimanius at 2016-01-22T17:20:02+01:00
Merged tcblab/relion-devel into master
- - - - -
6db4efac by Sjors Scheres at 2016-01-22T16:26:15+00:00
provide alias from running and better checking uniq aliases
- - - - -
c0aacc29 by Sjors Scheres at 2016-01-22T16:55:28+00:00
small bug alias of continued refine jobs
- - - - -
c69e1975 by Bjoern Forsberg at 2016-01-25T10:24:30+01:00
Single/double-compatibility fix
- - - - -
6e7d69c7 by Bjoern Forsberg at 2016-01-25T10:25:14+01:00
Fixes to cmake external project build-process
- - - - -
cc3c83ba by Bjoern Forsberg at 2016-01-25T10:36:12+01:00
Adding new FLTK-dep. files to GUI-omission list
- - - - -
b47fca5d by Bjoern Forsberg at 2016-01-25T10:55:32+01:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
68ac5e6a by Bjoern Forsberg at 2016-01-26T15:40:01+01:00
Removed BackProjection-specific GPU-streams
fixes issue #82
- - - - -
09881e67 by Dari Kimanius at 2016-01-26T16:25:17+01:00
Added memory some tracing code.
- - - - -
8e0383b9 by Dari Kimanius at 2016-01-26T16:26:13+01:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
4c1ecb90 by Bjoern Forsberg at 2016-01-26T17:03:17+01:00
Adding missing syncs which are needed in theory
Their absence has not yet caused errors, but could, if we are unlucky.
Refer to issue #82 for an example of errors and symptoms.
- - - - -
4480e299 by Sjors Scheres at 2016-01-27T09:20:34+00:00
better explanation import coordinate files
- - - - -
79345b17 by Dari Kimanius at 2016-01-27T12:36:16+01:00
Major revision of BP and WAVG kernels to dramatically reduce image size depenedent memory footprint.
Weight averaged translations are now produced on the fly inside the BP-kernel.
- - - - -
e067aeda by Sjors Scheres at 2016-01-27T12:47:39+00:00
fl_mesage window with metatdata
- - - - -
61059459 by Sjors Scheres at 2016-01-27T12:48:04+00:00
more info on imports
- - - - -
c988b590 by Sjors Scheres at 2016-01-27T12:50:30+00:00
toggle continue issue postprocessing repaired
- - - - -
ef3830ca by Bjoern Forsberg at 2016-01-27T13:58:58+01:00
New syntax for device-affinity of ranks/threads
Syntax examples and explanation, assuming a machine with 4 GPUs;
A)
--gpu
No indices will distribute threads and ranks as uniformly as possible
B)
mpirun -n 3 ... -j 2 --gpu 0,2:1,3
A colon delimits each rank, and the integers denote the devices to be
used for that rank. In this case, the number of integers match the
number of threads so the mapping will be (slave-rank/thread):
(1/0)->0
(1/1)->2
(2/0)->1
(2/1)->3
C)
mpirun -n 4 ... -j 2 --gpu 0,2:1,3
As B), but the slave-rank which has no indices is treated as in A),
resulting in the same mapping as B), and additionally
(3/0)->2
(3/1)->1 (uniform distribution of this rank over all devices)
D)
mpirun -n 4 ... -j 3 --gpu 2:2:1,3
Slave ranks 1 and 2 are restricted to device 2, slave 3 is restricted to
devices 1 and 3. All ranks simply distribute themselves uniformly across
their restrictions, which results in a higher memory-load on device 2, and
no utilization of device 0 (note that each rank resident on a device has
shared memory object for its threads on that device, so device 2 has roughly
twice the memory load of device 1, in this case):
(1/0)->2
(1/1)->2
(1/2)->2
(2/0)->2
(2/1)->2
(2/2)->2
(3/0)->1
(3/1)->3
(3/2)->1
- - - - -
f09ffa81 by Sjors Scheres at 2016-01-27T13:00:31+00:00
slightly blue background for bad-contrasty screens
- - - - -
d70b8b8e by Bjoern Forsberg at 2016-01-27T14:15:00+01:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
c2ec12fb by Dari Kimanius at 2016-01-27T14:27:17+01:00
Reduced the memroy footprint for wdiff2s_AA and wdiff2s_XA.
- - - - -
96ea75a5 by Dari Kimanius at 2016-01-27T14:28:02+01:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
b9bb4f37 by Sjors Scheres at 2016-01-27T14:15:24+00:00
Merged in tcblab/relion-devel-tcblab (pull request #12)
Norm-correction fix, advanced device assignment options and an order of magnitude reduced memory footprint
- - - - -
f2ba50eb by Bjoern Forsberg at 2016-01-27T15:20:59+01:00
device-id manipulation put into function (cleanup)
- - - - -
83b3e654 by Sjors Scheres at 2016-01-27T14:59:00+00:00
Merged tcblab/relion-devel into master
- - - - -
ff1f7723 by Bjoern Forsberg at 2016-01-27T21:09:46+01:00
non-mpi refine has new dev-affinity syntax
- - - - -
0a1cb272 by Bjoern Forsberg at 2016-01-27T21:18:00+01:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
7146e5b6 by Bjoern Forsberg at 2016-01-28T11:40:51+01:00
MPI-enabled autopicker uses GPUs
Same syntax as refinement, that is colon-separated indices for ranks.
As there is no thread-parallelism in the autopicker, comma-separated
indices are not necessary, and have no effect:
Default behaviour:
Requires NO indices to be suppied, and is device = rank % #devices,
so that if there are 3 devices and 5 threads, they will map as
(rank->device)
0->0
1->1
2->2
3->0
4->1
When mapping three ranks to three devices, threads (comma-sep) mean
nothing:
--gpu 0:1:2 <=> --gpu 0,0:1,2:2
To place two ranks on device 0 and one rank on device 1,
--gpu 0:0:1
Currently, the GPU-autopicker requires memory proportional to the
number of orientations (--ang 5 => 72 orientations), which can be
quite substantial. A single rank can handle the case where these
do not fit into the GPU memory, but ranks are NOT aware of each
other, meaning that multiple ranks may cause out-of-memory errors
in the present implementation. This will likely be fixed in the
future, so as to reduce memory usage and/or make ranks mutually
aware.
- - - - -
ccfa0da1 by Bjoern Forsberg at 2016-01-28T11:54:11+01:00
dev-id parsing uses relion-native text-to-int func
- - - - -
2bce4098 by Bjoern Forsberg at 2016-01-28T14:20:56+01:00
fixes to default gpu_ids in non-mpi refine
- - - - -
3cde65d4 by Dari Kimanius at 2016-01-28T15:15:19+01:00
GPU memory allocated by custom allocator is readjusted according to the total memory assigned by the user, accounting for additional allocations.
- - - - -
d106437c by Dari Kimanius at 2016-01-28T15:17:21+01:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
47b90a04 by Sjors Scheres at 2016-01-28T15:05:05+00:00
merged again with Shaoda
- - - - -
ffd821d8 by Dari Kimanius at 2016-01-28T16:53:11+01:00
Improved allocation semantics and user output info.
- - - - -
37020dae by Sjors Scheres at 2016-01-28T16:28:39+00:00
bug with extraction of movieparticles from Refine3D/data.star files: retain rlnOriginalParticleName!
- - - - -
a61f9c01 by Bjoern Forsberg at 2016-01-29T09:32:06+01:00
powerClass on GPU, getFTandCTFs 10-20% faster
- - - - -
5d7187db by Bjoern Forsberg at 2016-01-29T11:25:02+01:00
selTranslate on GPU, memcpy-cleanup needed
- - - - -
42faccfc by Bjoern Forsberg at 2016-01-29T15:10:09+01:00
getFTandCTFs mem-cpy cleanup and some acceleration
- - - - -
0b80ae67 by Bjoern Forsberg at 2016-01-29T16:25:19+01:00
Minor improvements to getFTandCTFs flow
- - - - -
ac7c0959 by Bjoern Forsberg at 2016-01-29T17:06:48+01:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
02cdd06e by Sjors Scheres at 2016-01-29T17:31:17+00:00
repaired bug in extraction of movies from refined data.star
- - - - -
27ac7af9 by Sjors Scheres at 2016-01-29T17:31:33+00:00
added small functionalities
- - - - -
7a8f4b62 by Bjoern Forsberg at 2016-02-01T11:10:43+01:00
Minor fixes for double-prec GPU build
- - - - -
f93279cc by Bjoern Forsberg at 2016-02-01T15:18:56+01:00
Loop thread-wise additive reduction across devices
Each thread adds it's contribution to the weights in a refinement-
iteration. Since threads on the same device share this object, the
weights get to be too large when threads in the same rank run on
the same device, since all add their joint contribution. Instead,
each such shared device-object should contribute.
this fixes issue #84
Note that issue #83 remains unfixed in this commit, such that the
extra normalization needs to be omitted in the GPU-code for a
direct CPU vs. GPU comparison to be valid.
- - - - -
7391d266 by Bjoern Forsberg at 2016-02-01T15:52:50+01:00
Added extra probability-normalization to CPU code
As implemented on the GPU-code in commits
f65b0b0 and 2b02a62
this fixes issue #83
- - - - -
5a289474 by Sjors Scheres at 2016-02-01T16:08:58+00:00
Merged in tcblab/relion-devel-tcblab (pull request #13)
Bugfixes, 2D-refinemnet speedup, and more precise custom allocation
- - - - -
3a252a41 by Sjors Scheres at 2016-02-01T17:26:07+00:00
Merged tcblab/relion-devel into master
- - - - -
225a9317 by Bjoern Forsberg at 2016-02-02T15:20:49+01:00
Ranks without device-ids use default settings
I overlooked this case before, somehow
- - - - -
e019d5a8 by Bjoern Forsberg at 2016-02-02T15:29:12+01:00
system-safe memory artihmetic for GPU-allocs
The requested memory was formulatyed as
size_t mem = ... ;
int allocSize = (int) mem - ... ;
GPUmalloc(allocSize);
where the int-cast of "mem" was unsafe, in that if "mem" was larger
than the system int, a negative value was returned. The new method
uses size_t for all these cases, and where subtraction is needed,
the minuend and subtrahend (yes, that's what they're called, I
looked it up) are compared for size, not their difference against 0.
this fixes issue #85
- - - - -
0bde04ee by Bjoern Forsberg at 2016-02-02T17:16:01+01:00
Merged tcblab/relion-devel into master
- - - - -
d1b6d64f by Sjors Scheres at 2016-02-02T17:16:46+00:00
import mrcs as 2d references
- - - - -
49b0d258 by Sjors Scheres at 2016-02-02T17:19:06+00:00
dont die if nodes dont have parent process
- - - - -
f8c61f42 by Sjors Scheres at 2016-02-02T17:19:29+00:00
display alias in current job entry
- - - - -
cf8403cb by Sjors Scheres at 2016-02-03T09:12:30+00:00
Merged in tcblab/relion-devel-tcblab (pull request #14)
Memory allocation fix, minor to device-id defaults in refine_mpi
- - - - -
ecd008f4 by Sjors Scheres at 2016-02-03T09:15:40+00:00
Merged tcblab/relion-devel into master
- - - - -
02ab7d5f by Sjors Scheres at 2016-02-03T09:16:16+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
7e6e1ec1 by Bjoern Forsberg at 2016-02-03T14:57:30+01:00
Added syncs for exp(weights) on classStreams
There is an issue (issue #87) which was behaving incredibly sporadic,
with this commit it starts to behave consistently.
- - - - -
f81a2bac by Bjoern Forsberg at 2016-02-03T15:00:18+01:00
Merged tcblab/relion-devel into master
- - - - -
35ac601d by Sjors Scheres at 2016-02-03T15:32:08+00:00
merge with Shaoda: display fourier TFs and repaired bug of translational searches with helices
- - - - -
e078aa55 by Sjors Scheres at 2016-02-04T13:01:04+00:00
repaired bug with reading empty pipeline
- - - - -
583357d0 by Sjors Scheres at 2016-02-04T13:01:35+00:00
repaired bug continue manualpick and recursive alias names
- - - - -
23730a7f by Sjors Scheres at 2016-02-04T17:48:37+00:00
replaced 12-digit uniqfilenames with consecutive job numbering. TODO: remove some code to keep functionality for 12-digit directories later
- - - - -
1eca7565 by Erato.bjoern at 2016-02-04T20:12:40+01:00
Launch error checks added
- - - - -
daf57c4f by Sjors Scheres at 2016-02-05T10:23:16+00:00
cleaned up buggy mess with pipeline filenames
- - - - -
3e9939ac by Erato.bjoern at 2016-02-05T14:07:28+01:00
Added launch error checks after all kernel calls
Some main kernel-launches use a thread-private cudaError_t
also repaired accidental deletion of allocator-functinalities in
previous commit
- - - - -
6aac6605 by Sjors Scheres at 2016-02-08T12:15:00+00:00
changed explanation mask
- - - - -
a4d21ce3 by Sjors Scheres at 2016-02-08T16:00:02+00:00
repaired various small bugs#
- - - - -
60cc113a by Bjoern Forsberg at 2016-02-10T10:36:19+01:00
size_t used instead of int in some locations
- - - - -
5ff142a2 by Sjors Scheres at 2016-02-10T14:03:13+00:00
changed resmap output
- - - - -
7649d777 by Sjors Scheres at 2016-02-10T14:03:32+00:00
try to prevent bus errors in last iteration of classification
- - - - -
42e1bb01 by Bjoern Forsberg at 2016-02-10T23:20:48+01:00
init-check of windowFourTransf. fixed
note: init-check of diff2fine NOT fixed
- - - - -
2a6d60df by Bjoern Forsberg at 2016-02-11T10:35:33+01:00
Fixed minor race-condition
this was reported by cuda-memcheck --tool racecheck,
but may not have been an issue, due to implicit thread
sequential execution.
- - - - -
b71506b0 by Sjors Scheres at 2016-02-11T14:32:30+00:00
merge again with Shaoda
- - - - -
e99e3a9d by Sjors Scheres at 2016-02-12T11:28:51+00:00
more output for finding errors with movie expansion
- - - - -
65aaa6b9 by Sjors Scheres at 2016-02-14T18:39:04+00:00
added MovieRefine job-type with movie extraction and reifnement combiend
- - - - -
8feb9303 by Sjors Scheres at 2016-02-14T20:35:58+00:00
started with polishing per micrograph
- - - - -
c99da710 by Sjors Scheres at 2016-02-14T20:55:24+00:00
now polishing (without defocus and beamtilt correction) seems to work per-micrograph
- - - - -
90c7b952 by Bjoern Forsberg at 2016-02-15T09:55:57+01:00
Size_t sprinkles + rm unused streams
- - - - -
48d756a5 by Bjoern Forsberg at 2016-02-15T10:04:02+01:00
Merge commit '90c7b95'
- - - - -
1195c448 by Bjoern Forsberg at 2016-02-15T12:26:08+01:00
Fixed non-custom-alloc comptibility
- - - - -
634a4d5d by Sjors Scheres at 2016-02-16T18:54:24+00:00
now per-batch polishing also working. Still minor problem reproducibility ber-batch refine!#
- - - - -
fa872c29 by Sjors Scheres at 2016-02-17T18:16:17+00:00
repaired bug restarting refine outputnodenames
- - - - -
f1722b1b by Sjors Scheres at 2016-02-17T18:22:01+00:00
added option to skip static bool for only randomising once to allow movie processing in batches
- - - - -
8ba36bd6 by Sjors Scheres at 2016-02-17T18:22:42+00:00
repaired small bug batch names
- - - - -
2ca465c2 by Sjors Scheres at 2016-02-17T18:27:22+00:00
now movie-processing in batches gives same results as not doing batches
- - - - -
51c71352 by Sjors Scheres at 2016-02-17T18:40:23+00:00
small change comment
- - - - -
19228c0b by Sjors Scheres at 2016-02-17T19:37:10+00:00
added phase-shift to CTF model
- - - - -
0d94cfb2 by Sjors Scheres at 2016-02-17T19:37:41+00:00
added avg_movie_frames to movie-refinement and polishing
- - - - -
e4bbdf3f by Sjors Scheres at 2016-02-17T20:26:13+00:00
dont allow selection for manualpick/display, added selection of coords file in subset jobtype
- - - - -
8d50a907 by Sjors Scheres at 2016-02-17T20:26:41+00:00
more helpful error message to finsd what's wrong in movie expansion
- - - - -
c662eddd by Bjoern Forsberg at 2016-02-18T14:33:23+01:00
Prep for mem-leak fixes
Purposely fragmented the mem-leak to allow future clean-up and/or
additional fixes to be smoother. On this commit a memleak is made
obvious under the default allocator. It was previously invisible
to cuda-memcheck, since there was no call to cudaDeviceReset().
A un unused function CudaGlobalPtr.resize() host is added, to be
used in a future commit.
- - - - -
de727209 by Bjoern Forsberg at 2016-02-18T14:37:13+01:00
Leak fix to default allocator.
The issue seems to be that the default template argument for any
CudaGlobalPtr is CuatomAlloc=True, regardless of whether the custom
allocator is used. This results in calls to cudaFree() being avoided.
It is unclear to me how the allocation can remain unaffected. None-
theless, the leak is fixed. Not only does cuda-memcheck report less
leaking memory, but the positive memory gradient (as reported by
nvidia-smi) is also reduced, in fact almost eliminated.
- - - - -
4f75ee40 by Bjoern Forsberg at 2016-02-18T16:03:06+01:00
Leak fix to default allocator II
I the problem fixed here appears to be that the temporary array
allocated in some cub/thurst-dedicated functions get marked for
deletion without actually getting deleted, as only the custom
allocator collects the garbage merked for deletion. When the
default allocator is used, this makes for a leak of temporary
storage.
This in part adresses issue #88
- - - - -
18b158b6 by Bjoern Forsberg at 2016-02-19T10:19:11+01:00
Leak fix to default allocator III
The problem herein fixed, is that the projectors and backprojectors
in the MlDeviceBundle were not cleared before program termination.
Even worse, they appear to have been NEVER cleared, so that upon a
new iteration, new texture-arrays were simply malloc:Ed and copied
into the texture space. For this purpose, both back- and projectors
are now cleared in MlDeviceBundle::resetData(), just as is the
ProjectionPlans. This should be verified to be correct procedure.
Also, this was tested with non-MPI compatibility, so the MPI-version
should be updated to comply.
- - - - -
a1da3eac by Sjors Scheres at 2016-02-20T17:02:20+00:00
continued debugging pipeline plus new option for cleaning job directories
- - - - -
012f56a3 by Sjors Scheres at 2016-02-20T18:15:08+00:00
load previous selection
- - - - -
911caa46 by Bjoern Forsberg at 2016-02-22T11:29:30+01:00
Memory requests-numbers are in bytes. All of them.
Alos, the requested and available memory are kept separate.
- - - - -
02789a16 by Dari Kimanius at 2016-02-26T09:55:58+01:00
Backproject wrap up fixed.
- - - - -
5fe6cf99 by Sjors Scheres at 2016-02-26T17:15:34+00:00
repaired bug backup_selection.star
- - - - -
3a229adf by Sjors Scheres at 2016-02-26T17:15:53+00:00
edit general notes
- - - - -
f52945cd by Sjors Scheres at 2016-02-26T17:16:16+00:00
bug if undefined maxres or negative bfac
- - - - -
33df062f by Sjors Scheres at 2016-02-26T17:16:39+00:00
repaired memory leak when clearing metadatatable
- - - - -
5e875dce by Sjors Scheres at 2016-03-04T13:55:19+00:00
keep debugging
- - - - -
2bb725c8 by Sjors Scheres at 2016-03-08T11:39:06+00:00
add solvent_fsc to GUI, and repair some bugs minor
- - - - -
b19e8524 by Sjors Scheres at 2016-03-08T11:39:41+00:00
repaired bug nr movie frames in movie extract
- - - - -
b0564b22 by Sjors Scheres at 2016-03-08T12:58:25+00:00
took out some old comments
- - - - -
9251fcdc by Sjors Scheres at 2016-03-08T13:02:50+00:00
no longer allow very small changes in orientations for auto-refine#
- - - - -
f2b7689f by Sjors Scheres at 2016-03-08T15:09:08+00:00
added debug_reconstruction option to write out data_real.mrc files every iteration
- - - - -
87b90736 by Dari Kimanius at 2016-03-09T11:49:56+01:00
Custom allcoator auto-free is functional when cached allocation is turned off.
- - - - -
00e82372 by Sjors Scheres at 2016-03-09T14:22:00+00:00
minor bugs in ctffind runner
- - - - -
36d5bc2a by Sjors Scheres at 2016-03-09T17:33:11+00:00
keep stacks open for reading multiple particles in one pool
- - - - -
cafce5bd by Sjors Scheres at 2016-03-09T18:00:09+00:00
forgot to switch off DEBUG_SIMULTANEUS_READ
- - - - -
d81d105b by Sjors Scheres at 2016-03-10T17:34:17+00:00
use stats instead of fopen for exist of file
- - - - -
750552bf by Sjors Scheres at 2016-03-10T17:34:48+00:00
open/close only new stacks when prereading all particles into RAM
- - - - -
e688a6c4 by Sjors Scheres at 2016-03-10T17:35:34+00:00
new Compute tabs for refine jobs, and do not remake .Nodes directory all the time in GUI
- - - - -
6906b22e by Sjors Scheres at 2016-03-10T17:35:57+00:00
repaired bug with rescaled particles and manually set angpix
- - - - -
a4566092 by Sjors Scheres at 2016-03-10T17:36:32+00:00
dont read particles into exp_imgs for subtomograms in order to save RAM
- - - - -
ea31508d by Sjors Scheres at 2016-03-11T15:19:39+00:00
repaired silly mistake of commenting out entire reconstruction of half2, oops
- - - - -
8537bb32 by Sjors Scheres at 2016-03-11T16:57:03+00:00
another silly mistake with the new reading of particles
- - - - -
cd6fc33d by Sjors Scheres at 2016-03-11T16:59:14+00:00
included eol
- - - - -
1dbfe628 by Sjors Scheres at 2016-03-14T15:23:34+00:00
use --pool 1 with preread_images
- - - - -
1d4ade9d by Sjors Scheres at 2016-03-14T15:24:01+00:00
added touch, copy and mktree
- - - - -
b8f1d91a by Sjors Scheres at 2016-03-14T15:24:34+00:00
added --no_parallel_disc_io before continue read statement
- - - - -
780204ab by Sjors Scheres at 2016-03-14T17:00:39+00:00
debugging new way of handling .Nodes
- - - - -
6a475160 by Sjors Scheres at 2016-03-15T17:24:32+00:00
removed stray cerr line
- - - - -
54937c1d by Sjors Scheres at 2016-03-15T17:25:05+00:00
pass maxres and FOM of CTF estimation to particles.star
- - - - -
7d07bf4f by Sjors Scheres at 2016-03-15T17:25:22+00:00
phase-shifts and new ctffind4 wrapper
- - - - -
9fa8c3d7 by Sjors Scheres at 2016-03-16T12:43:53+00:00
get rid of --preread_images error messages
- - - - -
3e39e465 by Sjors Scheres at 2016-03-16T12:44:24+00:00
open only optimiser.star files inside current output directory
- - - - -
3a7b742f by Dari Kimanius at 2016-03-16T16:45:25+01:00
Byte guards can be activated in custom-allocator with -DCustomAllocMemGuards=ON during build.
CA prints stack-trace of the allocation call if guard bytes corruption is detected during free calls.
Current implementation is suboptimal due to use of memset and memcopy instead of kernel-calls on a specialized stream.
- - - - -
c4e6baee by Dari Kimanius at 2016-03-16T18:34:57+01:00
Demangled output from custom memguard check.
- - - - -
a16c260a by Dari Kimanius at 2016-03-16T18:35:41+01:00
Fixed faulty if-statment inside windowFourierTransform2.
- - - - -
8a5959dd by Sjors Scheres at 2016-03-17T07:54:34+00:00
repaired bug new browse button when default dir is NULL
- - - - -
882ebd85 by Sjors Scheres at 2016-03-17T14:05:43+00:00
repaired bug found by Shaoda that imagesize became unequal for 2 halves because ave_Pmax was on different sides of 0.1
- - - - -
48832c27 by Sjors Scheres at 2016-03-17T17:26:33+00:00
added mpiexec flag for better restartying upon request by Jake
- - - - -
25c9a06e by Sjors Scheres at 2016-03-17T17:27:44+00:00
dont always read backup selection file
- - - - -
18de002b by Sjors Scheres at 2016-03-17T17:28:10+00:00
double-click to open run.out and run,.err
- - - - -
ace18b42 by Sjors Scheres at 2016-03-17T17:29:02+00:00
now movie-refinement cornelius doesnt crash anymore, results are still not good!
- - - - -
31c7a8d0 by Bjoern Forsberg at 2016-03-18T15:14:07+01:00
Fix tocoarse-pass GPU-refinement scale-correction
fixes issue #91
- - - - -
24f26c17 by Sjors Scheres at 2016-03-18T17:01:55+00:00
minor changes
- - - - -
b2beedc9 by Sjors Scheres at 2016-03-21T14:27:05+00:00
further debugging of movie-refinement
- - - - -
48d8cd62 by Bjoern Forsberg at 2016-03-22T10:17:27+01:00
rm:ed GPU-legacy check in windowFourierTransform2
- - - - -
503a35ee by Sjors Scheres at 2016-03-22T11:33:32+00:00
added 2D plotting options by including CPlot2D
- - - - -
b6b8913d by Sjors Scheres at 2016-03-22T11:33:58+00:00
added more plotting options
- - - - -
211d5010 by Sjors Scheres at 2016-03-22T14:36:24+00:00
added visualisation of logfile.pdf on GUI for postprocessing and polishing
- - - - -
d0d15ed8 by Sjors Scheres at 2016-03-22T14:43:06+00:00
redirect gs output to devnull
- - - - -
5bc7f076 by Sjors Scheres at 2016-03-23T15:13:13+00:00
added a background image
- - - - -
a9c5f081 by Sjors Scheres at 2016-03-23T15:21:05+00:00
dont put forgot button if background image does not exist
- - - - -
5764786e by Sjors Scheres at 2016-03-23T16:29:16+00:00
hide background image when not needed
- - - - -
28bc5ae5 by Sjors Scheres at 2016-03-24T11:55:18+00:00
Merged in tcblab/relion-devel-tcblab (pull request #15)
Master
- - - - -
92725142 by Sjors Scheres at 2016-03-24T11:55:38+00:00
Merged tcblab/relion-devel into master
- - - - -
51e0a0bc by Bjoern Forsberg at 2016-03-24T12:56:28+01:00
Merged tcblab/relion-devel into master
- - - - -
3080254c by Sjors Scheres at 2016-03-24T11:57:02+00:00
minor changes
- - - - -
bdbbaa66 by Sjors Scheres at 2016-03-24T11:57:14+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
5e09b0da by Bjoern Forsberg at 2016-03-31T14:48:43+02:00
Indexing fix to pdf_orientations
When a particle was classified as contributing to more than
a single reconstructed class, the directions/orientations
were always selected starting from the beginning of the
array holding the direction-indices for all classes. This
mistake was made ONLY in the summation of the orientational
distribution, not the backprojection.
Essentially, each image contributed correctly to the
distribution of directions for the FIRST class it was class-
ified as, and contributed incorrectly to the distribution
of directions for all subsequent classes it was assigned to.
This means that class 0 always had the correct distribution
(up until and including 2nd iter), whereas other classes where
roughly correct. The latter because
a) many images were assigned as NOT contributing to class 0,
in which case they contributed correctly to the first class
they WERE assigned to.
b) the orientational assignment is roughly the same, irrespective
of class, at low convergence ( low class differentiation).
this fixes issue #95
- - - - -
9d16f49c by Bjoern Forsberg at 2016-04-01T13:28:50+02:00
Merged tcblab/relion-devel into master
- - - - -
74c24898 by Sjors Scheres at 2016-04-04T09:29:02+01:00
Merged in tcblab/relion-devel-tcblab (pull request #16)
Indexing fix to pdf_orientations
- - - - -
33704771 by Sjors Scheres at 2016-04-04T09:29:20+01:00
Merged tcblab/relion-devel into master
- - - - -
7ed6a773 by Sjors Scheres at 2016-04-04T09:31:31+01:00
added warning for empty micrographs
- - - - -
dbf0f36e by Sjors Scheres at 2016-04-04T09:32:00+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
7d9e8796 by Sjors Scheres at 2016-04-06T17:40:55+01:00
added ctffind4.1 wrapper, to be tested...
- - - - -
e8693c68 by Sjors Scheres at 2016-04-06T17:41:21+01:00
added functionality for flowcharts in the gui
- - - - -
c45c9798 by Dari Kimanius at 2016-04-06T18:44:58+02:00
The GPU-memory distribution over ranks is now aware of the MPI-context and can handle multiple ranks allocating on the same device.
The flag --gpu_memory_per_mpi_rank is changed to --free_gpu_memory that sets the amount of memory left free on each device after allocation. Default is zero.
- - - - -
1411dc65 by Dari Kimanius at 2016-04-06T18:47:29+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
f1c6e12d by Dari Kimanius at 2016-04-06T18:49:56+02:00
Merged tcblab/relion-devel into master
- - - - -
a12ac3d2 by Sjors Scheres at 2016-04-07T10:29:16+01:00
Merged in tcblab/relion-devel-tcblab (pull request #17)
The GPU-memory distribution over ranks is now aware of the MPI-context and can handle multiple ranks allocating on the same device.
- - - - -
2a933634 by Sjors Scheres at 2016-04-07T10:40:11+01:00
added --check_duplicates in joinstar
- - - - -
53b1a95a by Sjors Scheres at 2016-04-07T10:40:28+01:00
more debugging
- - - - -
c8908479 by Sjors Scheres at 2016-04-07T10:40:44+01:00
try and resolve strange conflict upon merge with Dari
- - - - -
60e108ff by Sjors Scheres at 2016-04-07T10:44:03+01:00
try and resolve strange conflict upon merge with Dari
- - - - -
26cbbbd4 by Sjors Scheres at 2016-04-07T10:44:26+01:00
Merged tcblab/relion-devel into master
- - - - -
e025c0de by Sjors Scheres at 2016-04-07T11:18:35+01:00
still trying to resolve conflict...
- - - - -
a5e7846b by Bjoern Forsberg at 2016-04-07T13:44:57+02:00
Fix to precision issue in offfset priors
The single-prec cast of small offset priors was made
before mean-normalization, so if all priors were sufficiently
small, they all got set to zero, resulting in no weights.
- - - - -
a34c5be7 by Dari Kimanius at 2016-04-07T14:25:33+02:00
Merged tcblab/relion-devel into master
- - - - -
8836f7cc by Sjors Scheres at 2016-04-07T14:16:04+01:00
Merged in scheres/relion-devel-lmb (pull request #18)
update main repo with relion-2.0 updates
- - - - -
cb787495 by Dari Kimanius at 2016-04-07T15:32:59+02:00
Merged tcblab/relion-devel into master
- - - - -
25690294 by Sjors Scheres at 2016-04-07T15:35:07+01:00
added true for reading of pipelines
- - - - -
0f524f31 by Dari Kimanius at 2016-04-07T17:33:38+02:00
Fixed an indexing error in getAllSquaredDifferencesCoarse occurring when nr_particls is non-zero.
- - - - -
706a8d8a by Sjors Scheres at 2016-04-08T12:15:52+01:00
also unknown branches, and rescale genera flowcharts
- - - - -
4b474bea by Sjors Scheres at 2016-04-08T15:58:18+01:00
no more movie.star in movierefine
- - - - -
33f0597c by Sjors Scheres at 2016-04-11T12:01:46+01:00
repaired bug combineAllWeights, nt thru disk
- - - - -
320dd003 by Sjors Scheres at 2016-04-11T12:02:42+01:00
Merged tcblab/relion-devel into master
- - - - -
1ad814b7 by Sjors Scheres at 2016-04-11T12:04:43+01:00
Merged in scheres/relion-devel-lmb (pull request #19)
Master
- - - - -
ef304b30 by Sjors Scheres at 2016-04-11T17:05:34+01:00
check for finished CTF programs
- - - - -
3d3ac803 by Sjors Scheres at 2016-04-11T17:06:34+01:00
changed proposed by Yunxiao Zhang from Stanford
- - - - -
896ceaf9 by Sjors Scheres at 2016-04-11T17:26:35+01:00
merged with Shaosa: allow local averaging of helical segments for priors
- - - - -
a8bed2e6 by Sjors Scheres at 2016-04-11T17:27:43+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
7c5945df by Sjors Scheres at 2016-04-11T17:29:11+01:00
Merged in scheres/relion-devel-lmb (pull request #20)
included reshape from Yunxiao Zhang and merged again with Shaoda
- - - - -
f39a6183 by Bjoern Forsberg at 2016-04-13T14:26:49+02:00
Merged tcblab/relion-devel into master
- - - - -
4cc6f0b7 by Sjors Scheres at 2016-04-13T16:47:34+01:00
errno was a special variable on MAC OSX
- - - - -
5a27577f by Sjors Scheres at 2016-04-13T16:48:15+01:00
now GPU-options on the GPU, also for autopick, ctffind and motioncorr
- - - - -
9049ab4e by Sjors Scheres at 2016-04-13T17:08:58+01:00
small changes
- - - - -
c4cbccc6 by Sjors Scheres at 2016-04-13T17:09:26+01:00
reproducible class assignments in 1st iteration when using threads
- - - - -
6e0e7352 by Sjors Scheres at 2016-04-13T17:10:06+01:00
Merged in scheres/relion-devel-lmb (pull request #21)
Master
- - - - -
465d856d by Sjors Scheres at 2016-04-14T09:25:06+01:00
removed unnecessary read
- - - - -
74101794 by Sjors Scheres at 2016-04-14T09:45:47+01:00
small changes in gui
- - - - -
4bf7a204 by Bjoern Forsberg at 2016-04-14T11:04:27+02:00
Merged tcblab/relion-devel into master
- - - - -
21b0b26a by Sjors Scheres at 2016-04-14T16:03:45+01:00
debugging shrink and gpu functionality of autopicker
- - - - -
1373a747 by Sjors Scheres at 2016-04-14T16:32:14+01:00
die when cannot write to pipeline.star
- - - - -
fab41d90 by Sjors Scheres at 2016-04-14T16:39:48+01:00
Merged tcblab/relion-devel into master
- - - - -
01973602 by Sjors Scheres at 2016-04-14T16:40:15+01:00
Merged in scheres/relion-devel-lmb (pull request #22)
Master
- - - - -
21aa4e32 by Sjors Scheres at 2016-04-14T16:41:37+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
c18f3161 by Sjors Scheres at 2016-04-14T16:42:58+01:00
for now no shrink with autopicker
- - - - -
68d626e8 by Bjoern Forsberg at 2016-04-15T08:11:09+02:00
gpu-side fix equal to commit c4cbccc
- - - - -
3bddc455 by Bjoern Forsberg at 2016-04-15T16:53:40+02:00
Merged tcblab/relion-devel into master
- - - - -
1983cda2 by Sjors Scheres at 2016-04-15T16:23:48+01:00
Merged tcblab/relion-devel into master
- - - - -
d65004d0 by Bjoern Forsberg at 2016-04-15T17:40:52+02:00
fix to --shrink autopicker like Sjors CPU-mods
A few small modifications were made to the CPU-side autopicker in
commit 21b0b26. This commit mimics these edits to replicate auto-
picking results when --shrink is set to values other than 1.
- - - - -
6ae05dbc by Sjors Scheres at 2016-04-15T16:55:09+01:00
Merged in tcblab/relion-devel-tcblab (pull request #23)
Master
- - - - -
c622632c by Sjors Scheres at 2016-04-15T16:55:21+01:00
Merged tcblab/relion-devel into master
- - - - -
824d93d4 by Sjors Scheres at 2016-04-15T17:02:28+01:00
added shiny_post.mrc to outputnodes of polish to keep job in running status until finished
- - - - -
3f0bbcda by Bjoern Forsberg at 2016-04-15T18:22:28+02:00
Removing excessive correction to autpicker stddev
- - - - -
3422fd9f by Bjoern Forsberg at 2016-04-15T18:24:08+02:00
Merged tcblab/relion-devel into master
- - - - -
162881b8 by Sjors Scheres at 2016-04-18T10:01:30+01:00
Merged in tcblab/relion-devel-tcblab (pull request #24)
Removing excessive correction to autpicker stddev
- - - - -
6b068595 by Sjors Scheres at 2016-04-18T10:01:46+01:00
Merged tcblab/relion-devel into master
- - - - -
ced057c9 by Sjors Scheres at 2016-04-18T10:02:49+01:00
better handling of gpu_ids
- - - - -
86d1869a by Sjors Scheres at 2016-04-19T13:03:39+01:00
resize non-square 2D images
- - - - -
a7b7de56 by Sjors Scheres at 2016-04-19T13:12:00+01:00
more drastic shrinking in autopicker, plus added pixel size of references
- - - - -
5e680375 by Sjors Scheres at 2016-04-19T13:12:35+01:00
Merged in scheres/relion-devel-lmb (pull request #25)
Master
- - - - -
48a42f5d by Sjors Scheres at 2016-04-19T14:26:47+01:00
Shaoda: 1D profiles in display of FFT and picking cross-overs for helical tubes
- - - - -
608dbdf0 by Sjors Scheres at 2016-04-19T14:42:03+01:00
further changes to autopicker
- - - - -
c6457395 by Sjors Scheres at 2016-04-19T14:43:35+01:00
Merged in scheres/relion-devel-lmb (pull request #26)
Master
- - - - -
7cb57760 by Sjors Scheres at 2016-04-19T16:18:17+01:00
Merged tcblab/relion-devel into master
- - - - -
8a2857cf by Sjors Scheres at 2016-04-19T17:01:03+01:00
back to less aggresive shrinking
- - - - -
af3c68bf by Bjoern Forsberg at 2016-04-19T19:12:55+02:00
-j >1 cmake build (make) waits for ext. fftw/fltk
- - - - -
846cb655 by Bjoern Forsberg at 2016-04-20T10:59:14+02:00
Highpass filter added to autopick, as in CPU-code
- - - - -
abda7fc7 by Bjoern Forsberg at 2016-04-20T11:00:25+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
6a9a39c9 by Sjors Scheres at 2016-04-20T14:36:58+01:00
back to less aggressive shrinking, now seems to work for non-square mics as well
- - - - -
61a580e3 by Sjors Scheres at 2016-04-20T14:38:09+01:00
Merged in scheres/relion-devel-lmb (pull request #27)
Master
- - - - -
481a607f by Bjoern Forsberg at 2016-04-20T15:49:55+02:00
Merged tcblab/relion-devel into master
- - - - -
29890011 by Sjors Scheres at 2016-04-20T15:47:43+01:00
Merged in tcblab/relion-devel-tcblab (pull request #28)
Master
- - - - -
1cd0524a by Sjors Scheres at 2016-04-20T15:47:57+01:00
Merged tcblab/relion-devel into master
- - - - -
ab4d60ce by Sjors Scheres at 2016-04-20T16:33:05+01:00
prevent empty plots
- - - - -
66adff04 by Sjors Scheres at 2016-04-20T16:33:25+01:00
warn when no angpix is available
- - - - -
a63ed67c by Sjors Scheres at 2016-04-20T16:38:22+01:00
small changes by Shaoda
- - - - -
4bdc5ad8 by Sjors Scheres at 2016-04-20T16:42:18+01:00
Merged in scheres/relion-devel-lmb (pull request #29)
Master
- - - - -
2982654e by Sjors Scheres at 2016-04-20T16:42:34+01:00
Merged tcblab/relion-devel into master
- - - - -
4086e544 by Sjors Scheres at 2016-04-21T10:57:54+01:00
warn if users tries to submit job with too many MPIs locally
- - - - -
8388dedf by Dari Kimanius at 2016-04-22T10:42:59+02:00
All devices are reset after each iteration.
- - - - -
66fb7eaf by Dari Kimanius at 2016-04-22T10:43:47+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
508ca1b4 by Dari Kimanius at 2016-04-22T17:33:59+02:00
Fixes an error with the bundle distribution amongst threads.
- - - - -
5c19c9f3 by Dari Kimanius at 2016-04-22T17:44:11+02:00
Merged tcblab/relion-devel into master
- - - - -
5387c728 by Sjors Scheres at 2016-04-22T16:56:54+01:00
Merged in tcblab/relion-devel-tcblab (pull request #30)
Hopefully fixes the "out of memory error"
- - - - -
fecb4f13 by Sjors Scheres at 2016-04-22T16:57:07+01:00
Merged tcblab/relion-devel into master
- - - - -
05209ad6 by Dari Kimanius at 2016-04-25T15:19:54+02:00
Added timing in preprocessing.cpp (extraction). Changed timer output to three decimal rounded seconds.
- - - - -
83752cd3 by Bjoern Forsberg at 2016-04-25T16:08:04+02:00
Autopicker timings
- - - - -
81b6c71e by Bjoern Forsberg at 2016-04-25T16:08:26+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
6878bb3a by Bjoern Forsberg at 2016-04-25T17:01:22+02:00
Removed fftw_cleanup causing autopicker issue
The cleanup destroyed all references to fftw-objects, not just the
local transformer. Now we instead let it go out of scope and get
handled by the destructor, which cleans up the local copy only.
This fixes issue #99 (but we still don't know why that issue does
not appear when using a single reference as input for autopicking)
- - - - -
aadccfa7 by Bjoern Forsberg at 2016-04-26T09:56:03+02:00
Removed unused transformers in autopicker
- - - - -
ca1fb4f4 by Bjoern Forsberg at 2016-04-26T10:46:42+02:00
Batches of ffts in autopicker returns correct psis
FOM maps and psi-metadata are correct when being forced to batch
psi-dependent fft-transoforms due to shortage of gpu-mem.
this fixes issue #101
- - - - -
7214040b by Bjoern Forsberg at 2016-04-26T11:36:13+02:00
Clearing up some mem-estimation confusion
Autpicker uses custom alloc for some object, and default for others.
The code was "confused" about what objects should be included in the
overhead-estimations for either one.
- - - - -
f7e0a57c by Dari Kimanius at 2016-04-26T13:35:43+02:00
Runtime setup of cache-mode for custom allocator.
- - - - -
cd58babe by Dari Kimanius at 2016-04-26T13:36:13+02:00
No allocator caching for autopick.
- - - - -
ec2b942c by Dari Kimanius at 2016-04-26T13:45:27+02:00
Fixed the CachedAlloc-cmake-option.
- - - - -
6da59a01 by Dari Kimanius at 2016-04-26T14:59:41+02:00
Projector moved to higher scope in autopick.
- - - - -
7eb579a8 by Dari Kimanius at 2016-04-26T15:12:42+02:00
Cuda FFT transformer objects moved up in scope in autopick.
- - - - -
8cb3965a by Bjoern Forsberg at 2016-04-26T15:36:28+02:00
Some prel prime-factor warnings in autopicker
- - - - -
a0b50a61 by Bjoern Forsberg at 2016-04-26T15:38:33+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
9e715fda by Sjors Scheres at 2016-04-27T09:15:16+01:00
add clear of transformer
- - - - -
a4adb71c by Sjors Scheres at 2016-04-27T09:16:48+01:00
continuation of refine jobs
- - - - -
5eedacb1 by Sjors Scheres at 2016-04-27T09:17:22+01:00
made function to print nodename
- - - - -
9d015867 by Sjors Scheres at 2016-04-27T09:18:00+01:00
added sorting with autopicked coords
- - - - -
1dc7817e by Sjors Scheres at 2016-04-27T09:20:18+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
71414ef2 by Dari Kimanius at 2016-04-27T10:29:56+02:00
Autopick cleanup and added some new helper-functions and method in cudaFFT-class.
- - - - -
b3ec4eaf by Bjoern Forsberg at 2016-04-27T10:59:46+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
f0e2ee33 by Sjors Scheres at 2016-04-27T10:50:13+01:00
removed cleanup from computeFourierTransform in projector, which solves problem in autopicker
- - - - -
7313b69f by Sjors Scheres at 2016-04-27T10:51:24+01:00
added option to copy all particle images to a local scratch directory
- - - - -
d8953ad9 by Sjors Scheres at 2016-04-27T10:52:27+01:00
Merged in scheres/relion-devel-lmb (pull request #31)
Master
- - - - -
87ad35d8 by Sjors Scheres at 2016-04-27T10:53:28+01:00
Merged tcblab/relion-devel into master
- - - - -
1762630d by Dari Kimanius at 2016-04-27T12:43:23+02:00
Merged tcblab/relion-devel into master
- - - - -
20511ec0 by Bjoern Forsberg at 2016-04-27T16:33:07+02:00
All objects in GPU-side autipicker are worksize
This affects the performance positively, as adjustments to FFT-sizes
based on prime-factors do not result in size-discrepancies. This also
allows us to remove an intermediate arrey for data resizing (FT-
windowing), resulting in a smaller memory footprint on the GPU.
In the present state, job segmentation is non-functional, but this
will be fixed in the next commit, which will restructure the psi-
batching in the GPU-code and clean up the autopicker code.
- - - - -
74222ff1 by Sjors Scheres at 2016-04-27T15:52:04+01:00
added unblur wrapper and various small things
- - - - -
84172c29 by Sjors Scheres at 2016-04-27T15:52:35+01:00
Merged in scheres/relion-devel-lmb (pull request #32)
Merged tcblab/relion-devel into master
- - - - -
9c3ee4d4 by Sjors Scheres at 2016-04-27T15:52:45+01:00
Merged tcblab/relion-devel into master
- - - - -
a117bcb9 by Sjors Scheres at 2016-04-27T15:52:58+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
4cfe8f76 by Sjors Scheres at 2016-04-27T15:54:35+01:00
Merged in scheres/relion-devel-lmb (pull request #33)
added unblur wrapper and various small things
- - - - -
1c1b2666 by Sjors Scheres at 2016-04-27T15:54:46+01:00
Merged tcblab/relion-devel into master
- - - - -
1d4f7fee by Bjoern Forsberg at 2016-04-27T17:52:42+02:00
Batching psi-iters in cudaFFT-method
This makes psi-batching functional again (was made non-functional in
the previous commit).
The psi-proportional transformer is now allocated fully, which does
increase memory usage somewhat during batching. Only the tranform
itself is therefore batched, which allows more expedient calcualtions
during batching, since no data needs to be moved.
Also, small changes to FFT-prime factor warnings.
- - - - -
0fcec494 by Bjoern Forsberg at 2016-04-27T17:57:09+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
42ca5560 by Sjors Scheres at 2016-04-27T17:03:43+01:00
debugging GPU IDs motioncorr
- - - - -
3706845b by Bjoern Forsberg at 2016-04-27T18:03:46+02:00
Merged tcblab/relion-devel into master
- - - - -
fd10178b by Bjoern Forsberg at 2016-04-27T19:45:06+02:00
Adding fudge-factor to FFT temp-space estimation
This is to stabilize the behaviour at high loads, where batching of
psis might become necessary tue to either/or
a) many orientations
b) large micrograph sizes
The cuda API estimation prior to plan allocation is apparently not
really accurate, so we can get stuck if we run out of memory. Also,
with the recently implemented mod to batching, there is minial
overhead of batching, so it is easier to just batch, even when it
might be technically possible by a small margin to not batch psis.
For now, the warning of batching will still be printed.
- - - - -
7c626085 by Bjoern Forsberg at 2016-04-27T19:49:35+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
fe4eef12 by Bjoern Forsberg at 2016-04-27T20:03:36+02:00
Fix to estimate-fudge
- - - - -
d213f1fe by Sjors Scheres at 2016-04-28T09:50:05+01:00
Merged in tcblab/relion-devel-tcblab (pull request #34)
Autopicker worksized + checking FFT-primefactors
- - - - -
8079de2f by Sjors Scheres at 2016-04-28T09:50:39+01:00
Merged tcblab/relion-devel into master
- - - - -
05ab019a by Bjoern Forsberg at 2016-04-28T14:42:16+02:00
Enabling fwd/reverese-only cudaFFT transformers
Since the large (possibly batched) transform in the autopicker is
always complex-to-real (backwards) only, we want to make a plan
only for te backwards transform, so that we only allocate temporary
space for it and not the forward transform.
- - - - -
e53cbf3b by Sjors Scheres at 2016-04-28T15:34:18+01:00
Shaoda's adaptation of shrink for helices in autopicker
- - - - -
e4538af8 by Sjors Scheres at 2016-04-28T15:34:37+01:00
Shaoda's debug of polishing for helices
- - - - -
249f9b93 by Sjors Scheres at 2016-04-28T15:36:21+01:00
copy data to scratch with explicit prevention of filling up scratch space
- - - - -
eab159b4 by Sjors Scheres at 2016-04-28T15:37:39+01:00
Merged in scheres/relion-devel-lmb (pull request #35)
Master
- - - - -
6eacf397 by Sjors Scheres at 2016-04-28T15:37:52+01:00
Merged tcblab/relion-devel into master
- - - - -
e982295b by Dari Kimanius at 2016-04-28T17:48:50+02:00
The exponentiate_weights_coarse kernel is batched over all classes.
- - - - -
8dc10b89 by Dari Kimanius at 2016-04-28T17:52:25+02:00
The GPU_MEMORY_OVERHEAD_MB subtraction from CA-cache scales with number of threads.
- - - - -
f272115b by Sjors Scheres at 2016-04-28T17:19:27+01:00
Shaoda found another modification for helical autopicker in cuda code
- - - - -
85508146 by Sjors Scheres at 2016-04-29T09:11:17+01:00
because MRC is stored in float, also preread images in float, not RFLOAT
- - - - -
22d3ee9d by Sjors Scheres at 2016-04-29T11:35:37+01:00
made one generic function to generate PDF logfiles
- - - - -
a25ba44e by Sjors Scheres at 2016-04-29T11:36:26+01:00
change psi-sampling in case of gpu to finer one that is multiple of 32
- - - - -
f93cf4c2 by Sjors Scheres at 2016-04-29T11:37:16+01:00
rescale_factor downscaling AND change dimensions automatically for optimal FFTs
- - - - -
c7b9c967 by Sjors Scheres at 2016-04-29T11:37:44+01:00
Merged in scheres/relion-devel-lmb (pull request #36)
Master
- - - - -
2648abf7 by Sjors Scheres at 2016-04-29T11:38:02+01:00
Merged tcblab/relion-devel into master
- - - - -
22674324 by Bjoern Forsberg at 2016-04-29T14:07:06+02:00
Batching encompasses temp, src and dest arrays
This reverts behaviour closer to a previous state, which now makes
the autopicker more stable in memory-limited cases.
- - - - -
fe183447 by Bjoern Forsberg at 2016-04-29T14:57:44+02:00
Autopicker cleanup and nicer warnings
warns if:
shrinking past lowpass
bad FFT-size based on prime factors
batches over psi-sampling
- - - - -
096d13a7 by Bjoern Forsberg at 2016-04-29T14:59:29+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
58f41777 by Dari Kimanius at 2016-04-29T15:54:18+02:00
Timing for projector plan.
- - - - -
5d1b8b3e by Bjoern Forsberg at 2016-04-29T18:17:16+02:00
A centering went missing in the last few commits
- - - - -
f59a9838 by Dari Kimanius at 2016-04-29T18:49:29+02:00
Removed the back and forth transformation between angle and matrix representation of sampling rotations.
30-40% speed-up in AutoRefine.
- - - - -
a967884a by Dari Kimanius at 2016-04-29T18:51:32+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
e176729d by Sjors Scheres at 2016-05-03T13:01:05+01:00
debugged copying to scratch
- - - - -
4db90e91 by Sjors Scheres at 2016-05-03T13:01:31+01:00
Merged in scheres/relion-devel-lmb (pull request #38)
debugged copying to scratch
- - - - -
f47addb8 by Dari Kimanius at 2016-05-03T14:16:40+02:00
Separated preprocessor constant for extraction timing.
- - - - -
1a394a19 by Bjoern Forsberg at 2016-05-03T14:34:48+02:00
temp revert of autopicker shrink control
(for pull-req)
- - - - -
97e64503 by Bjoern Forsberg at 2016-05-03T14:35:28+02:00
Merged tcblab/relion-devel into master
- - - - -
4af35e43 by Bjoern Forsberg at 2016-05-03T14:45:28+02:00
reinstating bug fix to shrink 0<x<1
- - - - -
ea20fe9c by Sjors Scheres at 2016-05-03T13:48:52+01:00
Merged in tcblab/relion-devel-tcblab (pull request #39)
Master
- - - - -
4ce98536 by Sjors Scheres at 2016-05-03T13:49:14+01:00
Merged tcblab/relion-devel into master
- - - - -
74706bdf by Sjors Scheres at 2016-05-03T13:58:32+01:00
minor tweaks to GUI
- - - - -
32f8afcd by Dari Kimanius at 2016-05-03T16:42:18+02:00
Better error-report in coarse-convert for zero number of significants.
- - - - -
33ae8de1 by Dari Kimanius at 2016-05-03T17:03:32+02:00
GPU acceleration of euler-matrix calculation.
- - - - -
1212333c by Dari Kimanius at 2016-05-03T17:04:19+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
03e1614c by Dari Kimanius at 2016-05-03T18:02:58+02:00
Fixed firstiter CC debug warnings.
- - - - -
7ffc813c by Bjoern Forsberg at 2016-05-03T20:09:10+02:00
stabilizing autopicker memory allocations
If one wants to runs insane combinations like 3600 oriennts of
4096x4096 mics with 7 refs, one CAN do it, but even on a K80 the
autopicker will need ~72 batches to do it (and it will be a lot
better at being insane than the the CPU-code).
- - - - -
57612678 by Bjoern Forsberg at 2016-05-03T20:15:34+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
6910e5cd by Sjors Scheres at 2016-05-04T13:47:47+01:00
go back to getImageScratchName in exp_model#
- - - - -
7a3123c1 by Sjors Scheres at 2016-05-04T13:48:20+01:00
fl_message instead of cout
- - - - -
34924260 by Sjors Scheres at 2016-05-04T13:48:52+01:00
go back to providing a movie STAR file as input for movierefine
- - - - -
a5711bbe by Sjors Scheres at 2016-05-04T14:02:31+01:00
replaced errno ddeclarations for err_no for osx compilation
- - - - -
5e01a052 by Sjors Scheres at 2016-05-04T14:14:23+01:00
pass also do ctfimage in copyParticles
- - - - -
424e0115 by Sjors Scheres at 2016-05-04T15:38:22+01:00
remove coords as outputnode from movierefine
- - - - -
ed63f4d4 by Sjors Scheres at 2016-05-04T15:38:37+01:00
siwtch off PP_Timing
- - - - -
3bef0ac1 by Sjors Scheres at 2016-05-04T16:17:53+01:00
synatx change warning
- - - - -
93bc2211 by Sjors Scheres at 2016-05-04T16:18:26+01:00
Merged in scheres/relion-devel-lmb (pull request #40)
Master
- - - - -
b4622f6d by Sjors Scheres at 2016-05-04T16:18:39+01:00
Merged tcblab/relion-devel into master
- - - - -
1f238730 by Sjors Scheres at 2016-05-04T17:20:40+01:00
dont resizebox flowcharts, leave comments for user to take out
- - - - -
309c6e0f by Sjors Scheres at 2016-05-04T17:22:21+01:00
dont resizebox flowcharts, leave comments for user to take out
- - - - -
06139057 by Erik Lindahl at 2016-05-05T22:47:50+02:00
Fix non-CUDA and OS X builds
- Introduced CUDA CMake option that in turn defines -DCUDA for those
compiles, and added #ifdef statements around CUDA calls.
- Fixed OS X linking, library, and detection issues.
- Fixed usage of sincos()/sincosf() with the clang compiler.
- Added correct FFTW libraries to relion_lib.
- Use internal CMake varibles instead of environment variables.
- Made the detection compatible with both lib and lib64 under the
cuda toolkit directory when searching for libraries.
There is still an issue with CUDA OS X builds where clang complains
about __any() being ambiguous. This appears to be due to other
Relion files including a header that in turn uses the entire std
namespace.
- - - - -
59c2eecf by Dari Kimanius at 2016-05-06T10:42:57+02:00
Better out of memory recovery with a large number of retries before error is thrown.
A bit smarter assignement of CA allocation size.
- - - - -
0e8b8a00 by Dari Kimanius at 2016-05-06T10:44:16+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
a89770bd by Dari Kimanius at 2016-05-06T11:25:44+02:00
Check if the new policy exists before setting it.
- - - - -
9f439dc4 by Dari Kimanius at 2016-05-06T11:57:21+02:00
Translations in diff2-coarse-kernel moved to an external kernel-function.
- - - - -
8158484f by Sjors Scheres at 2016-05-06T15:51:41+01:00
put warning of psi change after its done
- - - - -
52e6571d by Sjors Scheres at 2016-05-06T15:52:55+01:00
small changes by Shaoda
- - - - -
0d905a99 by Sjors Scheres at 2016-05-06T16:57:02+01:00
cut large workflows in pieces
- - - - -
7c5edd3a by Erik Lindahl at 2016-05-07T18:13:32+02:00
Fixed OS X FFTW3 and FLTK builds
Fixed library suffixes and detection. This should
resolve the issue with fftw3.h not being found for
local builds.
- - - - -
400a167e by Erik Lindahl at 2016-05-07T18:35:29+02:00
Fixed namespace errors related to CPlot2D
The CPlot2D header previous had a "using namespace std" on
global file scope, which caused a chain effect where every single
file had all routines in the standard namespace. This in turn caused
a clash between internal clang and CUDA symbols. Fixed by properly
prefixing names instead.
- - - - -
6b025e57 by Erik Lindahl at 2016-05-07T18:46:59+02:00
Added std:: prefixes to source code
A few additional files needed std:: prefixes
now that CPlot2D.h no longer pulls in all of the
std namespace.
- - - - -
8302946f by Erik Lindahl at 2016-05-07T19:31:07+02:00
Fix Cuda on OS X
Hide circular dependency errors on OS X to enable Cuda builds.
Added flags to hide warnings about unresolved symbols at link time
due to circular dependencies between gpu_utils and main relion_lib,
and avoid using linux rpath settings on OS X.
- - - - -
77d2c8db by Erik Lindahl at 2016-05-07T21:26:59+02:00
Fix fltk build inconsistency Linux/Mac
Bumped required cmake version to 2.8.12, and use a patch
instead of a custom fltk-1.3.3 version for the OS X issues.
- - - - -
444f1a0d by Erik Lindahl at 2016-05-07T21:44:00+02:00
Removed autoconf files
- - - - -
f78f59a2 by localadmin at 2016-05-09T14:12:40+02:00
Warnw if non-cuda compile tries to use GPUs
- - - - -
9af66057 by localadmin at 2016-05-09T14:20:04+02:00
Warns if cuda-verison is too low
- - - - -
a9bfa093 by Dari Kimanius at 2016-05-09T16:17:54+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
8cbcf6cf by Dari Kimanius at 2016-05-10T10:33:04+02:00
Fixed issue with building without shared libs.
Also some clean-up.
- - - - -
f89f4af5 by localadmin at 2016-05-10T10:54:51+02:00
Reengineered cuda-code timings
Now activated in benchmark build. At present only works
with a single thread and rank, due to filename clashes
to be fixed in next commit.
- - - - -
9f628dc5 by localadmin at 2016-05-10T13:52:38+02:00
Multi-reank /-threaded timing files
under precompile symbol TIMING_FILES, bewae of LARGE output for
realistic runs; averaging not yet in place.
Works for refine and autopicker (mpi-verisons included)
- - - - -
e3f86fb8 by localadmin at 2016-05-10T13:54:59+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
e6904478 by Dari Kimanius at 2016-05-10T14:33:41+02:00
Translations generated pixel-wise on demand within the diff2-fine kernel.
- - - - -
848b2a00 by Dari Kimanius at 2016-05-10T14:47:39+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
Conflicts:
src/gpu_utils/cuda_ml_optimiser.cu
- - - - -
d57bab8d by Dari Kimanius at 2016-05-11T11:10:38+02:00
CC-kernels support in kernel pixel-wise translations.
- - - - -
8741f653 by Bjoern Forsberg at 2016-05-11T12:23:35+02:00
CenterFFT 3D introduced (unused)
- - - - -
647446ed by Dari Kimanius at 2016-05-11T16:30:28+02:00
In kernel pixel-wise translations implemented for wavg and backproject kernels.
- - - - -
14544feb by Bjoern Forsberg at 2016-05-12T15:00:03+02:00
cuda FFTs can be 3D
- - - - -
8a541c81 by Bjoern Forsberg at 2016-05-12T15:43:14+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
653c3f08 by Bjoern Forsberg at 2016-05-12T16:26:42+02:00
Fix to profiling build
- - - - -
0f573929 by Bjoern Forsberg at 2016-05-12T19:57:49+02:00
Another profiling fix (markers in nnvp)
- - - - -
842b29a5 by Dari Kimanius at 2016-05-13T12:24:33+02:00
Clean-up and superficial optimization of wavg-kernel.
- - - - -
7f64df71 by Dari Kimanius at 2016-05-13T13:23:43+02:00
BP-kernel optimization/clean-up.
- - - - -
8a940d67 by Dari Kimanius at 2016-05-13T14:16:34+02:00
Translations clean-up.
Removed cuda-tranlsation-class and the use of it.
- - - - -
e4303857 by Sjors Scheres at 2016-05-13T13:50:25+01:00
raise an error if one of the halfsets is empty for helices
- - - - -
6c416bd5 by Sjors Scheres at 2016-05-13T13:50:54+01:00
cut flowcharts longer than 10 steps
- - - - -
92dfabf4 by Sjors Scheres at 2016-05-13T13:51:31+01:00
new converegnce behaviour in auto-refine
- - - - -
46d9d09e by Sjors Scheres at 2016-05-13T13:51:50+01:00
added joinStar for movies
- - - - -
8d3ace76 by Sjors Scheres at 2016-05-13T14:24:57+01:00
further development of joinSTAR
- - - - -
9985dc68 by Sjors Scheres at 2016-05-13T14:25:26+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
705766f7 by Sjors Scheres at 2016-05-13T15:05:21+01:00
Merged in scheres/relion-devel-lmb (pull request #41)
Master
- - - - -
83680dec by Sjors Scheres at 2016-05-13T15:05:38+01:00
Merged tcblab/relion-devel into master
- - - - -
29b30623 by Dari Kimanius at 2016-05-13T18:36:10+02:00
Merge https://bitbucket.org/tcblab/relion-devel
Conflicts:
src/motioncorr_runner.cpp
- - - - -
c145ad9c by Sjors Scheres at 2016-05-14T09:56:17+01:00
Merged in tcblab/relion-devel-tcblab (pull request #42)
Better build environment and profiling, 30-40% faster autorefine and on demand translations.
- - - - -
487dacb0 by Sjors Scheres at 2016-05-14T09:56:44+01:00
Merged tcblab/relion-devel into master
- - - - -
88ec6443 by Bjoern Forsberg at 2016-05-16T11:43:02+02:00
Modified the automatic device assignment
Now mapps threads more intuitively and in some cases
more benefitially in terms of memory;
-n 4 -j 6 on 4 devices (before)
0 2 1 3 0 2
1 3 0 2 1 3
0 2 1 3 0 2
(each device has 3 ranks)
-n 4 -j 6 on 4 devices (now)
0 0 0 0 0 1
1 1 1 2 2 2
2 3 3 3 3 3
(each device has 2 ranks)
If we preferred, we could also choose to "unbalance" the device-
load to further reduce default memory footprint;
0 0 0 0 0(0)
1 1 1 2 2 2
(3)3 3 3 3 3
(1 device has 2 ranks, both others have 1)
This point is sort of dumb though, since one device is still memory-
limited, and these launch-configs are bad to begin with.
- - - - -
763c69a4 by Sjors Scheres at 2016-05-16T14:00:43+01:00
hopefulyl fixed inconsistent pipeline.star when executing scheduled jobs
- - - - -
15fa87b6 by Sjors Scheres at 2016-05-16T14:01:05+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
56266374 by Dari Kimanius at 2016-05-16T15:51:55+02:00
Memory management enhancements and an allocator bug-fix.
- - - - -
9692aaa0 by Dari Kimanius at 2016-05-16T16:02:27+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
cafcaa29 by Bjoern Forsberg at 2016-05-16T16:15:44+02:00
no lockfile for scratch_dir in non-mpi refine
this fixes issue #102
- - - - -
522e8643 by Bjoern Forsberg at 2016-05-16T16:52:14+02:00
Autpicker fix
An oversight in preliminiary testing caused ctf correction to be
needed, as a GPU-kernel used a ctf-image as a multiplier regard-
less of ctf-correction or not.
this fixes issue #103
- - - - -
31f6bd2c by Bjoern Forsberg at 2016-05-17T09:57:13+02:00
Extra careful about allocation 0-size cudaGlbPtr
- - - - -
d2aacd0f by Bjoern Forsberg at 2016-05-17T09:57:58+02:00
All setDevice() are guarded against bad ordinals
- - - - -
c1d9651a by Sjors Scheres at 2016-05-17T09:23:22+01:00
fixed remaking of logfile for continuationn run
- - - - -
4f017cd0 by Bjoern Forsberg at 2016-05-17T10:54:01+02:00
Fix to debug build
(amends commit 31f6bd2)
- - - - -
51e85e4d by Dari Kimanius at 2016-05-17T11:52:02+02:00
Fix to Issue #105.
- - - - -
1783c54d by Bjoern Forsberg at 2016-05-17T11:59:36+02:00
Additional fix to issue #105
- - - - -
051f6581 by Bjoern Forsberg at 2016-05-17T12:00:00+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
2acd4d01 by Dari Kimanius at 2016-05-17T13:38:11+02:00
Proper application exit for refine and autopick.
Instead of exit() just a return-statement is issued at the end of the main function.
This to make sure all object destrcutors are called.
A good source for this is: http://stackoverflow.com/questions/30250934/how-to-end-c-code
- - - - -
8937748e by Björn Forsberg at 2016-05-17T16:28:51+02:00
Merged tcblab/relion-devel into master
- - - - -
4649e5ed by Sjors Scheres at 2016-05-17T16:30:21+01:00
Merged in tcblab/relion-devel-tcblab (pull request #43)
Several important memory fixes
- - - - -
b070c02b by Sjors Scheres at 2016-05-17T16:30:37+01:00
Merged tcblab/relion-devel into master
- - - - -
c2befc2b by Erato.bjoern at 2016-05-18T18:08:24+02:00
Fix to non-default dev-ids & mem-OH-calc
possibly unnecessary low-OH devicesync added for safety
- - - - -
dca6054b by Bjoern Forsberg at 2016-05-18T18:55:57+02:00
Device is set before reset
- - - - -
0bb1789d by Björn Forsberg at 2016-05-19T08:57:19+02:00
Merged tcblab/relion-devel into master
- - - - -
4b209fb5 by Sjors Scheres at 2016-05-19T08:33:29+01:00
Merged in tcblab/relion-devel-tcblab (pull request #44)
Master
- - - - -
fe49cdb3 by Sjors Scheres at 2016-05-19T08:33:44+01:00
Merged tcblab/relion-devel into master
- - - - -
71decc80 by Sjors Scheres at 2016-05-19T16:16:51+01:00
print usage to stdout, so that error message is more visible in stderr window
- - - - -
18590647 by Sjors Scheres at 2016-05-19T16:17:22+01:00
added scratch_dir to refine jobs
- - - - -
1a29676e by Sjors Scheres at 2016-05-19T16:18:22+01:00
delete non-finished outputnodes when marking a job as finsihed
- - - - -
865c24f3 by Bjoern Forsberg at 2016-05-20T11:28:51+02:00
Explicit indexing actually happens on mpi_refine
- - - - -
74562128 by Bjoern Forsberg at 2016-05-20T11:30:08+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
5148526a by Sjors Scheres at 2016-05-20T10:46:40+01:00
Merged in tcblab/relion-devel-tcblab (pull request #45)
Explicit indexing actually happens on mpi_refine
- - - - -
dd4b29ef by Björn Forsberg at 2016-05-20T11:51:45+02:00
Merged tcblab/relion-devel into master
- - - - -
6cb7aeca by Erik Lindahl at 2016-05-22T22:10:42+02:00
Install gui background xpm to library directory
The relion_maingui program will first look for the file
in the configured library install directory, and then in
the source directory.
- - - - -
724ac140 by Erik Lindahl at 2016-05-23T09:01:53+02:00
Fix locally built FLTK installation
Install to the target library directory, so RELION works
after installation even if the source directory is
removed.
- - - - -
28c40ae9 by Sjors Scheres at 2016-05-23T11:57:13+01:00
allow saving from manualpick window
- - - - -
bee40d76 by Sjors Scheres at 2016-05-23T11:57:31+01:00
back to more stringent auto-refine convergence criteria
- - - - -
237b40bb by Sjors Scheres at 2016-05-23T11:58:19+01:00
Merged in scheres/relion-devel-lmb (pull request #46)
Master
- - - - -
27eabec9 by Dari Kimanius at 2016-05-24T15:58:16+02:00
Merged tcblab/relion-devel into master
- - - - -
f60e7ef9 by Bjoern Forsberg at 2016-05-25T11:56:35+02:00
Fix to batches over psi-orientations in AP
- - - - -
dbbd7b58 by Bjoern Forsberg at 2016-05-25T11:57:27+02:00
Reduces overhead of re-planning FFTs when batching
by checking against the *requested* size last time batched transforms
were requested, instead of the then found possible batchsize.
- - - - -
8bc307d9 by Bjoern Forsberg at 2016-05-25T11:59:49+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
31241154 by Sjors Scheres at 2016-05-26T15:04:04+01:00
merge with Shaoda#
- - - - -
f4e30fa8 by Sjors Scheres at 2016-05-26T15:04:27+01:00
dont warn zero overlap
- - - - -
05fefb0e by Sjors Scheres at 2016-05-26T15:04:47+01:00
write all notes to output file
- - - - -
b02557e0 by Sjors Scheres at 2016-05-26T15:05:15+01:00
main route up comes from re-extracted particles in flowchart
- - - - -
3487890a by Sjors Scheres at 2016-05-26T15:05:35+01:00
more error output
- - - - -
90431dd9 by Sjors Scheres at 2016-05-26T15:06:01+01:00
better help in help-button
- - - - -
4589d458 by Dari Kimanius at 2016-05-26T19:50:00+02:00
Timings in projector.cpp
- - - - -
0b777e7f by Dari Kimanius at 2016-05-26T21:43:26+02:00
Mirrored computeFourierTransformMap from ml_model into cuda-files.
- - - - -
754d372f by Sjors Scheres at 2016-06-08T15:04:27+01:00
Merged in tcblab/relion-devel-tcblab (pull request #47)
Master
- - - - -
55796125 by Sjors Scheres at 2016-06-08T15:04:52+01:00
Merged tcblab/relion-devel into master
- - - - -
8cd9fbda by Sjors Scheres at 2016-06-08T15:22:23+01:00
dont die with errors from unblur execution
- - - - -
f79d22fe by Sjors Scheres at 2016-06-08T15:22:43+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
e6e57ef0 by Sjors Scheres at 2016-06-09T10:03:50+01:00
also use master!
- - - - -
d0f1158d by Bjoern Forsberg at 2016-06-13T12:01:37+02:00
fix (only for small arrays?)
- - - - -
e3bd152b by Bjoern Forsberg at 2016-06-13T15:15:58+02:00
Adaptive OS used in fine
This should not be needed, but if oversampling is not used (i.e. if
there is only a single pass), then data-objects are incompatible with
storeWS in the GPU-version. We escape this by making a redundant calc
in a symbolic fine-pass that is NOT oversampled, but then we need
to use adaptive OS instead of hardcoding an oversampling in the fine
pass.
- - - - -
3768c86a by Sjors Scheres at 2016-06-14T09:27:52+01:00
allow import coordinates from pipeline directories, and dont let on-the-fly processing die with errors, write warnings instead
- - - - -
ef5b1397 by Sjors Scheres at 2016-06-14T12:40:15+01:00
Merged in scheres/relion-devel-lmb (pull request #48)
Master
- - - - -
fdcc4c88 by Sjors Scheres at 2016-06-15T15:15:52+01:00
Shaoda changed some default values for helical autopicking
- - - - -
9bc92f24 by Sjors Scheres at 2016-06-16T10:04:00+01:00
only save x,y coordinates in manualpicking to avoid problems with picking extra particles#
- - - - -
5b850ada by Sjors Scheres at 2016-06-16T10:04:58+01:00
Merged in scheres/relion-devel-lmb (pull request #49)
Master
- - - - -
bf7b9080 by Sjors Scheres at 2016-06-16T10:05:08+01:00
Merged tcblab/relion-devel into master
- - - - -
e50ec109 by Dari Kimanius at 2016-06-16T13:25:54+02:00
Fixed better memory related warning and debug messages
- - - - -
9058ee23 by Bjoern Forsberg at 2016-06-16T14:19:17+02:00
CudaAutopicking uses MPI correctly
- - - - -
f21ef0c7 by Sjors Scheres at 2016-06-16T16:09:10+01:00
small bug in manualpicker and in main gui
- - - - -
5a49ce71 by Sjors Scheres at 2016-06-16T16:09:39+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
760cf4b0 by Sjors Scheres at 2016-06-16T16:12:22+01:00
Merged in scheres/relion-devel-lmb (pull request #50)
small bug in manualpicker and in main gui
- - - - -
0aceea1f by Sjors Scheres at 2016-06-16T16:12:39+01:00
Merged tcblab/relion-devel into master
- - - - -
4772a6d9 by Bjoern Forsberg at 2016-06-16T17:59:31+02:00
fix to issue #116
- - - - -
a381f141 by Bjoern Forsberg at 2016-06-16T18:00:24+02:00
fix valgrind error
- - - - -
693a59ef by Bjoern Forsberg at 2016-06-16T18:04:30+02:00
Fixes under some flags
--skip_align
--skip_rotate
These flags do not work perfectly yet, but this commit does
partially address issue #113
- - - - -
71968f04 by Bjoern Forsberg at 2016-06-16T18:07:07+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
1e607c86 by Bjoern Forsberg at 2016-06-16T18:17:49+02:00
temp non-working state for merge
- - - - -
8a80a707 by Björn Forsberg at 2016-06-16T18:18:35+02:00
Merged tcblab/relion-devel into master
- - - - -
1c46878e by Bjoern Forsberg at 2016-06-16T18:18:48+02:00
Put back working Autopicker in GPU-MPI
- - - - -
9986679b by Bjoern Forsberg at 2016-06-16T18:19:37+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
97bf428a by Bjoern Forsberg at 2016-06-16T18:51:45+02:00
include
- - - - -
9ef10182 by Sjors Scheres at 2016-06-17T14:43:54+01:00
provide random seed for autopicking
- - - - -
6af1b0c9 by Sjors Scheres at 2016-06-17T15:04:11+01:00
Merged in scheres/relion-devel-lmb (pull request #52)
provide random seed for autopicking
- - - - -
6b4552ef by Sjors Scheres at 2016-06-17T15:04:25+01:00
Merged tcblab/relion-devel into master
- - - - -
6f73c1d1 by Björn Forsberg at 2016-06-17T19:06:38+02:00
Merged tcblab/relion-devel into master
- - - - -
62a9b703 by Sjors Scheres at 2016-06-17T18:15:30+01:00
Merged in tcblab/relion-devel-tcblab (pull request #51)
Master
- - - - -
f7d98788 by Björn Forsberg at 2016-06-17T19:18:28+02:00
Merged tcblab/relion-devel into master
- - - - -
07e9f02f by Bjoern Forsberg at 2016-06-21T10:57:52+02:00
fix to fedora-like linktime requirements
- - - - -
50d7e5ed by Bjoern Forsberg at 2016-06-21T11:32:47+02:00
fix compilation error
- - - - -
c80642be by Bjoern Forsberg at 2016-06-21T11:33:12+02:00
A bunch of single/double compatibilities
- - - - -
3a1a6e0c by Bjoern Forsberg at 2016-06-21T11:36:07+02:00
minor build-time print fix
- - - - -
1070902d by Bjoern Forsberg at 2016-06-21T11:36:34+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
9a6cbe60 by Bjoern Forsberg at 2016-06-21T12:27:28+02:00
Fix to recent change in sum_weight initalization
- - - - -
fac5670a by Bjoern Forsberg at 2016-06-22T14:08:55+02:00
CC-kernels (coarse less smem-dependent)
This adresses a reported issue in the relion2-beta and is effectually
a response to the original implementation which hadi these deficiencies
noted (see commit 6c2a43a).
- - - - -
111aa805 by Erik Lindahl at 2016-06-22T22:33:19+02:00
Enable SSE/AVX when building FFTW ourselves
Detect if the architecture is x86, and enable both SSE/SSE2
as well as AVX if it is. FFTW will time routines and pick
the best supported alternative at runtime.
- - - - -
b8ff6568 by Dari Kimanius at 2016-06-27T10:09:21+02:00
Better handling of the different build cases: no-GUI, no-CUDA
Fixes beta-repo issue nr 12
- - - - -
2007cccc by Bjoern Forsberg at 2016-06-27T10:24:05+02:00
Fix to ctf-values being zero
- - - - -
5fba995e by Bjoern Forsberg at 2016-06-27T10:24:47+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
29886201 by Bjoern Forsberg at 2016-06-27T12:03:37+02:00
Fixed compile-time warnings, mostly GUI
some non-returns in std::sys -calls
- - - - -
e5db9cc5 by Bjoern Forsberg at 2016-06-28T14:01:39+02:00
Scripts are installed with the right permissions.
- - - - -
a48bdf54 by Bjoern Forsberg at 2016-06-30T17:53:49+02:00
Adding 3D-kernels use in getFTandCtfs
- - - - -
be83a51f by Bjoern Forsberg at 2016-06-30T17:54:31+02:00
Fixes to movie-refinement
- - - - -
31da19a4 by Sjors Scheres at 2016-07-01T11:36:43+01:00
Merged tcblab/relion-devel into master
- - - - -
ecf74998 by Sjors Scheres at 2016-07-01T11:59:43+01:00
Merged in tcblab/relion-devel-tcblab (pull request #53)
Master
- - - - -
6f96227a by Sjors Scheres at 2016-07-01T12:06:27+01:00
Merged tcblab/relion-devel into master
- - - - -
05b46b49 by Sjors Scheres at 2016-07-01T12:10:52+01:00
allow helical_outer_diameter change in continuation jobs and set phase-shift in ml_optimiser.cpp and .cu
- - - - -
7a44b582 by Sjors Scheres at 2016-07-01T12:11:30+01:00
dont set pool to 1 when prereading images
- - - - -
ba8ee84e by Sjors Scheres at 2016-07-01T12:12:01+01:00
resolve issue of hanging GUI when deleting job without output nodes
- - - - -
dbcf2fce by Sjors Scheres at 2016-07-01T12:12:45+01:00
Merged in scheres/relion-devel-lmb (pull request #54)
Master
- - - - -
27943e42 by Bjoern Forsberg at 2016-07-01T13:14:44+02:00
Merged tcblab/relion-devel into master
- - - - -
274e389a by Sjors Scheres at 2016-07-01T12:18:55+01:00
repairing typos...
- - - - -
be518603 by Sjors Scheres at 2016-07-01T12:19:43+01:00
Merged in scheres/relion-devel-lmb (pull request #55)
repairing typos...
- - - - -
22a0b8df by Bjoern Forsberg at 2016-07-01T13:20:30+02:00
Merged tcblab/relion-devel into master
- - - - -
6342123f by Sjors Scheres at 2016-07-01T12:24:40+01:00
Merged tcblab/relion-devel into master
- - - - -
5ba45764 by Sjors Scheres at 2016-07-01T13:52:38+01:00
dont die GUI if fields are empty, raise warnings instead; fixes 2.0-beta issue #34
- - - - -
d085c5ae by Sjors Scheres at 2016-07-01T13:53:01+01:00
Merged in scheres/relion-devel-lmb (pull request #56)
dont die GUI if fields are empty, raise warnings instead; fixes 2.0-beta issue #34
- - - - -
fffebc5f by Sjors Scheres at 2016-07-01T13:53:23+01:00
Merged tcblab/relion-devel into master
- - - - -
9f2e3515 by Sjors Scheres at 2016-07-01T14:48:11+01:00
resolving issue #36: setting angpix correctly from STAR file in manualpicker
- - - - -
dca38edf by Sjors Scheres at 2016-07-01T14:49:11+01:00
Merged in scheres/relion-devel-lmb (pull request #57)
resolving issue #36: setting angpix correctly from STAR file in manualpicker
- - - - -
b63410c5 by Sjors Scheres at 2016-07-01T14:49:23+01:00
Merged tcblab/relion-devel into master
- - - - -
8621d09b by Dari Kimanius at 2016-07-01T16:26:16+02:00
Massive commit for added support for helical refinement.
Warnings removes for non-support for gpu runs on helical refinement.
Slicker code in and around precalculateShiftedImagesCtfsAndInvSigma2s.
- - - - -
3ea5a860 by Dari Kimanius at 2016-07-01T16:29:26+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
24098756 by Dari Kimanius at 2016-07-01T16:30:26+02:00
Merged tcblab/relion-devel into master
- - - - -
66f206f5 by Bjoern Forsberg at 2016-07-01T17:19:36+02:00
Multinode jobs get devices set based on host-share
- - - - -
3f668448 by Bjoern Forsberg at 2016-07-01T17:20:52+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
cf8ab0ab by Bjoern Forsberg at 2016-07-05T09:22:22+02:00
fix to single/double choice.
In this case, even single prec. was less than single. This fixes
issue #114 and possibly also issue #113
- - - - -
42801af9 by Bjoern Forsberg at 2016-07-05T09:24:02+02:00
Addintional single/double fix
(not symptomatic, but could be in future, so fixed as precaution)
- - - - -
65030586 by Sjors Scheres at 2016-07-05T09:35:41+01:00
Merged in tcblab/relion-devel-tcblab (pull request #58)
Added support for helical refinement.
- - - - -
cdde80e9 by Sjors Scheres at 2016-07-05T09:35:54+01:00
Merged tcblab/relion-devel into master
- - - - -
6a43823a by Bjoern Forsberg at 2016-07-05T19:12:54+02:00
Merged tcblab/relion-devel into master
- - - - -
587758c3 by Dari Kimanius at 2016-07-06T16:19:11+02:00
Minor cmake related cleanup and bugfixes.
- - - - -
12d97136 by Erato.bjoern at 2016-07-06T21:08:09+02:00
fix to device id MPI_send in MPI refine
- - - - -
88e9bb57 by Sjors Scheres at 2016-07-07T09:19:08+01:00
Merged in tcblab/relion-devel-tcblab (pull request #59)
Master
- - - - -
a0218a93 by Sjors Scheres at 2016-07-07T09:20:26+01:00
Merged tcblab/relion-devel into master
- - - - -
0096315f by Sjors Scheres at 2016-07-07T09:44:15+01:00
resolving issue 18 in relion-2-beta
- - - - -
b8c541c1 by Sjors Scheres at 2016-07-07T09:45:01+01:00
Merged in scheres/relion-devel-lmb (pull request #60)
resolving issue 18 in relion-2-beta
- - - - -
5169b0a3 by Sjors Scheres at 2016-07-07T09:45:16+01:00
Merged tcblab/relion-devel into master
- - - - -
25c28f7b by Bjoern Forsberg at 2016-07-07T11:07:06+02:00
Merged tcblab/relion-devel into master
- - - - -
b77837d1 by Bjoern Forsberg at 2016-07-07T11:09:24+02:00
movie-compatiblity amendments
related to issue #119, and solves major symptoms, but may have
lingering issues.
- - - - -
4d45beb2 by Bjoern Forsberg at 2016-07-07T11:10:45+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
b08f2b6a by Sjors Scheres at 2016-07-07T10:20:39+01:00
new io of ctffind4.1 according to Alexis Rohou
- - - - -
89cce59b by Sjors Scheres at 2016-07-07T10:21:30+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
f4d5c0ff by Sjors Scheres at 2016-07-07T10:21:57+01:00
Merged in scheres/relion-devel-lmb (pull request #61)
new io of ctffind4.1 according to Alexis Rohou
- - - - -
34b885a5 by Sjors Scheres at 2016-07-07T11:39:40+01:00
repaired bug manualpicking reported by Rafa Leiro at LMB
- - - - -
d5e71a17 by Sjors Scheres at 2016-07-07T11:40:03+01:00
Merged tcblab/relion-devel into master
- - - - -
3a5f8653 by Sjors Scheres at 2016-07-07T11:40:22+01:00
Merged in scheres/relion-devel-lmb (pull request #62)
repaired bug manualpicking reported by Rafa Leiro at LMB
- - - - -
68f6c7ae by Sjors Scheres at 2016-07-07T11:40:44+01:00
Merged tcblab/relion-devel into master
- - - - -
b979959b by Sjors Scheres at 2016-07-08T10:42:18+01:00
resolving issue 15 of relion-2.0-beta
- - - - -
fa2751d0 by Sjors Scheres at 2016-07-08T10:42:55+01:00
Merged in scheres/relion-devel-lmb (pull request #63)
resolving issue 15 of relion-2.0-beta
- - - - -
53b54113 by Sjors Scheres at 2016-07-08T10:43:08+01:00
Merged tcblab/relion-devel into master
- - - - -
70ac67d6 by Sjors Scheres at 2016-07-08T11:16:45+01:00
also have multiple subdirectories (for subtomogram averaging) treated OK
- - - - -
29c3e1f9 by Sjors Scheres at 2016-07-08T11:17:59+01:00
Merged in scheres/relion-devel-lmb (pull request #64)
also have multiple subdirectories (for subtomogram averaging) treated OK
- - - - -
56d2f818 by Sjors Scheres at 2016-07-08T11:18:49+01:00
Merged tcblab/relion-devel into master
- - - - -
186ffad6 by Sjors Scheres at 2016-07-12T10:01:24+01:00
adapted for latest ctffind4.1.3 from Alexis
- - - - -
2f53e248 by Sjors Scheres at 2016-07-12T10:04:03+01:00
better io for errors with helical search ranges
- - - - -
d7187f6e by Sjors Scheres at 2016-07-12T10:04:57+01:00
Merged in scheres/relion-devel-lmb (pull request #65)
Master
- - - - -
f3ad6d45 by Sjors Scheres at 2016-07-12T10:54:36+01:00
better indentation
- - - - -
c580bfab by Dari Kimanius at 2016-07-12T14:19:57+02:00
Improving accuracy by moving the exponentiated diff2-values towards the center of the single precision numerical range.
This is done by permitting resulting values above one, i.e. positive exponents.
- - - - -
f6d6a8eb by Dari Kimanius at 2016-07-12T14:25:07+02:00
Merged tcblab/relion-devel into master
- - - - -
6e9da1c1 by Sjors Scheres at 2016-07-12T13:43:34+01:00
Merged in tcblab/relion-devel-tcblab (pull request #66)
Master
- - - - -
5ce48150 by Sjors Scheres at 2016-07-12T13:43:54+01:00
Merged tcblab/relion-devel into master
- - - - -
bbcbd3f8 by Sjors Scheres at 2016-07-14T14:17:09+01:00
carry on pitch and tube id for helical segment extraction
- - - - -
c31ca860 by Sjors Scheres at 2016-07-14T14:18:15+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
a863ac04 by Bjoern Forsberg at 2016-07-18T10:20:33+02:00
max_weight is based on coarse/fine weights
This amends/fixes issue #119:
max weight was occationally larger than the sum of weights, which
was algorithmically ok but resulted in meta-data probabilites
larger than 1. We now raise segfault if this happens, and collect
max_weight(s) and sum_weight(s) from the same array of weights,
i.e. it should never happen.
Also, max_weights are stored for all iparts since they are now
fetched during convert, and subsequently used in storeWS.
- - - - -
0c7770f3 by Bjoern Forsberg at 2016-07-18T10:26:02+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
Conflicts:
src/gpu_utils/cuda_ml_optimiser.h
- - - - -
68ea4ae3 by Sjors Scheres at 2016-07-18T09:54:13+01:00
added local resolution option in relion
- - - - -
2540fe9f by Sjors Scheres at 2016-07-18T10:06:22+01:00
small changes in help boxes
- - - - -
a98cecf7 by Sjors Scheres at 2016-07-18T10:08:12+01:00
allow threads for resmap
- - - - -
6eea0884 by Sjors Scheres at 2016-07-18T16:51:12+01:00
small changes in locres estimation
- - - - -
125957e1 by Sjors Scheres at 2016-07-18T16:59:32+01:00
small changes in locres estimation 2
- - - - -
d9880c85 by Bjoern Forsberg at 2016-07-19T10:27:19+02:00
min_diff is only updated when fine is meaningful
This fixes the remaining discrepancies in issue #119
Also, LogLikeli is correct using avg_diff subtraction, by changing
definition of dLL accordingly. NOTE: sum_weight is still uncorrected,
so there appears to still be "hidden" discrepancies in LogLikeli
which are completely overshadowed by the diff- and noise-terms.
Also, I changed mean_diff -> avg diff, because having both min_diff
and mean_diff is just confusing.
- - - - -
96181c52 by Bjoern Forsberg at 2016-07-19T10:41:52+02:00
Merged tcblab/relion-devel into master
- - - - -
d8108066 by Sjors Scheres at 2016-07-19T09:48:19+01:00
Merged in tcblab/relion-devel-tcblab (pull request #67)
Movie-refinement works and diff-subtraction corrected
- - - - -
a9bc027a by Sjors Scheres at 2016-07-19T09:49:24+01:00
Merged tcblab/relion-devel into master
- - - - -
605f5e4d by Sjors Scheres at 2016-07-19T12:32:54+01:00
removed options of relion locres
- - - - -
a5f8b011 by Sjors Scheres at 2016-07-19T12:33:12+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
d3526254 by Sjors Scheres at 2016-07-19T12:34:58+01:00
defaults to 25A in locres
- - - - -
8d585dae by Bjoern Forsberg at 2016-07-19T13:59:56+02:00
mpi includes are included correctly and sys-indep
related to a FIXME from way back when: 290e722
- - - - -
4da3e2c4 by Bjoern Forsberg at 2016-07-19T14:01:30+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
72f493a3 by Sjors Scheres at 2016-07-19T14:21:22+01:00
added helical to symmetry expand
- - - - -
10f5a0b2 by Sjors Scheres at 2016-07-19T14:22:48+01:00
added option to remove NaNs
- - - - -
8197df5c by Sjors Scheres at 2016-07-20T11:08:18+01:00
dont try to read fn_in as a STAR file if it isnt a STAR file
- - - - -
6e5b04ba by Sjors Scheres at 2016-07-20T12:43:46+01:00
import/export blocks of scheduled jobs
- - - - -
fe5971f4 by Dari Kimanius at 2016-07-21T11:54:22+02:00
Fixes Issue #117
- - - - -
937b8a4b by Dari Kimanius at 2016-07-21T16:31:56+02:00
Average square diff (avg_diff2) is set in the fine pass as well.
- - - - -
cf9430d0 by Bjoern Forsberg at 2016-07-21T17:07:10+02:00
mixed multi-rank (non)default gpu sel.syntax works
- - - - -
2157871d by Bjoern Forsberg at 2016-07-22T10:19:07+02:00
Fix to runtime installation path
librelion was installed, but the installed binaries did not reference it.
- - - - -
89f0e1f1 by Sjors Scheres at 2016-07-22T10:38:25+01:00
Merged in tcblab/relion-devel-tcblab (pull request #68)
Further improved numerical accuracy
- - - - -
a4694545 by Sjors Scheres at 2016-07-22T10:38:41+01:00
Merged tcblab/relion-devel into master
- - - - -
29826fd3 by Dari Kimanius at 2016-07-25T13:54:40+02:00
Clarified warning message.
- - - - -
b592231e by Dari Kimanius at 2016-07-25T18:04:56+02:00
Added a flag for running asymmetrical padding factor for projector and backprojector models.
If the hidden flag --asymmetric_padding is specified the BP padding is +1 that of the projector.
The purpose of this is to be used in combination with --pad 1 to reduce the GPU memory requirenments.
- - - - -
d75a37b7 by Bjoern Forsberg at 2016-07-26T17:47:12+02:00
draft subset flag
- - - - -
5e2b4a32 by Sjors Scheres at 2016-07-26T17:18:25+01:00
added extra_gctf_args
- - - - -
552bea95 by Sjors Scheres at 2016-07-26T17:29:52+01:00
Merged in scheres/relion-devel-lmb (pull request #69)
Local-resolution in RELION, import/export blocks of jobs, and some fixes
- - - - -
44a08692 by Sjors Scheres at 2016-07-26T17:30:10+01:00
Merged tcblab/relion-devel into master
- - - - -
6c0cffee by Bjoern Forsberg at 2016-07-27T09:11:41+02:00
Runtime cuda compute capability capacity control
- - - - -
125857f9 by Bjoern Forsberg at 2016-07-27T09:12:33+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
a734a0f8 by Bjoern Forsberg at 2016-07-27T09:54:59+02:00
Merge branch 'subset'
- - - - -
b95e8d04 by Bjoern Forsberg at 2016-07-27T10:08:50+02:00
Fix to subset during split random halves
- - - - -
9b304896 by Dari Kimanius at 2016-07-27T11:43:04+02:00
New hidden flag "maximum_significants" than can be used to set the maximum allowed significant weights in the coarse pass.
- - - - -
408c368b by Dari Kimanius at 2016-07-27T11:44:22+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
e563f1a4 by Bjoern Forsberg at 2016-07-27T11:55:43+02:00
GPU sel. for exactly one rank used by all ranks
before on 2-gpu system:
mpirun -n 3 -j 2 --gpu 0
slave 1, thread 0 -> device 0
slave 1, thread 1 -> device 0
slave 2, thread 0 -> device 1
slave 2, thread 1 -> device 1
mpirun -n 3 -j 2 --gpu 0:0
slave 1, thread 0 -> device 0
slave 1, thread 1 -> device 0
slave 2, thread 0 -> device 0
slave 2, thread 1 -> device 0
new behaviour:
mpirun -n 3 -j 2 --gpu 0
slave 1, thread 0 -> device 0
slave 1, thread 1 -> device 0
slave 2, thread 0 -> device 0
slave 2, thread 1 -> device 0
- - - - -
66515d09 by Bjoern Forsberg at 2016-07-27T11:59:01+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
bc884c31 by Dari Kimanius at 2016-07-27T12:18:05+02:00
Diffrentiating hidden from invalid flags in output.
- - - - -
061d6af9 by Bjoern Forsberg at 2016-07-27T14:58:08+02:00
Merged tcblab/relion-devel into master
- - - - -
0fd1b1b9 by Sjors Scheres at 2016-07-27T15:31:14+01:00
Merged in tcblab/relion-devel-tcblab (pull request #70)
Feature bundle
- - - - -
0aa963f5 by Sjors Scheres at 2016-07-27T15:31:37+01:00
Merged tcblab/relion-devel into master
- - - - -
37be342e by Sjors Scheres at 2016-07-27T18:01:39+01:00
added multiple schedulers, and their parallel execution to pipeline
- - - - -
4fbb9586 by Sjors Scheres at 2016-07-27T18:02:38+01:00
added option to apply fourier-mask to projector (debug3) and added functionality to skip gridding in the reconstruction
- - - - -
6e2681d7 by Sjors Scheres at 2016-07-27T18:10:52+01:00
Merged in scheres/relion-devel-lmb (pull request #71)
Master
- - - - -
46153760 by Bjoern Forsberg at 2016-07-27T21:23:29+02:00
Fix to subset progress bar
- - - - -
4337d7f4 by Sjors Scheres at 2016-07-28T10:24:40+01:00
repaired bug with continuing refine jobs
- - - - -
ac920664 by Sjors Scheres at 2016-07-28T10:41:32+01:00
forgot to add skip_gridding to copy
- - - - -
5c1fb017 by Sjors Scheres at 2016-07-28T10:43:23+01:00
removed some spurious attempts to auto-uopdate window
- - - - -
234b9603 by Sjors Scheres at 2016-07-28T10:43:43+01:00
Merged tcblab/relion-devel into master
- - - - -
b259b68b by Sjors Scheres at 2016-07-28T10:44:20+01:00
Merged in scheres/relion-devel-lmb (pull request #72)
Master
- - - - -
47da6231 by Sjors Scheres at 2016-07-28T10:44:34+01:00
Merged tcblab/relion-devel into master
- - - - -
7139743e by Sjors Scheres at 2016-07-28T11:13:29+01:00
added functionality to automatically update stdout and joblists every 2 seconds
- - - - -
6ad0a6dc by Sjors Scheres at 2016-07-28T12:25:09+01:00
go back to old position in joblists scrollbars after autoupdating
- - - - -
b89fef97 by Dari Kimanius at 2016-07-28T17:11:14+02:00
Merged tcblab/relion-devel into master
- - - - -
cbe4724f by Bjoern Forsberg at 2016-07-28T17:15:33+02:00
Editable filename in displayer + default rev. sort
- - - - -
e585684b by Bjoern Forsberg at 2016-07-28T17:16:32+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
25c0dae6 by Dari Kimanius at 2016-07-28T17:55:59+02:00
The elegant (less fun) way of handling hidden flags.
- - - - -
b9a8878e by Dari Kimanius at 2016-07-28T17:57:20+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
31570412 by Bjoern Forsberg at 2016-07-29T12:56:45+02:00
Proposed defocus displa in manualpick display
- - - - -
f13ed309 by Bjoern Forsberg at 2016-07-29T12:57:35+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
38bbafed by Bjoern Forsberg at 2016-07-29T13:49:54+02:00
minor additional fixes to manualpick display
- - - - -
d534569a by Sjors Scheres at 2016-08-01T12:09:18+01:00
remove error message with GPU found
- - - - -
2d823806 by Sjors Scheres at 2016-08-01T12:09:41+01:00
control refresh rate and idle-die time of main GUI
- - - - -
e408ad56 by Sjors Scheres at 2016-08-01T12:10:12+01:00
Merged in scheres/relion-devel-lmb (pull request #73)
Master
- - - - -
cb1db5ae by Sjors Scheres at 2016-08-01T12:10:57+01:00
Merged tcblab/relion-devel into master
- - - - -
5f7a0bb7 by Sjors Scheres at 2016-08-01T14:25:45+01:00
repaired bug deleting last job
- - - - -
84dbf6bb by Erato.bjoern at 2016-08-02T07:56:05+02:00
Revert "Editable filename in displayer + default rev. sort"
This reverts commit cbe4724f1568f7d46ed4546cc86954f960029fbd.
- - - - -
9d3417b0 by Bjoern Forsberg at 2016-08-02T08:44:41+02:00
Merged tcblab/relion-devel into master
- - - - -
a2fd216d by Dari Kimanius at 2016-08-03T10:15:44+02:00
Removed legacy code in getFourierTransformsAndCtfs.
One transformer is enough, removed the other one and reuseing the first one.
Device init of fouriers and reals in transformers are not required, also removed.
- - - - -
8ba706fd by Dari Kimanius at 2016-08-03T10:23:33+02:00
Change flag name from maximum_significants to maxsig, for simplicity
- - - - -
ae268c12 by Sjors Scheres at 2016-08-03T10:04:13+01:00
Merged in tcblab/relion-devel-tcblab (pull request #74)
suggested version v2.0.b11 in beta repo
- - - - -
ce514556 by Sjors Scheres at 2016-08-03T12:17:39+01:00
Merged tcblab/relion-devel into master
- - - - -
f12d5cd9 by Bjoern Forsberg at 2016-08-03T17:57:31+02:00
Auto-refine does not use subset
This avoids issues with the subset-option which unfortunately broke
auto-refinement in commit d75a37b.
- - - - -
accb2bed by Dari Kimanius at 2016-08-03T18:17:40+02:00
fail-safe mode introduced to handle "filtered-size-zero" error and recover from it.
- - - - -
1d87c1f3 by Dari Kimanius at 2016-08-03T18:18:56+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
f7123d84 by Bjoern Forsberg at 2016-08-03T18:33:03+02:00
dditional fix progress_bar mpi subset stuff
- - - - -
6aa0fdc2 by Bjoern Forsberg at 2016-08-03T18:33:25+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
c5924838 by Bjoern Forsberg at 2016-08-03T18:55:56+02:00
Merged tcblab/relion-devel into master
- - - - -
b28fb498 by Erato.bjoern at 2016-08-03T20:29:48+02:00
inactiveate randomisation during subset selection
- - - - -
29b4817e by Sjors Scheres at 2016-08-04T11:33:59+01:00
allow float-values for padding_factor in projectors
- - - - -
f00ae3a9 by Sjors Scheres at 2016-08-04T11:34:32+01:00
repaired small bug reading when setting alias
- - - - -
b09c05b5 by Sjors Scheres at 2016-08-04T11:38:23+01:00
added options for flipX and flipY
- - - - -
03fd26b1 by Sjors Scheres at 2016-08-04T11:38:50+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
d6a865da by Sjors Scheres at 2016-08-04T11:42:58+01:00
Merged in scheres/relion-devel-lmb (pull request #76)
Master
- - - - -
60b3d48b by Sjors Scheres at 2016-08-04T12:25:00+01:00
added comment why closing isnt necessary
- - - - -
35112f4a by Bjoern Forsberg at 2016-08-04T13:52:28+02:00
minor revert, no functional changes
- - - - -
ff01b592 by Sjors Scheres at 2016-08-04T13:07:33+01:00
Merged in tcblab/relion-devel-tcblab (pull request #75)
Auto-refine bugfix + refine failsafe mode
- - - - -
803b3898 by Sjors Scheres at 2016-08-04T13:07:47+01:00
Merged tcblab/relion-devel into master
- - - - -
e3b83e63 by Sjors Scheres at 2016-08-04T13:08:09+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
3b357c88 by Bjoern Forsberg at 2016-08-04T14:08:43+02:00
Merged tcblab/relion-devel into master
- - - - -
146f3807 by Sjors Scheres at 2016-09-02T09:31:57+01:00
dont die if header wsa set incorrectly and max < min
- - - - -
3e681abc by Sjors Scheres at 2016-09-02T09:49:03+01:00
repaired bug not updating particle count first micrograph
- - - - -
7de18b11 by Sjors Scheres at 2016-09-02T15:52:36+01:00
resolve bug movie-processing reported by Shaoda
- - - - -
80c38203 by Sjors Scheres at 2016-09-02T17:40:48+01:00
specify that relion GUI has died
- - - - -
be4e3453 by Sjors Scheres at 2016-09-02T17:42:09+01:00
mention bug non-unique directories in browse button; added motioncor2 wrapper to GUI; added movie_name to refine-continue for Shaodas movie refinement bug
- - - - -
8eaa185d by Sjors Scheres at 2016-09-02T17:42:30+01:00
added motioncor2 wrapper
- - - - -
7e676fd5 by Sjors Scheres at 2016-09-02T17:43:13+01:00
Merged in scheres/relion-devel-lmb (pull request #77)
Master
- - - - -
0ecf8dc1 by Sjors Scheres at 2016-09-02T17:43:29+01:00
Merged tcblab/relion-devel into master
- - - - -
3550701e by Sjors Scheres at 2016-09-05T10:05:58+01:00
dont allow highres_exp limit to change upon continuation
- - - - -
693fc6ca by Sjors Scheres at 2016-09-05T10:06:23+01:00
repair bug in cancelling job and .lock file reported by Shaoda
- - - - -
5eb9e82a by Sjors Scheres at 2016-09-05T10:06:43+01:00
added motioncor2 to mpi version of run_motioncorr
- - - - -
4ba2b7e5 by Sjors Scheres at 2016-09-06T10:40:23+01:00
added option to autopick with gaussian-blob reference
- - - - -
a7a68525 by Sjors Scheres at 2016-09-06T10:40:40+01:00
run motioncor2 three times in case it fails
- - - - -
2522d388 by Sjors Scheres at 2016-09-06T10:41:05+01:00
Merged in scheres/relion-devel-lmb (pull request #78)
Master
- - - - -
f54bd53e by Sjors Scheres at 2016-09-06T10:41:17+01:00
Merged tcblab/relion-devel into master
- - - - -
13efa09b by Dari Kimanius at 2016-09-09T10:26:13+02:00
Merged tcblab/relion-devel into master
- - - - -
20146fc1 by Sjors Scheres at 2016-09-09T09:34:48+01:00
allow changing gpu ids upon continuation motioncorr
- - - - -
1e89e238 by Sjors Scheres at 2016-09-09T09:35:06+01:00
only move after successful motioncor2
- - - - -
c5b18326 by Sjors Scheres at 2016-09-09T09:35:34+01:00
Merged in scheres/relion-devel-lmb (pull request #79)
Master
- - - - -
54174cd0 by Sjors Scheres at 2016-09-09T09:35:51+01:00
Merged tcblab/relion-devel into master
- - - - -
0973cff0 by Sjors Scheres at 2016-09-09T09:38:04+01:00
only move after successful motioncor2 partb
- - - - -
a02157ff by Sjors Scheres at 2016-09-09T09:38:59+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
c3586ba2 by Sjors Scheres at 2016-09-09T09:39:29+01:00
Merged in scheres/relion-devel-lmb (pull request #80)
only move after successful motioncor2 partb
- - - - -
f64e993a by Sjors Scheres at 2016-09-09T09:39:39+01:00
Merged tcblab/relion-devel into master
- - - - -
2dee3145 by Dari Kimanius at 2016-09-09T10:40:36+02:00
Fixed compilation error
- - - - -
65762a08 by Sjors Scheres at 2016-09-09T12:48:16+01:00
trying to resolve conflict with Dari
- - - - -
027d0275 by Sjors Scheres at 2016-09-09T12:48:47+01:00
Merged in scheres/relion-devel-lmb (pull request #82)
trying to resolve conflict with Dari
- - - - -
6c0db491 by Sjors Scheres at 2016-09-09T12:49:40+01:00
Merged in tcblab/relion-devel-tcblab (pull request #81)
Fixed compilation error
- - - - -
ba4c8a7c by Sjors Scheres at 2016-09-09T12:49:59+01:00
Merged tcblab/relion-devel into master
- - - - -
c57436ea by Sjors Scheres at 2016-09-09T13:06:46+01:00
remove 3 attempts motioncor2 as it no longer is needed
- - - - -
6b4341d7 by Sjors Scheres at 2016-09-09T13:07:16+01:00
Merged in scheres/relion-devel-lmb (pull request #83)
remove 3 attempts motioncor2 as it no longer is needed
- - - - -
2b1fd9fd by Sjors Scheres at 2016-09-09T13:07:30+01:00
Merged tcblab/relion-devel into master
- - - - -
c17c14d8 by Dari Kimanius at 2016-09-09T14:37:17+02:00
Merged tcblab/relion-devel into master
- - - - -
c7545316 by Sjors Scheres at 2016-09-12T09:21:48+01:00
repaired bug lowpassfilter non-squared micrographs
- - - - -
3f0c13b9 by Sjors Scheres at 2016-09-12T09:22:11+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
b35e5432 by Sjors Scheres at 2016-09-12T09:32:46+01:00
changed default motioncor2 executable
- - - - -
2b01e4c6 by Sjors Scheres at 2016-09-13T14:29:01+01:00
repaired bug non-squared images lowpassfiltermap
- - - - -
acb7aa09 by Sjors Scheres at 2016-09-13T14:29:42+01:00
trying to debug scheduler with rafa
- - - - -
fd2d488e by Sjors Scheres at 2016-09-13T14:30:05+01:00
check at end whether all output files are there or not
- - - - -
5fc94665 by Sjors Scheres at 2016-09-16T12:47:07+01:00
reapired bug in progress bar MPI picking on GPU
- - - - -
5d104b1d by Sjors Scheres at 2016-09-16T12:47:50+01:00
added citations unblur motioncor2
- - - - -
63fecbfa by Sjors Scheres at 2016-09-16T12:48:13+01:00
more decsription for deletion lock file
- - - - -
e1182707 by Sjors Scheres at 2016-09-16T12:48:37+01:00
repaired bug motioncor2 wrapper
- - - - -
1cfc0690 by Sjors Scheres at 2016-09-22T11:38:21+01:00
add option to reuse_scratch
- - - - -
bc84a352 by Sjors Scheres at 2016-09-26T13:22:25+01:00
bug passing pixelsize to motioncorr2
- - - - -
5582e8f6 by Sjors Scheres at 2016-09-26T13:28:08+01:00
Merged in scheres/relion-devel-lmb (pull request #84)
Master
- - - - -
5a737ab1 by Sjors Scheres at 2016-09-26T14:31:31+01:00
resolves issue 110
- - - - -
0da468fb by Sjors Scheres at 2016-09-26T14:31:54+01:00
resolves issue 111
- - - - -
9c8f917d by Sjors Scheres at 2016-09-27T17:21:39+01:00
added time to logfile of scheduler
- - - - -
cfcc40f7 by Sjors Scheres at 2016-09-27T17:22:21+01:00
better explanation frac_range
- - - - -
3630b070 by Sjors Scheres at 2016-09-28T10:02:39+01:00
Merged in scheres/relion-devel-lmb (pull request #85)
Master
- - - - -
fdf61d4e by Bjoern Forsberg at 2016-09-28T11:12:15+02:00
Merged tcblab/relion-devel into master
- - - - -
05856307 by Sjors Scheres at 2016-09-28T15:24:14+01:00
added hposition to scrollbar resets in running/finsihed jobs lists
- - - - -
fa174c12 by Sjors Scheres at 2016-09-28T15:24:56+01:00
warning about softer mask postprocessing
- - - - -
39294fe3 by Sjors Scheres at 2016-09-28T15:25:28+01:00
make frac_range work with helical particle expand
- - - - -
d33a21a8 by Sjors Scheres at 2016-09-28T15:25:52+01:00
added fourier filter
- - - - -
274080f1 by Sjors Scheres at 2016-10-04T16:42:03+01:00
reset optimiser.star input for refine jobs, and de-link coords suffix files for auto and manual picking (issue nr 114)
- - - - -
553cea3d by Sjors Scheres at 2016-10-04T16:42:54+01:00
Merged in scheres/relion-devel-lmb (pull request #86)
Master
- - - - -
791d4fbe by Sjors Scheres at 2016-10-05T12:57:18+01:00
default psistep of 6 degrees, so GPU will do 5.7
- - - - -
a28890ba by Sjors Scheres at 2016-10-05T14:37:42+01:00
Merged tcblab/relion-devel into master
- - - - -
b3cdbc74 by Sjors Scheres at 2016-10-10T15:37:25+01:00
repaired bug with D1 symmetry, reported by Jaime Martin-Benito
- - - - -
3d722f4b by Sjors Scheres at 2016-10-12T10:13:55+01:00
use install instead of mkdir to create scratch dir with correct permissions
- - - - -
1484086b by Sjors Scheres at 2016-10-12T10:15:19+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
33634691 by Sjors Scheres at 2016-10-12T10:21:08+01:00
save min_dedicated to file
- - - - -
799a5f6b by Sjors Scheres at 2016-10-12T12:01:30+01:00
check for empty autoopick references, issue 124
- - - - -
d0ae7160 by Sjors Scheres at 2016-10-13T16:14:20+01:00
only print estimated memory if verb=2
- - - - -
8131c5b4 by Dari Kimanius at 2016-10-14T16:15:48+02:00
More precise error reports. And better exception handling in threads.
- - - - -
beee00f0 by Dari Kimanius at 2016-10-14T16:24:16+02:00
Merged tcblab/relion-devel into master
- - - - -
040e0ecc by Sjors Scheres at 2016-10-14T14:45:33+00:00
Merged in tcblab/relion-devel-tcblab (pull request #87)
More precise error reports. And better exception handling in threads.
- - - - -
fc9d3a56 by Sjors Scheres at 2016-10-17T09:07:46+01:00
Merged tcblab/relion-devel into master
- - - - -
fa91c71e by Sjors Scheres at 2016-10-20T10:11:16+01:00
added -Bft and -InitDose to motioncor2 options
- - - - -
83c66959 by Sjors Scheres at 2016-10-20T10:12:49+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
26250a8b by Sjors Scheres at 2016-10-20T09:19:14+00:00
Merged in scheres/relion-devel-lmb (pull request #88)
Master
- - - - -
356e8ef4 by Sjors Scheres at 2016-10-20T11:00:08+01:00
reapired bug last_frame and gpu passing for motioncor2
- - - - -
9f7b2351 by Sjors Scheres at 2016-10-20T10:00:47+00:00
Merged in scheres/relion-devel-lmb (pull request #89)
reapired bug last_frame and gpu passing for motioncor2
- - - - -
fe596c57 by Sjors Scheres at 2016-10-20T11:02:11+01:00
Merged tcblab/relion-devel into master
- - - - -
828e7662 by Sjors Scheres at 2016-11-08T09:28:44+00:00
added 3D rotations to subtract option
- - - - -
d67e01fc by Sjors Scheres at 2016-11-08T09:39:02+00:00
resolved GUI buggy behaviour with symlinks, issues #152, 137
- - - - -
c2010e70 by Sjors Scheres at 2016-11-08T09:39:36+00:00
resolved GUI buggy behaviour with symlinks, issues #152, 137
- - - - -
0d209617 by Sjors Scheres at 2016-11-08T09:40:07+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
6abd8e78 by Sjors Scheres at 2016-11-08T09:40:47+00:00
Merged in scheres/relion-devel-lmb (pull request #90)
Master
- - - - -
ec5766ca by Sjors Scheres at 2016-11-08T09:41:09+00:00
Merged tcblab/relion-devel into master
- - - - -
dedba476 by Sjors Scheres at 2016-11-09T16:45:19+00:00
repair issue #157 with mpirun in front of echo commands
- - - - -
0a4baf21 by Sjors Scheres at 2016-11-09T16:45:34+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
47eb3455 by Sjors Scheres at 2016-11-09T16:47:49+00:00
resolves issue #122, problems with helix_toolbox on some systems
- - - - -
cc51c494 by Sjors Scheres at 2016-11-09T16:48:20+00:00
Merged in scheres/relion-devel-lmb (pull request #91)
Master
- - - - -
701c383b by Sjors Scheres at 2016-11-09T16:48:32+00:00
Merged tcblab/relion-devel into master
- - - - -
bb8482bd by Sjors Scheres at 2016-11-14T14:30:15+00:00
repaired bug introduced 5 days ago
- - - - -
5f1f600c by Sjors Scheres at 2016-11-14T14:35:25+00:00
Merged in scheres/relion-devel-lmb (pull request #92)
repaired bug introduced 5 days ago
- - - - -
05a2b184 by Sjors Scheres at 2016-11-14T14:36:01+00:00
Merged tcblab/relion-devel into master
- - - - -
dc760071 by Sjors Scheres at 2016-11-14T15:45:27+00:00
bug was still not fixed...
- - - - -
f657321e by Shaoda He PI Sjors Scheres added 15092014 at 2016-11-15T16:33:59+00:00
changes by Shaoda in helical code
- - - - -
556fc43f by Sjors Scheres at 2016-11-15T16:37:28+00:00
Merged in scheres/relion-devel-lmb (pull request #93)
Master
- - - - -
2551b029 by Sjors Scheres at 2016-11-15T16:37:46+00:00
Merged tcblab/relion-devel into master
- - - - -
2b70a71b by Dari Kimanius at 2016-11-21T12:45:10+01:00
Added support for up to 1024 coarse translation.
- - - - -
86934259 by Dari Kimanius at 2016-11-21T12:45:54+01:00
Merged tcblab/relion-devel into master
- - - - -
4092f7f0 by Sjors Scheres at 2016-11-21T16:53:20+00:00
Merged in tcblab/relion-devel-tcblab (pull request #94)
Added support for up to 1024 coarse translation.
- - - - -
5aa3c963 by Sjors Scheres at 2016-11-21T16:53:50+00:00
Merged tcblab/relion-devel into master
- - - - -
4a40f43b by Bjoern Forsberg at 2016-11-23T15:58:11+01:00
Fix to automatic GPU mapping for multi-node jobs
- - - - -
f6157c7c by Sjors Scheres at 2016-11-23T16:13:25+00:00
repaired bug local averaging in helices
- - - - -
56418b2c by Sjors Scheres at 2016-11-23T16:14:50+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
212f9dd3 by Sjors Scheres at 2016-11-23T16:15:50+00:00
Merged in scheres/relion-devel-lmb (pull request #95)
repaired bug local averaging in helices
- - - - -
46327310 by Sjors Scheres at 2016-11-23T16:16:18+00:00
Merged tcblab/relion-devel into master
- - - - -
60630fc4 by Sjors Scheres at 2016-11-24T11:43:02+00:00
make readonly version of main GUI
- - - - -
cbba8460 by Sjors Scheres at 2016-11-24T11:43:55+00:00
Merged in scheres/relion-devel-lmb (pull request #96)
make readonly version of main GUI
- - - - -
dd5838f7 by Sjors Scheres at 2016-11-24T11:44:05+00:00
Merged tcblab/relion-devel into master
- - - - -
f71f79c5 by Dari Kimanius at 2016-11-24T17:51:51+01:00
Merged tcblab/relion-devel into master
- - - - -
30ef4bf2 by Sjors Scheres at 2016-11-24T17:06:24+00:00
Merged in tcblab/relion-devel-tcblab (pull request #97)
Fix to automatic GPU mapping for multi-node jobs
- - - - -
80ca2ed9 by Sjors Scheres at 2016-11-24T17:06:35+00:00
Merged tcblab/relion-devel into master
- - - - -
40299364 by Sjors Scheres at 2016-11-29T08:41:19+00:00
resolves issue 171
- - - - -
a51d7907 by Sjors Scheres at 2016-11-29T08:41:40+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
eedf6c7e by Sjors Scheres at 2016-11-29T08:42:24+00:00
Merged in scheres/relion-devel-lmb (pull request #98)
resolves issue 171
- - - - -
94eed338 by Sjors Scheres at 2016-11-29T08:42:46+00:00
Merged tcblab/relion-devel into master
- - - - -
d344374f by Bjoern Forsberg at 2016-11-29T17:39:05+01:00
Allow direct use of _model.star as --ref input
- - - - -
d2fbe70c by Bjoern Forsberg at 2016-11-29T17:39:44+01:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
877f40d8 by Dari Kimanius at 2016-11-30T18:08:50+01:00
Patch for reported issue, when building coarse kernel in double precision..
- - - - -
d4bf630b by Dari Kimanius at 2016-11-30T18:09:52+01:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
cbdf88e7 by Bjoern Forsberg at 2016-12-01T13:50:24+01:00
SoftMask kernel is 3D
inital kernel (commit 2000c5b) did anticipate 3D input data, but
disabled it so as to save cycles (?).
- - - - -
eb9f44ce by Bjoern Forsberg at 2016-12-01T17:09:12+01:00
GFTAC entirely compliant with 3D-input
NOTE : subtomo still not working, but endof GetFTsAndCtfs show the
same Fimg and Fctf as evidenced by spotchecking.
- - - - -
d843468a by Bjoern Forsberg at 2016-12-02T09:10:25+01:00
Cleaning GFTAC
- - - - -
5033b8c7 by Bjoern Forsberg at 2016-12-02T10:23:59+01:00
template PowerClass kernel on 2D/3D
- - - - -
0ad4d607 by Bjoern Forsberg at 2016-12-02T10:52:14+01:00
prep subtomo-GPU; trans_z device arrays
- - - - -
ddab97be by Dari Kimanius at 2016-12-07T17:02:31+01:00
Rearrangements and clean-up in cuda convert
- - - - -
cdc6da35 by Bjoern Forsberg at 2016-12-07T20:03:55+01:00
GASD-coarse is compliant with 3D-input
Correct raw weights are given after first pass, using a simple testcase.
Template diffcoarse-kernel on 3d input data to preserve 2d input perf.
Label some x-y-z index extraction as optimization TODO, but since we
have not profiled it under subtomo, I don't know the importnace of it..
Also reduction of default softmask block size due to excessive resources
during debug session. Its a fairly light kenrel (I think) and boosting
the block size isn't the way to optimize it.
- - - - -
fd3f6149 by Bjoern Forsberg at 2016-12-07T20:04:08+01:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
61954915 by Bjoern Forsberg at 2016-12-08T15:08:30+01:00
collect2jobs + wavg are
collect2jobs wrapped to hide template args. Forced to use extern
shared memory allocation to avoid uneccessary smem alloc. It should
be fine since all vars are expanded compile-time.
wavg is untested but compiles and runs with new template var added.
- - - - -
f2b19e4b by Bjoern Forsberg at 2016-12-08T17:02:43+01:00
Backproject 3D template on 3D input
More instances of index checks that seem excessive, removed with
reference to the fact that we do fine without them in diff_coarse.
Prelim. tests (ctest no. 5 + subtomo_case) indicate this is ok,
but this should be furhter tested on much larger cases.
also: untested complementary changes to BP_2D
- - - - -
54f32652 by Bjoern Forsberg at 2016-12-08T17:03:08+01:00
Cleanup
- - - - -
ecae2615 by Bjoern Forsberg at 2016-12-08T19:03:47+01:00
Fix implicit cast in runCollect2jobs wrapper
- - - - -
9f1ef2a5 by Dari Kimanius at 2016-12-08T19:13:48+01:00
Added double-mode
- - - - -
696f4860 by Sjors Scheres at 2016-12-09T09:09:57+00:00
resolving issue #131
- - - - -
49fff584 by Sjors Scheres at 2016-12-09T09:10:15+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
89d47d90 by Sjors Scheres at 2016-12-09T09:10:56+00:00
Merged in scheres/relion-devel-lmb (pull request #99)
resolving issue #131
- - - - -
3b0c798c by Sjors Scheres at 2016-12-09T09:11:17+00:00
Merged tcblab/relion-devel into master
- - - - -
49a3f5c7 by Dari Kimanius at 2016-12-09T15:07:34+01:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
f07d6abe by Bjoern Forsberg at 2016-12-09T17:04:38+01:00
Fix to subtomo z-trans copying
also just allocate excessively for now; translations are not the biggest
of arrays...
- - - - -
6626c723 by Dari Kimanius at 2016-12-12T18:18:04+01:00
Fixed firstiter_cc on GPU
Fixed issue #122
- - - - -
1ba98aef by Bjoern Forsberg at 2016-12-15T16:54:30+01:00
fix to BP in 3D input
- - - - -
c789f7e1 by Bjoern Forsberg at 2016-12-16T10:29:19+01:00
Critical bugfix to diff coarse
diff2 coarse cuda kernel broken during 3D input compatibility work
(commit cdc6da3), last functional commit is commit ddab97b.
- - - - -
5f3b6c51 by Bjoern Forsberg at 2016-12-16T10:30:50+01:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
78f9e28f by Bjoern Forsberg at 2016-12-16T13:16:56+01:00
fix to reduced sum in backprojection
- - - - -
ccd2e5fe by Bjoern Forsberg at 2016-12-16T16:33:27+01:00
range check in some kernels fixed
conservatively set to match CPU code maximally
- - - - -
5aa8e27f by Bjoern Forsberg at 2016-12-19T10:09:21+01:00
3D input CC-kernels added
- - - - -
bf990d4e by Bjoern Forsberg at 2016-12-19T11:45:36+01:00
Softmask kernel split
Much more efficient, especially for large images (also 3D input)
- - - - -
2cf78b02 by Bjoern Forsberg at 2016-12-19T17:01:52+01:00
removed apparently unnecessary D2H cpy
- - - - -
3f68fb31 by Bjoern Forsberg at 2016-12-20T13:10:42+01:00
Added specilaization to compile-consts
also, identified issue #123
- - - - -
1709fb5a by Bjoern Forsberg at 2016-12-20T16:18:36+01:00
Further block-size tuning available for 3D input
Display notification about 3D input data on GPU as dev feature.
- - - - -
c988fbf8 by Bjoern Forsberg at 2016-12-20T16:24:45+01:00
Merged tcblab/relion-devel into master
- - - - -
cee76cd8 by Bjoern Forsberg at 2016-12-23T15:26:10+01:00
Fix to cmake 3.7.x compile
- - - - -
387c2e1b by Bjoern Forsberg at 2016-12-23T15:27:09+01:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
1c51b3b4 by Sjors Scheres at 2017-01-03T09:26:02+00:00
Merged in tcblab/relion-devel-tcblab (pull request #100)
subtomo on GPU, CC-fix, and _model.star as --ref
- - - - -
36e9360f by Sjors Scheres at 2017-01-03T09:26:34+00:00
Merged tcblab/relion-devel into master
- - - - -
7302dfdc by Sjors Scheres at 2017-01-03T11:30:40+00:00
fix issue #178 always_cc only CPU part now works. GPU still gives segfault.
- - - - -
9946e0c6 by Sjors Scheres at 2017-01-03T12:00:40+00:00
check reference size before noise estimation
- - - - -
68ccdec7 by Sjors Scheres at 2017-01-03T12:05:47+00:00
check if output directory exists before doing anything else
- - - - -
5071b968 by Sjors Scheres at 2017-01-03T12:56:53+00:00
only show limited number of images in multiview displayer
- - - - -
1bdc981d by Sjors Scheres at 2017-01-03T13:06:17+00:00
default random_seed is 1 for autopicker
- - - - -
bdb912ef by Sjors Scheres at 2017-01-03T15:00:41+00:00
allow --angpix_ref in sorting
- - - - -
aa5de763 by Sjors Scheres at 2017-01-06T09:46:03+00:00
changed long long int for unsigned long long in response to issue #180
- - - - -
d901cba9 by Sjors Scheres at 2017-01-06T09:46:27+00:00
minor change in I/O
- - - - -
ae7431a7 by Sjors Scheres at 2017-01-06T09:47:34+00:00
Merged in scheres/relion-devel-lmb (pull request #101)
Master
- - - - -
efddedb8 by Bjoern Forsberg at 2017-01-07T11:49:56+01:00
Merged tcblab/relion-devel into master
- - - - -
44eb96dc by Sjors Scheres at 2017-01-13T12:56:36+00:00
for compilation with C++ 11, as suggested by Javier Vargas
- - - - -
e652a863 by Sjors Scheres at 2017-01-13T12:57:03+00:00
Merged tcblab/relion-devel into master
- - - - -
a3a30a53 by Sjors Scheres at 2017-01-13T16:28:04+00:00
to solve problems with particles.star files as in issue #143
- - - - -
680303a6 by Bjoern Forsberg at 2017-01-16T14:57:46+01:00
Fix to euler matrix init
identified by Sheng Fu
- - - - -
c2ccadff by Bjoern Forsberg at 2017-01-16T14:58:37+01:00
CC-diffKernels use single-prec sqrtf when possible
Identified by Sheng Fu
- - - - -
97fe4e10 by Bjoern Forsberg at 2017-01-16T16:00:46+01:00
Only include re-extracted mics in new particles
Previously, the _extract.star files of all mics included during
re-extraction are joined to form a new particles.star. If re-
extraction writes into a directory with existing _extract.star
files (which pertain to mics which have no particles for re-
extraction) then the joined particles.star will conatin these
pre-existing particels as well. This can lead to the unintentional
inclusion of additional particles to those selected for re-
extraction.
This commit only includes the particles from _extract.star-files
which were found to contain re-extracted particles, and thus only
the particles selected for re-extraction. This increases stability
during uninteded use without affecting "correct" behaviour under
intended use.
- - - - -
301c0f85 by Bjoern Forsberg at 2017-01-16T16:57:17+01:00
Merged tcblab/relion-devel into master
- - - - -
2a8bea8a by Bjoern Forsberg at 2017-01-16T16:58:30+01:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
1c25bd5c by Bjoern Forsberg at 2017-01-16T22:00:52+01:00
Use env-vars for c-/xx compilers
As suggested by Johan Hattne on 2016-12-15
- - - - -
9efb1500 by Bjoern Forsberg at 2017-01-17T10:03:39+01:00
Allow parallel writes of fom-maps
- - - - -
2002222d by Bjoern Forsberg at 2017-01-17T10:18:14+01:00
Limit number of fom-maps without manual override
If one tries to write >20 fom-maps, the run is aborted and the user
is required to continue/re-run with the addition of --no_fom_limit.
This prevents accidental cluttering and high I/O loads, since the
previous commit enabled parallel fom writes.
- - - - -
797cae85 by Dari Kimanius at 2017-01-18T14:53:10+01:00
Fixed issue with skip_align
- - - - -
b2afb009 by Bjoern Forsberg at 2017-01-23T19:04:06+01:00
stated fom limit is actually default
- - - - -
fc54eeae by Sjors Scheres at 2017-01-24T11:13:41+00:00
Merged in tcblab/relion-devel-tcblab (pull request #102)
Master
- - - - -
6945cbdc by Sjors Scheres at 2017-01-24T11:13:57+00:00
Merged tcblab/relion-devel into master
- - - - -
3f699868 by Sjors Scheres at 2017-01-26T14:46:43+00:00
last additions from Shaoda to stable release
- - - - -
5a83d9dc by Sjors Scheres at 2017-01-26T14:47:04+00:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel
- - - - -
91024134 by Sjors Scheres at 2017-01-26T14:57:01+00:00
Merged tcblab/relion-devel into master
- - - - -
02eba242 by Bjoern Forsberg at 2017-01-26T16:48:21+01:00
Merged tcblab/relion-devel into master
- - - - -
d61272b9 by Bjoern Forsberg at 2017-01-31T17:21:16+01:00
Update ctests
- - - - -
3f150f1b by Bjoern Forsberg at 2017-01-31T17:22:00+01:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
d0fe1cae by Bjoern Forsberg at 2017-02-02T14:32:08+01:00
file->about + window displays version
- - - - -
140b8980 by Bjoern Forsberg at 2017-02-02T14:33:52+01:00
included math.h in marcos.h for compile support
- - - - -
8ae9a5d6 by Bjoern Forsberg at 2017-02-02T15:57:41+01:00
Added (tergeted) Benchmark timing for Maximization
- - - - -
19607e85 by Bjoern Forsberg at 2017-02-02T16:18:52+01:00
oversampling can be set >1
ctest no. 1 works GPUsingle/CPUdouble (prec.)
this fixes issue #96, but further tests are required
- - - - -
7b8adbcb by Bjoern Forsberg at 2017-02-02T19:22:44+01:00
Some segfaults replaced by nicer errors
- - - - -
5c0e9cb4 by Bjoern Forsberg at 2017-02-02T22:08:45+01:00
More segfaults clarified
- - - - -
f68e1ffe by Sjors Scheres at 2017-02-03T11:23:18+00:00
corrected typo in error message
- - - - -
744decbb by Sjors Scheres at 2017-02-03T11:31:20+00:00
express ranges as fractions of helical rise
- - - - -
4bae6013 by Sjors Scheres at 2017-02-03T11:31:47+00:00
red color for devel-gui
- - - - -
8a7c1c42 by Sjors Scheres at 2017-02-03T11:36:43+00:00
added first version of Stochastic Gradient Descent (SGD)
- - - - -
3fbb8e20 by Sjors Scheres at 2017-02-03T12:18:56+00:00
repaired bug in normal auto-refine
- - - - -
be14326a by Sjors Scheres at 2017-02-03T12:45:16+00:00
add options for denovo model to Class3D
- - - - -
c67ee6a3 by Sjors Scheres at 2017-02-03T12:45:32+00:00
forgot semi-colon
- - - - -
8f483b8a by Sjors Scheres at 2017-02-03T12:46:49+00:00
small edit of jobwindow class3D denovo ref#
- - - - -
013198d9 by Sjors Scheres at 2017-02-03T17:11:11+00:00
added max_effective for SGD and extra break when autorefine converges
- - - - -
2894ade9 by Bjoern Forsberg at 2017-02-04T14:15:47+01:00
Removed legacy cmake option cufft
- - - - -
82972b34 by Bjoern Forsberg at 2017-02-06T15:02:16+01:00
Nicer error messages
- - - - -
5fe8942d by Bjoern Forsberg at 2017-02-06T18:38:17+01:00
Device mapping occurs before noise estimation
- - - - -
ede4456c by Bjoern Forsberg at 2017-02-06T20:19:55+01:00
draft warn - box dims and GPUmem
- - - - -
272f75c4 by Bjoern Forsberg at 2017-02-07T10:52:10+01:00
more verbose and helpful image-size warning
- - - - -
3dcef86a by Bjoern Forsberg at 2017-02-07T11:43:22+01:00
REDME github formatted
- - - - -
79cc7043 by Bjoern Forsberg at 2017-02-12T17:45:06+01:00
Extra timings
- - - - -
346890eb by Bjoern Forsberg at 2017-02-12T18:06:11+01:00
output ppref after expSetup
- - - - -
2775a734 by Bjoern Forsberg at 2017-02-12T18:16:31+01:00
Mod output to .dat
easier to compare directly
- - - - -
8abf24a3 by Bjoern Forsberg at 2017-02-12T20:23:30+01:00
Added ref files
- - - - -
2eb4d68d by Bjoern Forsberg at 2017-02-13T10:20:58+01:00
One rank does expectationSetup() + MPI Bcast
- - - - -
aa0d3c5f by Bjoern Forsberg at 2017-02-13T11:24:24+01:00
Working for nr_slaves = nr_classes
- - - - -
5fb4d274 by Bjoern Forsberg at 2017-02-13T13:39:18+01:00
Working (but ~untested) PPref slave-division
- - - - -
581d2b1d by Bjoern Forsberg at 2017-02-13T16:12:10+01:00
Slave comm working
- - - - -
823bd253 by Bjoern Forsberg at 2017-02-13T17:34:41+01:00
Move PPrefRank-control to initialization
- - - - -
c42a4e99 by Bjoern Forsberg at 2017-02-13T17:51:40+01:00
Fixes to non-mpi 2D classification
- - - - -
ebc94dc6 by Bjoern Forsberg at 2017-02-13T17:56:56+01:00
Mute PPref .dat print unless debug
- - - - -
8459899a by Bjoern Forsberg at 2017-02-15T14:02:54+01:00
Update about-tab & version number display
- - - - -
004fa4ec by Bjoern Forsberg at 2017-02-15T14:09:46+01:00
Merge branch 'scaling'
- - - - -
3e923303 by Bjoern Forsberg at 2017-02-15T14:40:42+00:00
Merged in tcblab/relion-devel-tcblab (pull request #103)
new ctests, more informative errors, and faster expectationSetup with MPI
- - - - -
1233bef3 by Bjoern Forsberg at 2017-02-15T21:10:31+01:00
continue append is 3-zero padded: e.g. _ct004
- - - - -
47465840 by Sjors Scheres at 2017-02-16T14:22:01+00:00
trying to resolve conflicts in merge
- - - - -
90d8412f by Sjors Scheres at 2017-02-16T14:24:16+00:00
trying to resolve conflicts in merge
- - - - -
43104746 by Sjors Scheres at 2017-02-16T14:35:22+00:00
trying to resolve conflicts ml_optimiser.cpp
- - - - -
3ee006e0 by Sjors Scheres at 2017-02-16T14:39:09+00:00
trying to resolve conflicts ml_optimiser.cpp
- - - - -
1f05b1e5 by Sjors Scheres at 2017-02-16T14:41:23+00:00
trying to resolve conflicts ml_optimiser.cpp
- - - - -
48795fe1 by Sjors Scheres at 2017-02-16T14:59:40+00:00
trying to resolve conflicts ml_optimiser_mpi.cpp
- - - - -
69837eff by Sjors Scheres at 2017-02-16T15:01:56+00:00
trying to resolve conflicts ml_optimiser_mpi.cpp
- - - - -
783119d3 by Sjors Scheres at 2017-02-16T15:14:06+00:00
trying to resolve conflicts ml_optimiser_mpi.cpp
- - - - -
d7a7841c by Sjors Scheres at 2017-02-16T15:15:49+00:00
trying to resolve conflicts ml_optimiser_mpi.cpp
- - - - -
6690f1f0 by Sjors Scheres at 2017-02-16T15:27:11+00:00
trying to resolve conflicts ml_optimiser_mpi.cpp
- - - - -
a53166d4 by Sjors Scheres at 2017-02-16T15:27:26+00:00
Merged tcblab/relion-devel into master
- - - - -
547786f8 by Sjors Scheres at 2017-02-16T15:35:09+00:00
merged with Swedes, put do_sgd back into initialse
- - - - -
6b596e12 by Sjors Scheres at 2017-02-16T15:38:10+00:00
repaired back again after merge with Swedes
- - - - -
913c6947 by Sjors Scheres at 2017-02-16T15:54:05+00:00
trying to merge things back together again
- - - - -
520f4e21 by Sjors Scheres at 2017-02-16T18:08:57+00:00
put sgd back in after merge
- - - - -
6a1f5c27 by Bjoern Forsberg at 2017-02-20T18:10:45+01:00
Safeset MRC header
- - - - -
5ddfed47 by Sjors Scheres at 2017-02-22T15:53:57+00:00
just before mergeing with Shaoda: strict_highres_sgd
- - - - -
2f45b182 by Sjors Scheres at 2017-02-22T15:56:22+00:00
merge with Shaoda: subtomogram averaging for helices
- - - - -
ad630c71 by Bjoern Forsberg at 2017-02-23T16:58:03+01:00
Merged tcblab/relion-devel into master
- - - - -
fc1a8c49 by Sjors Scheres at 2017-02-24T11:31:39+00:00
repaired bug with None-reference and CTF in initial noise spectrum
- - - - -
bea9b091 by Sjors Scheres at 2017-02-24T11:31:57+00:00
remove comment from mask
- - - - -
8f0302dd by Sjors Scheres at 2017-02-24T11:33:47+00:00
Bjoern says this prevents stalls
- - - - -
128749b7 by Sjors Scheres at 2017-02-24T11:34:42+00:00
repaired typo
- - - - -
34030c95 by Sjors Scheres at 2017-02-24T11:37:25+00:00
still have to add MPI_COMM_SLAVES
- - - - -
e37f45e7 by Sjors Scheres at 2017-02-24T16:02:21+00:00
option for groups in motioncor2
- - - - -
220a10e7 by Sjors Scheres at 2017-02-27T11:01:57+00:00
added subset options to Class2D GUI
- - - - -
ca989856 by Sjors Scheres at 2017-02-27T11:05:03+00:00
Merged in scheres/relion-devel-lmb (pull request #104)
Master
Approved-by: Sjors Scheres
- - - - -
a88273b2 by Sjors Scheres at 2017-02-27T11:15:26+00:00
die if do_sgd and do_gpu
- - - - -
3b5a63ab by Sjors Scheres at 2017-02-27T11:16:42+00:00
Merged in scheres/relion-devel-lmb (pull request #105)
die if do_sgd and do_gpu
- - - - -
c35458a8 by Sjors Scheres at 2017-02-27T11:17:01+00:00
Merged tcblab/relion-devel into master
- - - - -
9a7cff1e by Bjoern Forsberg at 2017-02-28T10:13:22+01:00
Lock-class for scope:d mutex locks
- - - - -
e38aa82e by Bjoern Forsberg at 2017-02-28T10:14:43+01:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
15482060 by Sjors Scheres at 2017-03-01T11:54:55+00:00
put initial seed generation back into after taking that out for SGD
- - - - -
43be3a74 by Sjors Scheres at 2017-03-01T11:55:57+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
cc9a6504 by Sjors Scheres at 2017-03-01T16:05:10+00:00
ignore FSC-true if unmasked FSC doesnt go below 0.8
- - - - -
84743fe0 by Sjors Scheres at 2017-03-01T16:05:50+00:00
ignore FSC-true if unmasked FSC doesnt go below 0.8
- - - - -
1d59b3df by Bjoern Forsberg at 2017-03-02T15:19:17+01:00
Syncs necessary for non-block classStreams
NOTE: use of non-block classStreams still OFF by default.
- - - - -
dc6ad4d1 by Sjors Scheres at 2017-03-02T15:18:34+00:00
deal with FSC never below randomize_at
- - - - -
815d4745 by Dari Kimanius at 2017-03-02T17:18:29+01:00
Each thread gets its own private cuda stream. Some thrust calls still runs on default stream.wq
Should reduce risk for internal CUDA thread-lockes.
- - - - -
8b59168c by Dari Kimanius at 2017-03-03T10:05:32+01:00
Replaced thrust calls with CUB. Speed-up of 15-20% in expectation.
- - - - -
0731b491 by Sjors Scheres at 2017-03-06T09:14:17+00:00
minor tweaks
- - - - -
b50f5ab0 by Sjors Scheres at 2017-03-06T09:14:39+00:00
forgot to initialise back exp_random_class_some_particles
- - - - -
bbc94475 by Sjors Scheres at 2017-03-06T09:15:31+00:00
Merged in scheres/relion-devel-lmb (pull request #106)
Master
- - - - -
007dc9c0 by Sjors Scheres at 2017-03-06T09:15:46+00:00
Merged tcblab/relion-devel into master
- - - - -
b8d1f3f8 by Bjoern Forsberg at 2017-03-07T14:55:24+01:00
make install OWN_FFTW/FLTK follows install prefix
As suggested by Merlin Hartley
- - - - -
fbbb4371 by Bjoern Forsberg at 2017-03-07T15:19:29+01:00
Merged tcblab/relion-devel into master
- - - - -
1629f12c by Bjoern Forsberg at 2017-03-07T16:10:21+01:00
Revert "continue append is 3-zero padded: e.g. _ct004"
This reverts commit 1233bef3f6cc44a46dc699e91c75ca8ac4be9eba.
- - - - -
509fe45f by Bjoern Forsberg at 2017-03-07T15:23:48+00:00
Merged in tcblab/relion-devel-tcblab (pull request #107)
v.2.0.4
- - - - -
c4c49ba4 by Sjors Scheres at 2017-03-07T15:24:01+00:00
Merged tcblab/relion-devel into master
- - - - -
619f6c61 by Sjors Scheres at 2017-03-08T09:29:21+00:00
adapted for LMB
- - - - -
875c2221 by Sjors Scheres at 2017-03-09T09:31:36+00:00
put back sp.iclass_max in cuda part of optimiser
- - - - -
1fd555ed by Sjors Scheres at 2017-03-09T09:32:25+00:00
updated wrappers to Gctf and Motioncor2, remove MMOTIONCORR and CTFFIND3 from GUI
- - - - -
f9d25a48 by Sjors Scheres at 2017-03-09T09:36:29+00:00
replace qsub.csh with relion_qsub.csh
- - - - -
694001e5 by Sjors Scheres at 2017-03-09T09:37:21+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
67385899 by Sjors Scheres at 2017-03-09T09:38:24+00:00
Merged in scheres/relion-devel-lmb (pull request #108)
Master
- - - - -
d13489e2 by Sjors Scheres at 2017-03-09T10:48:28+00:00
add trainingset saving, off by default
- - - - -
3c752436 by Sjors Scheres at 2017-03-09T17:00:15+00:00
allow restarts of subset-calculations
- - - - -
ffa1ed40 by Bjoern Forsberg at 2017-03-11T14:28:30+01:00
args via --help/-h/<no flags>
- - - - -
671130e2 by Sjors Scheres at 2017-03-13T09:17:27+00:00
Merged tcblab/relion-devel into master
- - - - -
a2c6cfd2 by Sjors Scheres at 2017-03-14T12:39:33+00:00
repaired bugs in SGD and made InitialModel GUI
- - - - -
e1111456 by Sjors Scheres at 2017-03-14T12:39:53+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
26af2d90 by Sjors Scheres at 2017-03-14T12:53:05+00:00
changed default sigma-fudge inimodel on GUI
- - - - -
47fea860 by Sjors Scheres at 2017-03-14T12:55:08+00:00
removed remnants SGD from Class3D jobtype GUI
- - - - -
25f8dc2e by Sjors Scheres at 2017-03-14T14:14:43+00:00
added new options from Shawn Zheng in Motioncor wrapper
- - - - -
2409bdc5 by Sjors Scheres at 2017-03-14T14:15:55+00:00
Merged in scheres/relion-devel-lmb (pull request #109)
Master
- - - - -
e4de2b48 by Sjors Scheres at 2017-03-14T14:16:41+00:00
switch off ask training again
- - - - -
d0691216 by Sjors Scheres at 2017-03-14T14:17:16+00:00
Merged in scheres/relion-devel-lmb (pull request #110)
switch off ask training again
- - - - -
7dffd784 by Sjors Scheres at 2017-03-15T10:59:56+00:00
added subset option to Class3D GUI
- - - - -
4a21ed12 by Sjors Scheres at 2017-03-15T11:00:17+00:00
Merged tcblab/relion-devel into master
- - - - -
5875e3d9 by Sjors Scheres at 2017-03-15T11:00:37+00:00
Merged in scheres/relion-devel-lmb (pull request #111)
added subset option to Class3D GUI
- - - - -
01b26fa5 by Sjors Scheres at 2017-03-15T11:00:54+00:00
Merged tcblab/relion-devel into master
- - - - -
e41d4816 by Bjoern Forsberg at 2017-03-18T22:58:24+01:00
More general help flags
- - - - -
82e82b58 by Sjors Scheres at 2017-03-22T15:23:17+00:00
set FSC solvent-corrected to zero beyond current size
- - - - -
fc9a7981 by Sjors Scheres at 2017-03-22T16:58:06+00:00
MAJOR OVERHAUL of the entire GUI, separate from pipeliner
- - - - -
165d0aa2 by Sjors Scheres at 2017-03-22T17:01:46+00:00
changed hearde inclusions new GUI files
- - - - -
0c0cbbec by Sjors Scheres at 2017-03-22T17:02:05+00:00
removed error printing
- - - - -
02ca74a5 by Sjors Scheres at 2017-03-23T11:49:28+00:00
further debugging new GUI
- - - - -
dd15c2ce by Sjors Scheres at 2017-03-23T15:04:29+00:00
small bug helical options
- - - - -
711c926b by Sjors Scheres at 2017-03-23T15:04:46+00:00
removed debugging statements
- - - - -
bb4da40b by Sjors Scheres at 2017-03-24T11:54:36+00:00
still getting out bugs from the new GUI
- - - - -
7593c1b8 by Sjors Scheres at 2017-03-24T16:20:59+00:00
keep debugging new GUI
- - - - -
97c4b9c3 by Sjors Scheres at 2017-03-24T17:21:45+00:00
better behaviour GUI after clicking Run button
- - - - -
636f7220 by Sjors Scheres at 2017-03-27T12:19:15+01:00
repair bug select jobtype from coords reported by Xian
- - - - -
55ca2b01 by Sjors Scheres at 2017-03-27T12:37:45+01:00
removes clear from gui_entries and gui_jobwindow to prevent segfaults
- - - - -
f2a49f0a by Sjors Scheres at 2017-03-27T13:52:45+01:00
group of d-ohelix in polish window
- - - - -
2abde687 by Sjors Scheres at 2017-03-27T13:53:22+01:00
group of d-ohelix in polish window
- - - - -
fb2bf979 by Sjors Scheres at 2017-03-27T16:58:20+01:00
added CTF toolbox programme
- - - - -
587c2917 by Sjors Scheres at 2017-03-27T17:17:44+01:00
added first peak intact to ctf toolbox
- - - - -
6250ae07 by Sjors Scheres at 2017-03-28T09:30:22+01:00
remove floatToString from inner_helix diameter
- - - - -
c3ecf61d by Sjors Scheres at 2017-03-28T09:43:12+01:00
repaired bug directory browse for files
- - - - -
463a7217 by Sjors Scheres at 2017-03-28T11:27:06+01:00
fixed bug initialise empty directory
- - - - -
0b95e4f4 by Sjors Scheres at 2017-03-28T16:16:25+01:00
repaired bug manualpick job from display_io_node
- - - - -
db2e9333 by Sjors Scheres at 2017-03-28T16:18:05+01:00
just renaming global_manualpickjob to manualpickjob
- - - - -
696bea86 by Sjors Scheres at 2017-03-28T16:29:51+01:00
initialise manualpickjob upon reading
- - - - -
20cb2267 by Sjors Scheres at 2017-03-29T10:11:00+01:00
removed classes and zero-mask options from the HUI for inimodel
- - - - -
14a0d20a by Sjors Scheres at 2017-03-29T10:11:54+01:00
removed classes from inimodel, and repaired bug scheduling
- - - - -
c19bc03f by Sjors Scheres at 2017-03-30T16:08:50+01:00
tried to fix bugs pipeline reported by Rafa
- - - - -
f411d339 by Bjoern Forsberg at 2017-04-05T09:51:42+02:00
Update cub to 1.6.4
- - - - -
07c68404 by Bjoern Forsberg at 2017-04-05T16:59:45+02:00
Gaussian blob good warn + good help
Gaussian-blob picker tells you what's wrong rahter
than a segfault
Autopicker listens to --help / -h / no-flags
- - - - -
2602f3c6 by Bjoern Forsberg at 2017-04-05T17:16:26+02:00
All apps refrain from printing usage on err-catch
instead, rely on --help / -h / <no flags> for usage
- - - - -
d3a6f954 by Bjoern Forsberg at 2017-04-05T17:17:47+02:00
CPU-code listens to --maxsig
- - - - -
54709bc7 by Bjoern Forsberg at 2017-04-06T20:22:31+02:00
Header organization
- - - - -
a8371e13 by Bjoern Forsberg at 2017-04-08T11:15:43+02:00
Merged tcblab/relion-devel into master
- - - - -
3ba3efc0 by Bjoern Forsberg at 2017-04-08T11:16:56+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
f70ed533 by Bjoern Forsberg at 2017-04-08T16:00:55+02:00
CPU code listens to maxsig (correctly)
- - - - -
f99be475 by Bjoern Forsberg at 2017-04-09T15:47:20+02:00
removed some old subtomo test files
- - - - -
133df200 by Bjoern Forsberg at 2017-04-09T16:41:15+02:00
Tentative SGP BP-kernel
- - - - -
edda4090 by Bjoern Forsberg at 2017-04-10T14:42:42+02:00
Fix to maxsig on CPU
- - - - -
4d7ffcdf by Bjoern Forsberg at 2017-04-10T14:44:16+02:00
Fix to SGD GPU kernel
CPU/GPU multithread BROKEN
- - - - -
d5aff3bd by Sjors Scheres at 2017-04-18T12:42:15+01:00
repaired bug not reading manualpickjob when jobtype Select
- - - - -
7cacd161 by Sjors Scheres at 2017-04-18T12:52:05+01:00
repaired bug setting alias upon running
- - - - -
c5c6adec by Sjors Scheres at 2017-04-18T12:52:31+01:00
small bug active gpuid line upon continue
- - - - -
90f0c30e by Sjors Scheres at 2017-04-18T12:04:52+00:00
Merged in scheres/relion-devel-lmb (pull request #112)
Master
- - - - -
8ab1d322 by Bjoern Forsberg at 2017-04-19T14:05:07+02:00
Merged tcblab/relion-devel into master
- - - - -
d7a29a4a by Bjoern Forsberg at 2017-04-20T09:23:42+02:00
Compile-time warnings fixed
- - - - -
59ff0a7d by Sjors Scheres at 2017-04-25T10:00:28+01:00
Merged tcblab/relion-devel into master
- - - - -
23e3ec9f by Sjors Scheres at 2017-05-05T16:46:54+01:00
dont deactive gpus in autopicking window
- - - - -
d5fe0a09 by Sjors Scheres at 2017-05-05T16:47:20+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
41811d50 by Dari Kimanius at 2017-05-10T12:07:59+02:00
Framework for a merged accelerated code base.
acc_ptr.h defines a handler that manages two pointers similar to CudaGlobalPtr.
Which of the pointers that is used is defined by the second template parameter.
Some additional functionallity have been added to make the management more general.
Example setup in "main" code-flow can be found on line 282-304 in cuda_ml_optimiser.cu
- - - - -
47c70820 by Sjors Scheres at 2017-05-12T09:20:13+01:00
add Fourier Completeness to output of model.star file
- - - - -
7697f489 by Sjors Scheres at 2017-05-12T08:21:03+00:00
Merged in scheres/relion-devel-lmb (pull request #113)
Master
- - - - -
59adf06d by Dari Kimanius at 2017-05-12T12:16:54+02:00
Expantion of the AccPtr to incorporate the realspace translation helper-kernels translate2D and translate3D.
Code now works with boh ACC_CUDA and ACC_CPU as template variable to getFourierTransformsAndCtfs.
- - - - -
964b373f by Dari Kimanius at 2017-05-12T15:08:07+02:00
Introduced a new image-datatype that extends the AccPtr class to further compress the code.
The mereged accelerated example code can be found in cuda_ml_optimiser.cu line 319-349.
Switching between ACC_CUDA and ACC_CPU on line 3091 in cuda_ml_optimiser.cu sets the code up to be compiled in either implementation.
- - - - -
721f1b94 by Bjoern Forsberg at 2017-05-12T19:09:25+02:00
Add check for solvent mask value before first iter
- - - - -
9b0234f7 by Sjors Scheres at 2017-05-15T09:52:58+01:00
distribute fourier_completeness over MPI slaves
- - - - -
2c6acc32 by Sjors Scheres at 2017-05-15T15:29:47+01:00
adding support for local symmetry by Shaoda
- - - - -
5563736b by Sjors Scheres at 2017-05-15T15:30:12+01:00
added extra option for helices to GUI
- - - - -
436fdc6b by Sjors Scheres at 2017-05-15T15:30:51+01:00
Merged tcblab/relion-devel into master
- - - - -
68e71afb by Sjors Scheres at 2017-05-15T14:31:45+00:00
Merged in scheres/relion-devel-lmb (pull request #114)
Master
- - - - -
15abd4f3 by Sjors Scheres at 2017-05-15T15:32:16+01:00
Merged tcblab/relion-devel into master
- - - - -
7775cca2 by Sjors Scheres at 2017-05-17T10:33:36+01:00
added functionality to project random, uniform agular distribution
- - - - -
9797057a by Sjors Scheres at 2017-05-18T17:20:35+01:00
going back to multibody. This works perfectly on phantom2d, 3d to be tested still....
- - - - -
06017a2d by Sjors Scheres at 2017-05-19T16:56:06+01:00
also checked 3D transformations of new multibody
- - - - -
0d5678e0 by Dari Kimanius at 2017-05-23T17:07:14+02:00
More consistent naming and a few new operators.
- - - - -
2c63f206 by Sjors Scheres at 2017-05-25T14:50:07+01:00
now multibody with solvent_correct_fsc works: multiple fsc_halves per model
- - - - -
37f30fe5 by Sjors Scheres at 2017-05-26T15:53:23+01:00
now multibody refinement continues from autorefine
- - - - -
e0cd0d9f by Sjors Scheres at 2017-05-31T08:58:03+01:00
reorient bild file
- - - - -
6e942148 by Sjors Scheres at 2017-05-31T14:45:19+01:00
added skip body option in multibody and kept debugging that
- - - - -
b6777ded by Charles W Congdon at 2017-05-31T12:39:57-07:00
Move cuda files to src/acc/cuda from src/gpu_utils
- - - - -
81f00bfd by Charles W Congdon at 2017-05-31T13:18:04-07:00
Get everything building again after the move
- - - - -
f160b822 by Charles W Congdon at 2017-05-31T13:51:47-07:00
Create accelerated generic common file
- - - - -
169ca4df by Sjors Scheres at 2017-06-01T11:54:48+01:00
warn if fndata_extract is empty
- - - - -
1c497ae8 by Sjors Scheres at 2017-06-01T11:55:13+01:00
always use pixsize for motioncor2
- - - - -
fca768d7 by Sjors Scheres at 2017-06-01T11:55:55+01:00
added random sort and maximum number of particles per class
- - - - -
41ab6d98 by Sjors Scheres at 2017-06-01T12:44:38+01:00
random sort and max particles per class
- - - - -
87074d13 by Sjors Scheres at 2017-06-01T12:44:55+01:00
max_subsets only within 1st iter and write out message
- - - - -
3263186c by Sjors Scheres at 2017-06-01T12:45:06+01:00
always pass pixsize
- - - - -
0cbd0c03 by Sjors Scheres at 2017-06-01T12:45:28+01:00
warn empty refined particles star file
- - - - -
f5025030 by Sjors Scheres at 2017-06-02T10:08:05+01:00
repaired bug with new max_patrs_per_class
- - - - -
c6f35c38 by Sjors Scheres at 2017-06-02T16:23:46+01:00
green buttons for relion-2.1
- - - - -
1b103208 by Sjors Scheres at 2017-06-02T16:24:16+01:00
bug reported by Rafa Leiro about picking extra particles in refined coordinates
- - - - -
b84f34ce by Sjors Scheres at 2017-06-02T17:17:03+01:00
set version to 2.1-beta and make better explanations SGD on GUI
- - - - -
387225ec by Bjoern Forsberg at 2017-06-02T18:44:29+02:00
Merged tcblab/relion-devel into master
- - - - -
12de721e by Bjoern Forsberg at 2017-06-03T09:09:47+00:00
Merged in tcblab/relion-devel-tcblab (pull request #115)
Master
- - - - -
be1e8e06 by Sjors Scheres at 2017-06-05T09:36:12+01:00
correct display of options with --continue
- - - - -
02a201e3 by Sjors Scheres at 2017-06-05T09:36:24+01:00
green run button
- - - - -
84a22294 by Sjors Scheres at 2017-06-05T09:36:47+01:00
added gpu for inimodel and better help text
- - - - -
5d1b9cbd by Sjors Scheres at 2017-06-05T09:37:11+01:00
display errors in refine program without giving any options
- - - - -
bf145c05 by Sjors Scheres at 2017-06-05T09:46:53+01:00
added SGD paper and motioncor2 paper in about
- - - - -
6c45d6cb by Sjors Scheres at 2017-06-05T10:50:38+01:00
let fsc_true become negative
- - - - -
d1c9aebe by Sjors Scheres at 2017-06-05T14:24:54+01:00
debugging SGD
- - - - -
c2091910 by Sjors Scheres at 2017-06-05T14:37:15+01:00
added ScLileLab logo to background image
- - - - -
f11e20dd by Sjors Scheres at 2017-06-06T09:34:51+01:00
set version to 2.1-beta-0
- - - - -
1580f59c by Charles W Congdon at 2017-06-06T11:07:55-07:00
Start segmenting code, fix build
- - - - -
e8aa459c by Charles W Congdon at 2017-06-09T13:33:35-07:00
WIP - segmenting CPU and GPU code
- - - - -
db4d3667 by Charles W Congdon at 2017-06-09T17:00:10-07:00
WIP: Convert the rest of the world to the new shared pointer class
- - - - -
7b6312d1 by Dari Kimanius at 2017-06-12T13:45:22+02:00
Merged tcblab/relion-devel into master
- - - - -
6a5d6aaa by Charles W Congdon at 2017-06-12T16:19:22-07:00
WIP: Clean build with CUDA on; not run-tested
- - - - -
2da5db4f by Charles W Congdon at 2017-06-13T14:16:07-07:00
Change CUDA_DOUBLE_PRECISION to ACC_DOUBLE_PRECISION
- - - - -
4db6d76f by Sheng Fu at 2017-06-13T15:52:20-07:00
Add CPU kernels
- - - - -
ba1658ce by Sheng Fu at 2017-06-13T15:53:09-07:00
Merge branch 'cpu-cuda-merge' of https://bitbucket.org/tcblab/relion-devel-tcblab into cpu-cuda-merge
- - - - -
634c4a71 by Charles W Congdon at 2017-06-14T11:32:47-07:00
Move single-particle expectation into common code
- - - - -
c5c583dd by Charles Congdon at 2017-06-14T17:03:11-07:00
Fix build errors and crashes in CUDA debug build
- - - - -
7f6ff726 by Charles W Congdon at 2017-06-15T15:46:25-07:00
Fix some error checking
- - - - -
5a665271 by Sheng Fu at 2017-06-19T16:06:00-07:00
check in the kernels that include sincos lookup table
- - - - -
b146b216 by Charles W Congdon at 2017-06-19T17:04:06-07:00
WIP: Try to start bringing the optimized CPU code back on line
- - - - -
3d539e1f by Charles W Congdon at 2017-06-19T17:04:52-07:00
Merge branch 'cpu-cuda-merge' of https://bitbucket.org/tcblab/relion-devel-tcblab into cpu-cuda-merge
- - - - -
6a7bc35b by Sjors Scheres at 2017-06-20T14:54:47+01:00
changes for relion-2.1-beta release
- - - - -
633b4fd2 by Sjors Scheres at 2017-06-20T14:55:12+01:00
merging with Shaoda
- - - - -
81692c6f by Charles W Congdon at 2017-06-20T11:08:55-07:00
Fix Intel MPI build, experiment with CUDA-free ALTCPU builds
- - - - -
ebc80d9c by Charles W Congdon at 2017-06-20T16:41:47-07:00
Drop some templating for AccPtr, start getting ALTCPU code to build
- - - - -
24552dab by Sjors Scheres at 2017-06-21T11:52:25+01:00
repair bug nonsquare mics for premultiplication CTF
- - - - -
0de16d04 by Charles W Congdon at 2017-06-21T12:51:35-07:00
Fix runtime error and selection of files to build
- - - - -
38725147 by Charles W Congdon at 2017-06-22T11:18:03-07:00
Bring new CPU path into ml_optimiser, plus more instrumentation
- - - - -
d9193690 by Charles W Congdon at 2017-06-23T11:44:59-07:00
Revert to using original code kernels for utilities, continue CPU merge work
- - - - -
3ba6807e by Charles W Congdon at 2017-06-23T16:51:41-07:00
Attempt to merge slightly further
- - - - -
94a5ecb0 by Charles W Congdon at 2017-06-26T16:47:58-07:00
Firewall ..onDevice functions in AccUtilities
- - - - -
f7dbb6c0 by Sjors Scheres at 2017-06-27T12:48:48+01:00
draw lines for start-end coordinates in helices
- - - - -
db894e92 by Charles W Congdon at 2017-06-27T16:26:53-07:00
Convert more CUDA calls into wrapper calls
- - - - -
f8938727 by Sjors Scheres at 2017-06-28T12:53:50+01:00
functionhelp in helix_toolbox
- - - - -
0ab7efe6 by Sjors Scheres at 2017-06-28T12:54:10+01:00
add --noscale and --no_norm in refine
- - - - -
b4699bce by Sjors Scheres at 2017-06-28T12:54:37+01:00
typo bining to bnning
- - - - -
3e4efa72 by Charles W Congdon at 2017-06-28T09:28:16-07:00
Convert another CUDA call, remove CUDA ifdefs
- - - - -
a91009bf by Bjoern Forsberg at 2017-06-28T18:44:05+02:00
image_handler --o fn.mrcs is sane
per-image ops on stacks used to be
--i in.mrcs --o out.mrcs -> in_out.mrcs.mrcs
--i in.mrcs --o out -> in_out.mrcs
now you get
--i in.mrcs --o out.mrcs -> out.mrcs
--i in.mrcs --o out -> in_out.mrcs
- - - - -
2030bbb2 by Bjoern Forsberg at 2017-06-28T18:47:41+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
a349386e by Charles W Congdon at 2017-06-28T16:19:05-07:00
Move to a shared projector_plan
- - - - -
da188757 by Bjoern Forsberg at 2017-06-29T10:02:16+02:00
displayer can lowpass 2D-stacks
- - - - -
eea00b17 by Sjors Scheres at 2017-06-29T15:17:03+01:00
bug reparted by takanori
- - - - -
f5facd5e by Charles W Congdon at 2017-07-07T14:00:43-07:00
Touched 3rd-rail - shared projector, backprojector, helpers and resulting fallout - only a little more refactoring needed on CUDA side
- - - - -
e4abe20e by Sjors Scheres at 2017-07-10T14:47:06+01:00
oldstyle GUI option added
- - - - -
8416d0b8 by Sjors Scheres at 2017-07-10T14:47:37+01:00
make angpix compulsory
- - - - -
f714945b by Sjors Scheres at 2017-07-10T14:49:14+01:00
added safeguard for gold-standard separation and made specific local searches angular distributionin healpix_sampling for multibody refinement
- - - - -
a2a6986b by Charles W Congdon at 2017-07-10T11:22:59-07:00
Complete initial refactoring, fix memory leak
- - - - -
8e7d03a9 by Charles W Congdon at 2017-07-10T12:02:30-07:00
Do away with AccT parameter - currently serving no use
- - - - -
d6fbf166 by Charles W Congdon at 2017-07-10T13:38:06-07:00
Clean up duplicated code on CUDA side
- - - - -
40492a4e by Charles W Congdon at 2017-07-10T15:58:41-07:00
Clean up, including unnecessary file
- - - - -
4b3ee5fe by Charles W Congdon at 2017-07-11T12:51:32-07:00
Early WIP to get CPU code building
- - - - -
f6219c7e by Sjors Scheres at 2017-07-12T09:38:04+01:00
repaired bug oldstyle gui filenames from browser
- - - - -
8b73d781 by Sjors Scheres at 2017-07-12T09:38:12+01:00
diameter in angstroms not pixels in help manualpicker
- - - - -
2e65be9f by Sjors Scheres at 2017-07-12T15:59:49+01:00
now multibody seems to work OK, but it is very slow because psi-prior has ben deactivated
- - - - -
28e8564f by Charles W Congdon at 2017-07-12T16:59:59-07:00
Clean up CPU build of everything but missing FFTW calls
- - - - -
e9d60a70 by Charles W Congdon at 2017-07-13T16:30:24-07:00
Solve a few link issues
- - - - -
bdd2f2de by Charles W Congdon at 2017-07-13T16:52:20-07:00
Optimized CUDA and CPU branches build cleanly. Now to go back and put functionality back in
- - - - -
ddcf679e by Sjors Scheres at 2017-07-17T16:12:35+01:00
merging with Shaoda's local symmetry
- - - - -
3549ca68 by Charles W Congdon at 2017-07-17T13:38:39-07:00
Added global support for lambda functions (c++11) in cmake config.
- - - - -
758a2c4c by Charles W Congdon at 2017-07-17T16:44:21-07:00
AccPtr always allocates aligned memory
- - - - -
ac84bac3 by Charles W Congdon at 2017-07-18T09:35:59-07:00
A more cmakey way of linking to TBB that resolves linking issues
- - - - -
79a296c4 by Charles W Congdon at 2017-07-19T11:03:50-07:00
Hook up optimized CPU projectors
- - - - -
44389285 by Charles W Congdon at 2017-07-19T15:12:54-07:00
Complete CUDACOMPLEX to ACCCOMPLEX transition
- - - - -
9fa2bff6 by Charles W Congdon at 2017-07-19T16:43:16-07:00
Visit a few TODO items
- - - - -
16aa54bc by Dari Kimanius at 2017-07-20T18:19:28+02:00
Compile bug fix
- - - - -
3550fb1d by Dari Kimanius at 2017-07-20T18:54:35+02:00
Fixed exception pointer bug.
- - - - -
6cb6eb17 by Charles W Congdon at 2017-07-20T14:05:12-07:00
Make some bugfixes, including some from maineline
- - - - -
21039fbd by Charles Congdon at 2017-07-20T14:45:32-07:00
Fix double-allocation
- - - - -
37a27024 by Dari Kimanius at 2017-07-21T12:53:00+02:00
Fix to compile issue with CUDA enabled combined with DoublePrec_ACC.
- - - - -
a37f2076 by Sjors Scheres at 2017-07-21T15:32:38+01:00
angpix optional again
- - - - -
a957d54d by Sjors Scheres at 2017-07-21T15:35:07+01:00
now multibody rotations are around (0,90,0) to avoid problems with psi-prior
- - - - -
88769db0 by Charles W Congdon at 2017-07-21T15:08:58-07:00
Bugfixes to produce CPU signs of life
- - - - -
7950b2e5 by Sjors Scheres at 2017-07-24T11:52:45+01:00
now multibody has correct translations and rotations?
- - - - -
7eba2039 by Sjors Scheres at 2017-07-24T11:58:20+01:00
small error in psi prior
- - - - -
3a6174e0 by Dari Kimanius at 2017-07-24T14:44:49+02:00
Fix segfault when compiled with CUDA enabled and DoublePrec_ACC=ON.
- - - - -
65bbc65e by Sjors Scheres at 2017-07-24T14:23:54+01:00
bug in origin pixel
- - - - -
97b7d874 by Sjors Scheres at 2017-07-24T14:24:30+01:00
write out BILD file with body COMs and rotation directions
- - - - -
250b0bb3 by Sjors Scheres at 2017-07-25T09:45:13+01:00
repaired bug with sigma2_rot for helices when starting from local searches
- - - - -
e5d1e53d by Sjors Scheres at 2017-07-25T09:45:39+01:00
changed --help for --function_help to not clash with args.cpp anymore
- - - - -
c0fcce98 by Sjors Scheres at 2017-07-25T09:48:57+01:00
added warning sort and max nr images
- - - - -
d0e6546d by Sjors Scheres at 2017-07-25T09:49:54+01:00
removed archive from GUI
- - - - -
3162db2a by Sjors Scheres at 2017-07-25T10:43:55+00:00
Merged in scheres/relion-devel-lmb (pull request #116)
Master
- - - - -
097af633 by Sjors Scheres at 2017-07-25T14:15:39+01:00
Merged tcblab/relion-devel into master
- - - - -
26c29c98 by Bjoern Forsberg at 2017-07-25T17:23:16+02:00
Merged tcblab/relion-devel into master
- - - - -
f6821c1a by Bjoern Forsberg at 2017-07-26T10:44:48+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
2caccd5c by Charles W Congdon at 2017-07-27T13:57:30-07:00
Memory initialization bugfixes, etc. for CPU
- - - - -
f1e639af by Charles W Congdon at 2017-07-28T13:28:37-07:00
Centralize CUDA subouts for CPU, use different back-projector
- - - - -
f159b98e by Charles W Congdon at 2017-07-28T14:53:19-07:00
A few bug fixes
- - - - -
6b34698d by Sjors Scheres at 2017-07-31T08:48:39+01:00
added 2D->1D (back)projectors
- - - - -
f220f48d by Sjors Scheres at 2017-07-31T08:52:05+01:00
added new functionality to make 2d (xy) initial model for amyloid-like helices
- - - - -
3697eae3 by Sjors Scheres at 2017-07-31T08:54:23+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
9729f4dc by Sjors Scheres at 2017-07-31T09:00:35+01:00
repaired bug in tilt prior multibodies
- - - - -
e348c266 by Sjors Scheres at 2017-07-31T09:01:01+01:00
added multibody-specific mask-radius
- - - - -
9fb6c050 by Sjors Scheres at 2017-07-31T08:03:34+00:00
Merged in scheres/relion-devel-lmb (pull request #117)
Master
- - - - -
bf80e591 by Sjors Scheres at 2017-07-31T09:03:54+01:00
Merged tcblab/relion-devel into master
- - - - -
9013c3e6 by Dari Kimanius at 2017-07-31T17:01:10+02:00
Alloc before init fixed.
- - - - -
3c85e973 by Sjors Scheres at 2017-08-01T18:02:23+01:00
back to red GUI for devel version
- - - - -
643b0c40 by Sjors Scheres at 2017-08-01T18:03:00+01:00
2D->1D (back)projector normalisation for greyscale in inimodel2d helix
- - - - -
fa3ea674 by Sjors Scheres at 2017-08-01T18:04:01+01:00
repair bug multidoy affecting class2D
- - - - -
f366ae15 by Sjors Scheres at 2017-08-01T18:04:26+01:00
also read rlnReferenceImage
- - - - -
f8c81e89 by Sjors Scheres at 2017-08-01T17:04:58+00:00
Merged in scheres/relion-devel-lmb (pull request #118)
Master
- - - - -
e3a60c78 by Sjors Scheres at 2017-08-01T18:05:10+01:00
Merged tcblab/relion-devel into master
- - - - -
449a6d9c by Charles W Congdon at 2017-08-03T16:47:00-07:00
Memory addressing and size fixes
- - - - -
17bbb85e by Charles W Congdon at 2017-08-04T08:55:15-07:00
Merge branch 'cpu-cuda-merge' of https://bitbucket.org/tcblab/relion-devel-tcblab into cpu-cuda-merge
- - - - -
b79021be by Sjors Scheres at 2017-08-07T09:27:02+01:00
repair bugs reported by Andres Hernandez and by Takanori
- - - - -
453f3756 by Sjors Scheres at 2017-08-07T09:27:57+01:00
repair bug reported by Andres Hernandez: delete a single job in any menu segfaults
- - - - -
4e74a8cb by Sjors Scheres at 2017-08-08T12:50:25+01:00
disable utf8 error messages
- - - - -
c998ad80 by Sjors Scheres at 2017-08-09T09:33:30+01:00
repair bug with pad1 reconstructions
- - - - -
a8b3c1e0 by Charles W Congdon at 2017-08-09T10:05:49-07:00
Tweaks and bug fixes
- - - - -
4caeaba9 by Charles W Congdon at 2017-08-09T13:23:26-07:00
First attempt to reduce number of ifdefs
- - - - -
1ef315ea by Sjors Scheres at 2017-08-10T10:58:19+01:00
dont allow saving stdout, as it is a temp file
- - - - -
30dc3bfe by Sjors Scheres at 2017-08-10T10:59:14+01:00
deal with overlapping masks in multibody refinement
- - - - -
d8cdaa82 by Sjors Scheres at 2017-08-10T15:43:01+01:00
print out cl errors without giving options
- - - - -
f38edadc by Sjors Scheres at 2017-08-10T15:43:35+01:00
repair bug in overlapping masks for multibodies
- - - - -
ca302bf4 by Sjors Scheres at 2017-08-10T14:47:30+00:00
Merged in scheres/relion-devel-lmb (pull request #119)
Master
- - - - -
94c8b0c7 by Sjors Scheres at 2017-08-10T15:47:45+01:00
Merged tcblab/relion-devel into master
- - - - -
dbf3e620 by Charles W Congdon at 2017-08-10T09:45:17-07:00
Clean up legacy code from CPU-side source
- - - - -
882cdef6 by Charles W Congdon at 2017-08-10T15:46:04-07:00
Remove compile warning building with ICC
- - - - -
833651ad by Dari Kimanius at 2017-08-11T15:54:58+02:00
Merged tcblab/relion-devel into master
- - - - -
b477078f by Bjoern Forsberg at 2017-08-11T16:18:12+02:00
Fix to zero-distribution classes @ reconstruct
Especially obvious for small-set / subset runs. E.g. segfault occurs
for a 600-particle 2D-classification on 10 2D-classes, whne using
~100 particles subsets. This is not of tremendeous importance, but
for debugging it is preferrable with stable execution.
- - - - -
ccb42f8f by Sjors Scheres at 2017-08-11T16:38:00+01:00
add program to analyse results from multibody refinement, and perform better subtraction for focussed refinements
- - - - -
21da208f by Charles W Congdon at 2017-08-11T10:38:02-07:00
Use accelerated back projector for 2D case
- - - - -
0c76679d by Bjoern Forsberg at 2017-08-14T11:10:59+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
2dd23dff by Charles Congdon at 2017-08-14T16:46:22-07:00
Add optional MKL build, parallelism to more serial sections
- - - - -
df446358 by Charles W Congdon at 2017-08-15T13:15:33-07:00
Fix linking issues
- - - - -
2293d665 by Bjoern Forsberg at 2017-08-16T11:27:40+02:00
check against clesses/bodies in PPref setup
- - - - -
5a253454 by Bjoern Forsberg at 2017-08-17T13:27:07+02:00
Default format .spi for file-write is consistent
running e.g. mask_create with "--o test" produced a spider-format
image named "test", rather than "test.spi". This is now fixed.
The code is a bit unclear on the distinction between "extension"
( traditionally "suffix" such as .spi ) and "format" ( auxiliary
specifier to override filename, such as :mrc). This should probably
be revised, but nobody want's to spend their time doing that.
- - - - -
b980948c by Bjoern Forsberg at 2017-08-19T19:47:38+02:00
Auto-refine bugfix
This fixes a critical bug affecting half-set (gold-standard) refinement.
Starting with version 2.0.4, each iteration of auto-refinement incorrectly
uses the same half-set reconstruction as for alignment of all data. The FSC
and resolution reported is however still based on comparing two half-set
reconstructions which contain completely separate image data. A
degree of reference-bias is however present. The effect is to report
a resolution higher than what should be under gold-standard conditions,
since, reconstructions are subject to a degree of over-fitting.
In order to claim a so-called "gold-standard" resolution, use any version
before and including 2.0.3, or after (including) this version.
- - - - -
4dea187b by Dari Kimanius at 2017-08-22T17:43:08+02:00
Fix to CPU side issue in runCenterFFT
- - - - -
47601608 by Charles W Congdon at 2017-08-22T14:20:15-07:00
Generalize centerFFT across serial and parallel invocations
- - - - -
2f219cf5 by Dari Kimanius at 2017-08-23T12:45:37+02:00
Number of failsafe attempts adjustable by user via new flag: --failsafe_threshold
- - - - -
c6e72eb2 by Dari Kimanius at 2017-08-23T13:04:32+02:00
Better error reporting for failsafe-mode.
- - - - -
81661d8e by Dari Kimanius at 2017-08-23T16:39:51+02:00
Critical bugfix to random halfs protocol in auto-refine.
- - - - -
cc792be3 by Charles W Congdon at 2017-08-24T16:00:10-07:00
Switch to unoptimized backprojector as an experiment
- - - - -
0e87d566 by Bjoern Forsberg at 2017-09-04T16:04:15+00:00
Merged in tcblab/relion-devel-tcblab (pull request #120)
Master
- - - - -
602c7adf by Sjors Scheres at 2017-09-04T17:04:48+01:00
Merged tcblab/relion-devel into master
- - - - -
631dcd9c by Sjors Scheres at 2017-09-05T09:21:09+01:00
after merge with Bjoerns stuff
- - - - -
8effdc92 by Sjors Scheres at 2017-09-05T08:22:08+00:00
Merged in scheres/relion-devel-lmb (pull request #121)
Master
- - - - -
7e093e5b by Sjors Scheres at 2017-09-05T09:22:26+01:00
Merged tcblab/relion-devel into master
- - - - -
2bf94987 by Dari Kimanius at 2017-09-10T23:13:24+02:00
Merged tcblab/relion-devel into master
- - - - -
22edb936 by Charles Congdon at 2017-09-11T15:38:22-07:00
Go with alternate backprojection
- - - - -
f98d344e by Takanori Nakane at 2017-09-18T10:07:39+01:00
Fix SEGV when opening .mrcs files (not .star files) with relion_display --gui
- - - - -
bf9f4bc7 by Takanori Nakane at 2017-09-20T15:25:43+01:00
Fix a bug that "relion_preprocess --operate_on" creates filename.mrcs.mrcs.
- - - - -
47b3a8a6 by Takanori Nakane at 2017-09-20T15:31:06+01:00
Set pixel size in --locres map.
This is my pull request on Github #266 that solves the issue #239.
- - - - -
2adb960c by Sjors Scheres at 2017-09-20T15:31:34+01:00
new flex_analyse functionalities
- - - - -
8e36ffad by Sjors Scheres at 2017-09-20T15:34:26+01:00
small bug in flex analyser
- - - - -
bdfd7969 by Sjors Scheres at 2017-09-21T13:53:28+01:00
keep debugging multibody, joinTwoHalves opn higher ranks than 1,2 only
- - - - -
caa91550 by Takanori Nakane at 2017-09-22T11:12:37+01:00
Fix a normalization bug in relion_preprocess when rescaling.
- - - - -
01b83b6b by tnakane at 2017-09-22T10:44:16+00:00
Merged in fix-preprocessor-bug (pull request #123)
Fix a normalization bug in relion_preprocess when rescaling.
- - - - -
57abfa93 by Sjors Scheres at 2017-09-22T16:58:13+01:00
how to deal with selfRound in multibody offsets
- - - - -
35bcb707 by Charles W Congdon at 2017-09-25T10:24:07-07:00
Remove unneeded link option when not using Intel MKL
- - - - -
1f8964df by Charles W Congdon at 2017-09-25T10:25:24-07:00
Merge branch 'cpu-cuda-merge' of https://bitbucket.org/tcblab/relion-devel-tcblab into cpu-cuda-merge
- - - - -
c174c756 by Sjors Scheres at 2017-09-27T13:49:37+01:00
added do_map_generate to flex_analyser
- - - - -
2f14f88b by Sjors Scheres at 2017-09-27T13:50:02+01:00
repaired bug for multibody last iertation
- - - - -
f3ea5abb by Sjors Scheres at 2017-09-27T13:50:19+01:00
repair bug negative numbers
- - - - -
76b5aadf by Sjors Scheres at 2017-09-27T13:50:43+01:00
add phase randomisation
- - - - -
b8cd8b56 by Sjors Scheres at 2017-09-28T09:40:27+01:00
fix bodies if they dont have angular accuracy
- - - - -
3e2ab46f by Sjors Scheres at 2017-09-28T09:41:10+01:00
Merged tcblab/relion-devel into master
- - - - -
4aab626f by Sjors Scheres at 2017-09-28T08:43:08+00:00
Merged in scheres/relion-devel-lmb (pull request #124)
Master
- - - - -
7258c213 by Sjors Scheres at 2017-09-28T09:43:32+01:00
Merged tcblab/relion-devel into master
- - - - -
433c4010 by Sjors Scheres at 2017-09-28T09:57:26+01:00
stop multibody angular decrease earlier
- - - - -
4d02fc15 by Bjoern Forsberg at 2017-09-29T16:16:04+02:00
Merged tcblab/relion-devel into master
- - - - -
d2d1721b by Sjors Scheres at 2017-10-02T09:36:43+01:00
cannot change padding factor during a multi-body refinement...
- - - - -
ac7cbbbc by Sjors Scheres at 2017-10-02T12:03:38+01:00
repaired bug phase randomisation
- - - - -
3bd28636 by Dari Kimanius at 2017-10-02T16:21:38+02:00
Clean-up
- - - - -
7aca98f0 by Dari Kimanius at 2017-10-02T17:04:07+02:00
Merge branch 'master' into cpu-cuda-merge
Conflicts:
src/acc/acc_ml_optimiser_impl.h
src/ml_optimiser.cpp
src/ml_optimiser.h
src/ml_optimiser_mpi.cpp
- - - - -
3ab12c16 by Dari Kimanius at 2017-10-04T19:04:00+02:00
Improved precision in convertAllSquaredDifferencesToWeights
- - - - -
cbff6c5d by Dari Kimanius at 2017-10-05T17:15:05+02:00
Improved precision in convertAllSquaredDifferencesToWeights on acc-cpu side
- - - - -
ccec0f54 by Dari Kimanius at 2017-10-06T11:26:20+02:00
Making code hardware-agnostic
- - - - -
7ef56845 by Dari Kimanius at 2017-10-06T12:14:48+02:00
Code made more hw-agnostic
- - - - -
4126e861 by Dari Kimanius at 2017-10-10T13:24:07+02:00
AccPtr usage clean-up
- - - - -
d64b4f3e by Dari Kimanius at 2017-10-10T18:04:15+02:00
Introducing AccPtrFactory and AccPtr with a state-variable for AccType
- - - - -
6c8c6e05 by Sjors Scheres at 2017-10-11T11:03:58+01:00
repair bug hanging multibody refinement
- - - - -
21bf310a by Dari Kimanius at 2017-10-11T12:05:49+02:00
Replaced metadata REPORT_ERROR with exception to track line number of issues in star-files
- - - - -
bceecd1a by Dari Kimanius at 2017-10-11T12:19:47+02:00
Fixed issue with nvcc not picking up the c++11 std.
- - - - -
d097273a by Takanori Nakane at 2017-10-17T11:16:12+01:00
Reset mu to zero when SGD or subset have completed.
This ensures rlnClassDisribution sum to 1.
- - - - -
40ba4a74 by Takanori Nakane at 2017-10-17T16:30:55+01:00
Fixed a coordinate bug in singleViewerCanvas, where the amount of scroll was corrected twice.
- - - - -
27755d8f by Dari Kimanius at 2017-10-17T23:09:37+02:00
Fixed max value issue
- - - - -
b89334d3 by Dari Kimanius at 2017-10-18T16:13:43+02:00
Further improve maximum exponenet management
- - - - -
ecdc1544 by Charles W Congdon at 2017-10-18T10:16:58-07:00
Replace CUDA_DOUBLE_PRECISION with ACC_DOUBLE_PRECISION
- - - - -
14e38ccb by Dari Kimanius at 2017-10-19T15:25:50+02:00
Clean-up Image
- - - - -
923850fb by Dari Kimanius at 2017-10-19T17:26:31+02:00
Removing warnings and rounding up some functions for mutual cuda/cpu acc enabled binary.
- - - - -
bff86a63 by Dari Kimanius at 2017-10-19T17:26:49+02:00
Merge branch 'cpu-cuda-merge' of https://bitbucket.org/tcblab/relion-devel-tcblab into cpu-cuda-merge
- - - - -
343b0d16 by Dari Kimanius at 2017-10-19T17:37:46+02:00
Dropping COMPLEXTEXTURE to reduce complexity for enabling dual CUDA/CPU builds.
- - - - -
fc03af21 by Charles W Congdon at 2017-10-19T11:45:03-07:00
All projectors share the same complex-number data structure for accelerated CPU
- - - - -
cdf18d3d by Charles W Congdon at 2017-10-19T11:45:55-07:00
Merge branch 'cpu-cuda-merge' of https://bitbucket.org/tcblab/relion-devel-tcblab into cpu-cuda-merge
- - - - -
3dc6589b by Charles W Congdon at 2017-10-19T16:20:52-07:00
Fix bug in CPU-side backprojectSGD
- - - - -
8dafa02a by Charles Congdon at 2017-10-20T09:28:02-07:00
Fix race condition in CPU MPI initialization
- - - - -
311f6963 by Charles Congdon at 2017-10-20T09:47:15-07:00
Remove unnecessary memory allocation/copies of data from CPU backprojectors
- - - - -
1b262263 by Charles W Congdon at 2017-10-20T11:21:47-07:00
Pass pointers by reference rather than value - sigh
- - - - -
67748168 by tnakane at 2017-10-23T08:18:29+00:00
Merged in reset_sgd_mu (pull request #2)
Reset mu to zero when SGD or subset have completed.
- - - - -
c8eeafba by tnakane at 2017-10-23T08:18:45+00:00
Merged in display_scroll_bug (pull request #1)
Fixed a coordinate bug in singleViewerCanvas, where the amount of scroll was corrected twice.
- - - - -
6fb53b8f by Charles W Congdon at 2017-10-23T12:43:07-07:00
Add multithreaded FFT operations to ml_optimiser_mpi, mirroring what is in ml_optimiser
- - - - -
45e08504 by Charles Congdon at 2017-10-23T14:44:49-07:00
Fix build error for CUDA target
- - - - -
399b1052 by Takanori Nakane at 2017-10-25T09:51:30+01:00
Fixed an integer overflow in relion_stack_create.
(calculation of the necessary memory size)
- - - - -
19d7d2cd by tnakane at 2017-10-25T08:57:16+00:00
Merged in fix-overflow-in-stack_create (pull request #3)
Fixed an integer overflow in relion_stack_create.
- - - - -
f35dfc82 by Takanori Nakane at 2017-10-25T12:55:18+01:00
"Show metadata" in relion_display now tells the total number of selected particles.
Also made clear that the metadata shown is of the last-clicked class, not all of the selected classes.
- - - - -
af877d2b by Charles W Congdon at 2017-10-26T14:51:18-07:00
Various tweaks to try to get CPU code to run
- - - - -
a1654cc5 by Charles W Congdon at 2017-10-26T14:51:58-07:00
Merge branch 'cpu-cuda-merge' of https://bitbucket.org/tcblab/relion-devel-tcblab into cpu-cuda-merge
- - - - -
9325939f by Dari Kimanius at 2017-10-27T12:19:07+02:00
Fixed compile issues
- - - - -
88b0fa6e by Dari Kimanius at 2017-10-27T12:28:32+02:00
Unwinded allAlloc back to accAlloc, we don't want to allocate unused RAM when on GPUs.
- - - - -
c92e154c by Dari Kimanius at 2017-10-27T12:40:29+02:00
Fixed issue with min value caused by the C++11 NOT so intuitive definition of min/max.
- - - - -
fcaeb97e by Takanori Nakane at 2017-10-27T14:16:02+01:00
Add -LogFile option to MotionCor2.
- - - - -
47623d8e by Charles W Congdon at 2017-10-27T11:00:17-07:00
Replace numeric_limits::min() with numeric_limits::lowest() for float/double
- - - - -
fd1c326c by tnakane at 2017-10-30T14:37:08+00:00
Merged in motioncor2-logfile (pull request #5)
Add -LogFile option to MotionCor2
- - - - -
71f09609 by Charles W Congdon at 2017-11-03T14:33:52-07:00
Limits and memory leak fixes
- - - - -
82d476d6 by Sjors Scheres at 2017-11-06T09:30:25+00:00
Merged in scheres/relion-devel-lmb (pull request #125)
Master
- - - - -
faec8f5e by Takanori Nakane at 2017-11-06T14:41:25+00:00
Fix boundary check for MRC stacks.
- - - - -
2672f153 by Takanori Nakane at 2017-11-06T16:26:41+00:00
Supported TIFF (only for reading)
- - - - -
1b21d1db by Takanori Nakane at 2017-11-06T16:28:15+00:00
Fixed typo
- - - - -
31a1e8a7 by Takanori Nakane at 2017-11-07T11:24:43+00:00
Respect readdata flag in TIFF reader
- - - - -
c12d2d77 by Takanori Nakane at 2017-11-07T11:32:56+00:00
relion_run_motioncorr can handle --last_frame_sum for TIFF images now
- - - - -
642f1fd8 by Sjors Scheres at 2017-11-07T16:16:16+00:00
repaired restarting multibody refinement
- - - - -
e88254cb by Sjors Scheres at 2017-11-07T16:16:42+00:00
repaired restarting multibody refinement 2
- - - - -
5117b884 by Sjors Scheres at 2017-11-07T16:17:07+00:00
debugging thread-safety FFTW
- - - - -
aceae608 by Sjors Scheres at 2017-11-07T16:17:34+00:00
expanded flex analyser
- - - - -
478e357a by Sjors Scheres at 2017-11-07T17:11:54+00:00
more tweaks for convergence of multibodies
- - - - -
62d9835e by Sjors Scheres at 2017-11-07T17:12:23+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
42ec39ea by Charles W Congdon at 2017-11-07T10:32:29-08:00
Fix unitialized memory on CUDA
- - - - -
6fef8688 by Dari Kimanius at 2017-11-08T14:28:13+01:00
Merged tcblab/relion-devel into master
- - - - -
38333825 by Dari Kimanius at 2017-11-08T14:40:24+01:00
Merge branch 'master' into cpu-cuda-merge
- - - - -
e25d93d2 by tnakane at 2017-11-08T15:09:33+00:00
Merged in improved_show_metadata (pull request #4)
"Show metadata" in relion_display now tells the total number of selected particles.
- - - - -
abbaca5e by Dari Kimanius at 2017-11-11T09:23:25+01:00
Texture switched off when ALTCPU is enabled.
- - - - -
df7b1ef0 by Dari Kimanius at 2017-11-13T08:36:19+01:00
Fixed cmake warning
- - - - -
953012e6 by Sjors Scheres at 2017-11-13T12:39:11+00:00
set nr_sign correctly for non multibody refinement
- - - - -
b157418f by Sjors Scheres at 2017-11-13T12:50:23+00:00
added Chris and Richard-style Ewald sphere correction. Still developmental...
- - - - -
ed9e701c by Sjors Scheres at 2017-11-13T12:50:44+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
29aabc8a by Dari Kimanius at 2017-11-13T17:03:54+01:00
Backprojector on shared memory using whole-kernel mutexes.
- - - - -
03443459 by Dari Kimanius at 2017-11-13T17:27:11+01:00
MPI support for shared memory backprojection.
- - - - -
24783ac0 by Sjors Scheres at 2017-11-14T12:21:41+00:00
mask and write out eigenvectors
- - - - -
7fd89e47 by Dari Kimanius at 2017-11-14T16:14:20+01:00
Code clean-up
- - - - -
6ddb7bf5 by Dari Kimanius at 2017-11-14T17:02:26+01:00
Fixed double free bug for MPI
- - - - -
d43b6bca by Charles W Congdon at 2017-11-14T14:41:41-08:00
Only free shared projectors/backprojectors inside do_cpu for MPI
- - - - -
68a2d2f1 by Sjors Scheres at 2017-11-15T11:34:48+00:00
Jasenko new metadata libraries....
- - - - -
84e76866 by Dari Kimanius at 2017-11-15T15:09:41+01:00
Granualar mutexes for CPU backproject kernels.
- - - - -
2057051a by Dari Kimanius at 2017-11-15T15:09:58+01:00
Merge branch 'cpu-cuda-merge' of https://bitbucket.org/tcblab/relion-devel-tcblab into cpu-cuda-merge
- - - - -
e92b6fb8 by Dari Kimanius at 2017-11-15T16:49:51+01:00
Fixed error on icc
- - - - -
d46a4187 by Sjors Scheres at 2017-11-16T09:13:23+00:00
dont do solvent flattening when doing multibody
- - - - -
eac1a885 by Sjors Scheres at 2017-11-16T09:13:51+00:00
Merged tcblab/relion-devel into master
- - - - -
ede23e08 by Charles Congdon at 2017-11-20T15:04:48-08:00
CUDA build should not use TBB
- - - - -
e3a26336 by Charles Congdon at 2017-11-20T16:30:04-08:00
Clearly label this as a test version of the code
- - - - -
bbcbc986 by Sjors Scheres at 2017-11-23T11:15:32+00:00
also do beamtilt correction on imagfe used for alignment
- - - - -
527dbc66 by Takanori Nakane at 2017-11-23T14:28:30+00:00
Fix Jasenko's MetaData container for single precision mode.
- - - - -
5de7aa42 by Takanori Nakane at 2017-11-24T12:14:24+00:00
Fix boundary check for MRC stacks.
- - - - -
2d4998cf by Sjors Scheres at 2017-11-27T16:21:29+00:00
Merged in scheres/relion-devel-lmb (pull request #126)
Master
- - - - -
4699d351 by Dari Kimanius at 2017-11-28T11:23:48+01:00
Merged tcblab/relion-devel into master
- - - - -
f924c87b by Dari Kimanius at 2017-11-28T11:32:25+01:00
Fix to missing sort-button in the displayer.
- - - - -
ce90586e by Dari Kimanius at 2017-11-28T11:34:43+01:00
Fix to github issue #291.
Issue description:
This line causes an error when compiling with clang on macOS Sierra with latest Xcode installed.
Please consider changing the comparison to page > nullptr.
- - - - -
5db48c21 by Dari Kimanius at 2017-11-28T11:44:49+01:00
Fixed help-output for the last binary with this issue.
- - - - -
41747061 by Dari Kimanius at 2017-11-28T13:57:45+01:00
More extensive information about actions related to external libs during cmake configuration.
- - - - -
8a74e925 by Dari Kimanius at 2017-11-29T09:04:07+00:00
Merged in tcblab/relion-devel-tcblab (pull request #127)
Fiixes, for release 2.1
- - - - -
0560f35a by Sjors Scheres at 2017-11-29T11:55:29+00:00
allow tau2_fudge in Refine3D
- - - - -
501ea28d by Sjors Scheres at 2017-11-29T11:55:50+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
6fa18c34 by Andreas Schenk at 2017-11-30T15:00:31+00:00
added write(DO_LOCK) to release lock file again in the case somebody tries to mark a not yet started job as finished
- - - - -
10fb779c by Takanori Nakane at 2017-11-30T15:21:12+00:00
Update README.md as suggested on GitHub with minor modification by Takanori.
- pull request #233 by Andrew Jewett
- pull request #218 by Shyam Saladi
- - - - -
2cfeb1ed by Bjoern Forsberg at 2017-11-30T16:40:50+01:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
2a41959e by Bjoern Forsberg at 2017-11-30T16:55:39+01:00
CUDA 9 compatibility
Fix to rarely used diff-kernel
Upgrade cub to 1.7.3
- - - - -
5b28b2cc by Bjoern Forsberg at 2017-11-30T17:01:20+01:00
Prevent int underflow in packing
resolves issue #180 (github)
- - - - -
10fd2e17 by Bjoern Forsberg at 2017-11-30T17:34:19+01:00
Subset reconstruction handled sgd-like
Reconstruction should only occur when there has been something
back-projected. Using SGD, subsets might be small enough for
that not to happen, which results in bad estimations for the
cumulative weight. Code was written to deal with this, but it
did not cover the use of subsets under conventional (non-SGD)
execution. Runs which used subsets in conjunction with --maxsig
could therefore crash. This fix should avoid this issue.
- - - - -
014c93f4 by Bjoern Forsberg at 2017-11-30T17:34:59+01:00
More general null-check of buttons
- - - - -
1dd8467b by Bjoern Forsberg at 2017-11-30T17:34:04+00:00
Merged in tcblab/relion-devel-tcblab (pull request #128)
Master
- - - - -
d05ec6a4 by Bjoern Forsberg at 2017-12-01T08:37:01+01:00
MPI-subset recon (and activate for non-SGD)
- - - - -
f9bf06f6 by Sjors Scheres at 2017-12-01T09:28:26+00:00
add tau2_fudge to refine3d, hidden option
- - - - -
cd5fecd0 by Sjors Scheres at 2017-12-01T09:28:35+00:00
last changes by Shaoda
- - - - -
8430f85d by Sjors Scheres at 2017-12-01T09:28:49+00:00
updated references in about window for release 2.1
- - - - -
69bf03bd by Bjoern Forsberg at 2017-12-01T13:11:31+01:00
Merged tcblab/relion-devel into master
- - - - -
51412cee by Bjoern Forsberg at 2017-12-01T13:29:12+01:00
tau2_fugde is NOT adjusted using non-SGD subsets
Mild confusion;
The issue that commit 10fd2e1 and d05ec6a was trying to solve was due
to the lack of an accuracy-check under the else()-case covering non-sgd
reconstruction. Those commits then removed the else-case to achieve this,
but also activated the tau2_fudge adjustment for non-SGD subset runs.
This commits reverts the latter effect, because tau2_fudge should ONLY
be subset-adjusted when using SGD, due to the update-formula used.
- - - - -
1fcdf2ef by Dari Kimanius at 2017-12-01T14:12:39+01:00
Fixed my strange formulations.
- - - - -
64814406 by Bjoern Forsberg at 2017-12-01T14:37:10+01:00
Separate FLTK-deps by gui-lib
As suggested by Matt Williams (https://github.com/milliams),
with minor alterations to be cmake 2.8-compatible
- - - - -
019abdd2 by Bjoern Forsberg at 2017-12-01T14:39:04+01:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
3bdc8f66 by Bjoern Forsberg at 2017-12-04T09:58:13+00:00
Merged in tcblab/relion-devel-tcblab (pull request #129)
Master
- - - - -
fb43d542 by Charles Congdon at 2017-12-04T09:47:41-08:00
Version 3.0 alpha first stable version
- - - - -
5b473f9b by Sjors Scheres at 2017-12-05T09:48:13+00:00
do recompute plans if pointers change
- - - - -
f73d2c51 by Sjors Scheres at 2017-12-05T09:48:38+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
ebcfd193 by Sjors Scheres at 2017-12-05T11:31:56+00:00
repaired bug windowFourierTransform ambiguity even/odd ydim
- - - - -
8cde6465 by Sjors Scheres at 2017-12-05T11:35:17+00:00
repaired single precision code with plans_are_set variable
- - - - -
426f21ef by Sjors Scheres at 2017-12-05T11:44:37+00:00
Merged in scheres/relion-devel-lmb (pull request #130)
Master
- - - - -
50289dda by Charles Congdon at 2017-12-08T10:54:46-08:00
Fixes for large images and memory diagnostics to run when building debug
- - - - -
0bc6c6bf by Dari Kimanius at 2017-12-11T16:56:23+01:00
Merged tcblab/relion-devel into master
- - - - -
879ae79f by Charles Congdon at 2017-12-13T10:42:22-08:00
Correct use of some existing timers, add a few more to improve coverage
- - - - -
0361ee35 by Charles Congdon at 2017-12-13T15:37:37-08:00
Actually initialize new timers
- - - - -
a9786149 by Dari Kimanius at 2017-12-14T13:33:31+01:00
Merge branch 'cpu-cuda-merge'
Conflicts:
CMakeLists.txt
src/acc/cuda/cub/agent/agent_spmv_csrt.cuh
src/acc/cuda/cub/agent/agent_spmv_row_based.cuh
src/acc/cuda/cub/device/dispatch/dispatch_spmv_csrt.cuh
src/acc/cuda/cub/device/dispatch/dispatch_spmv_row_based.cuh
src/apps/CMakeLists.txt
src/metadata_container.cpp
src/metadata_table.cpp
- - - - -
74ca8b66 by Takanori Nakane at 2017-12-14T13:43:39+00:00
Fixed a bug in relion_image_handler --threshold_above and --threshold_below when re-windowing at the same time.
- - - - -
a2cc1a0b by Sjors Scheres at 2017-12-14T15:42:25+00:00
removed 2 instances of usage in ml_optimiser_mpi
- - - - -
09a63b3f by Sjors Scheres at 2017-12-14T15:43:17+00:00
Merged tcblab/relion-devel into master
- - - - -
05d21ee7 by Sjors Scheres at 2017-12-14T15:52:13+00:00
Merged in scheres/relion-devel-lmb (pull request #131)
removed 2 instances of usage in ml_optimiser_mpi
- - - - -
a9f5e9e6 by Takanori Nakane at 2017-12-14T15:55:42+00:00
Synced AUTHORS to that of the GUI.
- - - - -
04fcbe0a by Dari Kimanius at 2017-12-14T17:40:25+01:00
Stable accelerated multibody port up to diff-coarse.
- - - - -
32b5f360 by Charles Congdon at 2017-12-14T15:53:51-08:00
Also add additional instrumentation to MPI side and reuse some existing timers
- - - - -
7cf7bf43 by Takanori Nakane at 2017-12-15T10:08:24+00:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel
- - - - -
27c7d1ef by Takanori Nakane at 2017-12-15T11:02:44+00:00
Merge branch 'master' into tiff
- - - - -
9f67ae00 by Takanori Nakane at 2017-12-15T12:17:29+00:00
Use "Slower, more exhaustive search" in CTFFIND 4.1 unless --fast_search is ON.
"Slower search" was OFF by default in CTFFIND 4.1.5, but turned out to be
less accurate (ref: http://grigoriefflab.janelia.org/ctffind4).
- - - - -
2d997c74 by Sjors Scheres at 2017-12-15T12:32:09+00:00
Merged ctffind4-wrapper into master
- - - - -
1be3f829 by Sjors Scheres at 2017-12-15T12:32:28+00:00
Merged tcblab/relion-devel into master
- - - - -
bdee9757 by Sjors Scheres at 2017-12-15T12:33:13+00:00
Merged in scheres/relion-devel-lmb (pull request #132)
Use "Slower, more exhaustive search" in CTFFIND 4.1 unless --fast_search is ON.
- - - - -
bf8fc493 by Takanori Nakane at 2017-12-19T11:58:48+00:00
Fixed a bug where broken texts appear in the GUI instead of hard-coded defaults when environmental values like RELION_QSUB_TEMPLATE are not set.
(The cause was that "char mydefault[]" is valid only within the scope of the "if" block)
- - - - -
651990f1 by Takanori Nakane at 2017-12-19T15:40:29+00:00
Check if the extract_size is even in relion_preprocess
- - - - -
46358bca by Dari Kimanius at 2017-12-20T19:01:54+01:00
Clean up diff-coarse
- - - - -
cdd6f119 by Dari Kimanius at 2017-12-20T19:02:10+01:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
dd3a2808 by Dari Kimanius at 2017-12-22T10:33:32+01:00
Merged tcblab/relion-devel into master
- - - - -
4b7eee88 by Dari Kimanius at 2017-12-22T11:00:12+01:00
Expanded multibody port to the fine-pass and storeWeightedSums.
- - - - -
a7059baa by Dari Kimanius at 2017-12-22T11:00:22+01:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
2a117707 by Dari Kimanius at 2017-12-22T11:25:29+01:00
GCC compile issue fixed
- - - - -
2d269901 by Dari Kimanius at 2017-12-22T11:26:51+01:00
Removed legacy files.
- - - - -
1ee0d28d by Dari Kimanius at 2018-01-01T19:16:52+01:00
Unified code with pointer bundle.
- - - - -
fb1bdc5d by Charles Congdon at 2018-01-03T14:22:23-08:00
Restore fix to diff2.cuh made in commit 2a41959 that was lost in merge
- - - - -
0b25e862 by Sjors Scheres at 2018-01-04T09:50:20+00:00
repair segfault display class withour sort
- - - - -
357f7b91 by Sjors Scheres at 2018-01-04T09:50:45+00:00
put version back to devel
- - - - -
ddad957d by Sjors Scheres at 2018-01-04T09:54:37+00:00
added flip in rotations and remove hand1 from reconstruction
- - - - -
eb711039 by Sjors Scheres at 2018-01-04T09:55:00+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
60a873b7 by Dari Kimanius at 2018-01-04T12:07:18+01:00
Array size issue fixed.
- - - - -
7fae5fc5 by Dari Kimanius at 2018-01-04T21:01:29+01:00
Matrix multiplication issue fixed.
- - - - -
4fd6b790 by Dari Kimanius at 2018-01-04T21:01:59+01:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
42297602 by Sjors Scheres at 2018-01-05T17:24:26+00:00
add laplacian-of-gaussian blob detection autopicker
- - - - -
4a7869a0 by Dari at 2018-01-08T13:22:31+01:00
Fixed issue with both required double- and single-precision FFTW-lib when base-relion and acc-relion are compiled with different precisions.
The find packaged function for FFTW is made more cmake-like by enabling the specification of components used to specify the need for the precision-types.
- - - - -
be646ec4 by Sjors Scheres at 2018-01-09T14:51:19+00:00
LoG-picker improved
- - - - -
5ca2ea9c by Sjors Scheres at 2018-01-09T14:51:53+00:00
updated autopicker
- - - - -
8426c1c9 by Dari Kimanius at 2018-01-10T19:37:28+01:00
Inverse and correction matrices applied in right order.
- - - - -
01206199 by Dari Kimanius at 2018-01-10T19:44:36+01:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
a31fea78 by Sjors Scheres at 2018-01-11T14:52:22+00:00
allow scheduling and executing pipeline jobs from the command line
- - - - -
fd344430 by Sjors Scheres at 2018-01-11T14:53:20+00:00
also for command-line scheduling of pipeline jobs
- - - - -
e6263546 by Sjors Scheres at 2018-01-11T14:53:40+00:00
repair verb bug autopicker
- - - - -
49b30509 by Sjors Scheres at 2018-01-12T16:06:37+00:00
Merged tcblab/relion-devel into master
- - - - -
30848d5c by Takanori Nakane at 2018-01-17T14:55:59+00:00
Started working on Micrograph class.
- - - - -
ff8d3604 by Takanori Nakane at 2018-01-17T15:45:46+00:00
Store global motion from MotionCor2 into a STAR file
- - - - -
aacfa039 by Takanori Nakane at 2018-01-17T16:12:15+00:00
Repair error output from run_motioncorr_mpi
- - - - -
1f172d02 by Sjors Scheres at 2018-01-17T16:24:35+00:00
added amyloid-specific functionality to autopicker
- - - - -
8a9c3ff9 by Sjors Scheres at 2018-01-17T16:25:01+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
cef9c590 by Takanori Nakane at 2018-01-17T16:51:04+00:00
Global motion can be saved from MPI version of run_motioncorr
- - - - -
01ab8b82 by Takanori Nakane at 2018-01-17T17:26:43+00:00
Force MotionCor2 to always take the first frame to be aligned as the origin of the motion
- - - - -
388fa054 by Dari Kimanius at 2018-01-17T20:27:54+01:00
Fixed issue #132, both single and double precision FFTW libs are actually built when required simultaneously.
- - - - -
10c95c2e by Dari Kimanius at 2018-01-17T20:55:09+01:00
Fix for undefined cmake variable warning.
- - - - -
b9775ee8 by Sjors Scheres at 2018-01-18T09:12:32+00:00
add option to recenter re-extract images on a 3D coordinate in the reference
- - - - -
f87dfd83 by Takanori Nakane at 2018-01-18T14:17:05+00:00
Distinguish unaligned frame(s).
- - - - -
2367931d by Takanori Nakane at 2018-01-18T14:26:30+00:00
Made micrograph-model STAR filenames consistent with the image filenames (replace . with _)
- - - - -
2b07e310 by Sjors Scheres at 2018-01-18T16:29:41+00:00
preparing to incorporate Jasenko's code
- - - - -
1b980546 by Dari Kimanius at 2018-01-18T17:36:17+01:00
Correct created number of projectors and backprojectors in MB.
- - - - -
8a126e43 by Dari Kimanius at 2018-01-18T17:56:07+01:00
Fix number significants
- - - - -
06bc925a by Dari Kimanius at 2018-01-18T17:08:01+00:00
Merged in tcblab/relion-devel-tcblab (pull request #133)
New accelerated code
- - - - -
790d8b96 by Sjors Scheres at 2018-01-18T17:08:52+00:00
fixed memory leak on GPU with modified autopicker
- - - - -
3207a31a by Dari Kimanius at 2018-01-18T18:10:24+01:00
Merged tcblab/relion-devel into master
- - - - -
28e14ee9 by Dari Kimanius at 2018-01-19T10:15:09+01:00
Forgot about this.
- - - - -
c9b63110 by Dari Kimanius at 2018-01-22T17:52:29+01:00
Deallocate unused CPU-memory allocations.
- - - - -
0236bc0d by Dari at 2018-01-23T23:04:46+01:00
Fixed issue in FindFFTW.cmake
- - - - -
dfb9455f by Takanori Nakane at 2018-01-24T14:03:28+00:00
decomposePipelineFileName before looking up the group table in relion_project
- - - - -
51df10df by Takanori Nakane at 2018-01-24T15:38:30+00:00
Fixed a bug where relion_image_handler skips the first image in a stack.
Also clarified the meaning of select_img in Image.write().
- - - - -
8bfd538d by Dari at 2018-01-25T10:51:47+01:00
Fixed yet another issue in FindFFTW.
- - - - -
34abfd86 by Dari at 2018-01-25T11:04:46+01:00
TBB include issue fixed.
- - - - -
14f7286c by Sjors Scheres at 2018-01-25T11:01:52+00:00
improved version of amyloid autopicker
- - - - -
58564657 by Sjors Scheres at 2018-01-25T11:02:49+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
8abc93e1 by Jasenko Zivanov at 2018-01-25T15:18:04+00:00
Merged master, tiff and jzivanov
- - - - -
9fbe6f0f by Jasenko Zivanov at 2018-01-25T16:38:08+00:00
merged master, tiff and jzivanov; adapted to C++11
- - - - -
42be736c by Sjors Scheres at 2018-01-25T17:07:50+00:00
repaired bug with new GUI jobtypes, revert back to numbers to avois lots of trouble in transition
- - - - -
5ae6d0d9 by Dari Kimanius at 2018-01-25T18:56:54+01:00
Fix for compile warnings and handle memory allocation failure.
- - - - -
b705fa74 by Dari Kimanius at 2018-01-25T19:45:33+01:00
Handling different number of projectors and backprojectors.
- - - - -
462f2644 by Jasenko Zivanov PI Sjors Scheres added 28022017 at 2018-01-26T00:25:48+00:00
Tilt_fit now uses the FSC as a weight
- - - - -
5797bd66 by Dari Kimanius at 2018-01-26T09:39:34+00:00
Merged in tcblab/relion-devel-tcblab (pull request #134)
Multibody fixes
- - - - -
00e76b85 by Dari Kimanius at 2018-01-26T13:14:58+01:00
Fixed out of bounds array access in symmetriseReconstructions for multibody.
- - - - -
6b2773f4 by Jasenko Zivanov at 2018-01-26T14:36:46+00:00
changed libRelion to static linking - this makes the programs archivable
- - - - -
2bc23443 by Dari Kimanius at 2018-01-26T15:27:33+00:00
Merged in tcblab/relion-devel-tcblab (pull request #135)
Please double-check this edit.
- - - - -
2376ad1b by Jasenko Zivanov at 2018-01-26T15:37:30+00:00
defocus and tilt fit now display the usage when called without mandatory arguments
- - - - -
e80290a8 by Jasenko Zivanov at 2018-01-26T17:11:37+00:00
tilt fit refactored and untested
- - - - -
0f28a341 by Takanori Nakane at 2018-01-26T17:46:09+00:00
Use Image.setStatisticsInHeader and Image.setSamplingRate to reduce code duplications
- - - - -
09adae07 by Dari Kimanius at 2018-01-29T15:50:52+01:00
Require TBB when compiling with ALTCPU and completely remove requirement when not.
- - - - -
4e30bb15 by Takanori Nakane at 2018-01-29T16:41:24+00:00
Added --half_maps to relion_postprocess to output post-processed half maps for validation
- - - - -
7ac8e998 by Jasenko Zivanov at 2018-01-29T19:50:36+00:00
motion_fit refactored, reconstruct_openmp updated
- - - - -
73d38ef2 by Jasenko Zivanov at 2018-01-29T19:51:01+00:00
motion_fit refactored, reconstruct_openmp updated
- - - - -
c2e7f37c by Jasenko Zivanov at 2018-01-29T20:25:50+00:00
reconstruct_openmp bugfix
- - - - -
e29b70b4 by Jasenko Zivanov at 2018-01-29T20:27:26+00:00
reconstruct_openmp bugfix 2
- - - - -
dece2347 by Dari Kimanius at 2018-01-30T02:43:57+01:00
Minor: an tbb include not wraped in ALTCPU.
- - - - -
f31badab by Dari Kimanius at 2018-01-30T03:44:41+01:00
Drop min required cmake version to 2.8
- - - - -
9515f1c3 by Dari Kimanius at 2018-01-30T04:14:41+01:00
Sjors' bug fix, including oversampled translations in storeWeightedSums also for multibody.
- - - - -
0b3bb326 by Dari Kimanius at 2018-01-30T08:23:44+00:00
Merged in tcblab/relion-devel-tcblab (pull request #136)
Drop compilation requirements and multibody bug fix
- - - - -
687a665b by Takanori Nakane at 2018-01-31T11:58:07+00:00
Fix CUDA autopicker (shouldn't write combinedCCF for non-helices)
- - - - -
165ed310 by Takanori Nakane at 2018-01-31T13:55:05+00:00
Merged in fix-autopicker (pull request #6)
Fix CUDA autopicker (shouldn't write combinedCCF for non-helices)
- - - - -
17e2c97f by Takanori Nakane at 2018-01-31T14:54:42+00:00
Keep track of binning in micrograph model
(NOTE that shifts and coordinates refer to UNBINNED pixels in the original movie, because binning should be done AFTER motion correction.)
- - - - -
98c13ab1 by Takanori Nakane at 2018-02-01T14:20:58+00:00
Our own implementation of motion correction.
Work in progress. Implemented global correction.
- - - - -
a0181494 by Jasenko Zivanov at 2018-02-01T14:23:04+00:00
merged with master, fixed helical_rise bug in reconstruct and reconstruct_openmp
- - - - -
59f4bcba by Takanori Nakane at 2018-02-01T15:57:33+00:00
Implemented patch based trajectory.
- - - - -
9f19c386 by Takanori Nakane at 2018-02-01T21:11:24+00:00
Merge branch 'master' into tiff
- - - - -
e47afc4a by Takanori Nakane at 2018-02-01T21:13:16+00:00
Merge branch 'master' into tiff
- - - - -
3cc1f87b by Takanori Nakane at 2018-02-02T10:11:19+00:00
Implemented polynomial fitting and binning.
- - - - -
c5a36321 by Takanori Nakane at 2018-02-02T11:29:05+00:00
Polynomial fitting is working. Real space interpolation next.
- - - - -
abf6cb06 by Takanori Nakane at 2018-02-02T14:40:08+00:00
Implemented Real space interpolation. (do we need gridding correction?)
- - - - -
92d76731 by Bjoern Forsberg at 2018-02-05T12:44:11+01:00
Trial commit curand Noise w/ img-printout
- - - - -
dcc1e8ca by Takanori Nakane at 2018-02-05T12:16:55+00:00
Cut out binning of non-square images as a function.
(Probably binning and shift should be moved to fftw.h eventually)
- - - - -
3d348095 by Bjoern Forsberg at 2018-02-05T13:51:31+01:00
Cleaned up new noise-mask code
- - - - -
9c7e18e3 by Takanori Nakane at 2018-02-05T16:31:13+00:00
Dose weighting. (WIP: Testing needed!!!)
- - - - -
af21aa5a by Takanori Nakane at 2018-02-06T13:52:20+00:00
Show stack trace in REPORT_ERROR.
- - - - -
f3a1a8ba by Takanori Nakane at 2018-02-06T14:32:17+00:00
Explained how to decipher the backtrace.
- - - - -
01f53553 by Bjoern Forsberg at 2018-02-07T09:55:54+01:00
Clean and make safe
- - - - -
b06be64c by Takanori Nakane at 2018-02-07T11:15:46+00:00
Read multiple frames from TIFF images
- - - - -
b329917c by Takanori Nakane at 2018-02-07T11:16:54+00:00
Supported MPI-based parallelisation in motion correction.
(WIP: TODO: output should respect "verbosity")
- - - - -
9d04870b by Sjors Scheres at 2018-02-07T15:17:33+00:00
added directional lowpass filter
- - - - -
965c4df4 by Sjors Scheres at 2018-02-07T15:19:49+00:00
also write out unmasked references in 2D classification
- - - - -
30f756e6 by Sjors Scheres at 2018-02-07T15:20:12+00:00
set default repeats to 1 in pipeliner
- - - - -
78d2ff37 by Sjors Scheres at 2018-02-07T15:20:41+00:00
add output directory ooption to reposition particle program
- - - - -
fef2d6fd by Sjors Scheres at 2018-02-07T15:21:17+00:00
repaired bug in relion_rconstruct rise was in pixel, not in Angstroms
- - - - -
d6f5e216 by Sjors Scheres at 2018-02-07T15:21:54+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
2214d041 by Sjors Scheres at 2018-02-07T15:47:47+00:00
forgot to remove directional filter from ml_optimiser.h constructor
- - - - -
10035738 by Sjors Scheres at 2018-02-07T17:33:44+00:00
update GUI windows upon reading default .gui files
- - - - -
a0730587 by Jasenko Zivanov at 2018-02-07T21:39:55+00:00
new motion_fit; not fully tested
- - - - -
75ea2db2 by Jasenko Zivanov at 2018-02-08T02:22:12+00:00
new motion_fit debugged and tested
- - - - -
8b847988 by Sjors Scheres at 2018-02-08T15:17:55+00:00
repaired bug in recenter of tomograms
- - - - -
bd256423 by Sjors Scheres at 2018-02-08T16:36:03+00:00
repaired bug re-extract with tomograms with Dustin
- - - - -
36e847a7 by Bjoern Forsberg at 2018-02-08T17:58:18+01:00
Reorg getCTFs-block and re-enable >1 iter on GPU
Rec_img auto-refinements work on acc-path (GPU tested, ctest #9)
- - - - -
320373e1 by Bjoern Forsberg at 2018-02-08T18:21:29+01:00
Compilation error
missed due to cached compile config.
- - - - -
d9ca9ad9 by Takanori Nakane at 2018-02-08T20:28:48+00:00
Flip the Y axis when reading a TIFF. (see code comments for the rationale)
- - - - -
c62057b6 by Jasenko Zivanov at 2018-02-09T04:20:39+00:00
fixed heap-corruption bug in motion_fit; motion_fit now produces near-identical results on unextracted images as it did on pre-extracted ones; split off parallel version of FourierTransformer
- - - - -
216cfb78 by Jasenko Zivanov at 2018-02-09T04:21:16+00:00
Merge remote-tracking branch 'origin/tiff' into jzivanov
- - - - -
d2d67296 by Jasenko Zivanov at 2018-02-09T04:29:40+00:00
removed y-flipping of tiffs from StackHelper since it is now done by the tiff-reader
- - - - -
d5b60029 by Jasenko Zivanov at 2018-02-09T04:32:01+00:00
added parallel FT class to repository
- - - - -
d8861715 by Takanori Nakane at 2018-02-09T11:25:06+00:00
Fixed dose weighting.
- - - - -
119dfff3 by Sjors Scheres at 2018-02-09T12:47:16+00:00
repaired error for permissions in saveTrainingSet
- - - - -
c4966559 by Sjors Scheres at 2018-02-09T12:47:37+00:00
allow sorting on selected subsets
- - - - -
5fc70951 by Takanori Nakane at 2018-02-09T14:33:16+00:00
Fix MRC header (pixel size) from ml_optimiser.
We had to call setSamplingRateInHeader(apix) only AFTER X/Y/Z SIZE were defined.
- - - - -
df1a3f86 by Takanori Nakane at 2018-02-09T14:37:46+00:00
Merge branch 'master' into tiff
- - - - -
c2a12e24 by Takanori Nakane at 2018-02-09T19:50:10+00:00
Added timing codes. CenterFFT is the bottleneck!
- - - - -
623d2acb by Jasenko Zivanov at 2018-02-09T20:28:41+00:00
recombine_frames refactored and made compatible with unextracted movies
- - - - -
4dd11283 by Takanori Nakane at 2018-02-09T20:32:24+00:00
Removed CenterFFT from argmax CCF. This greatly improves global alignment performance.
(But not patch alignment performance; related to cache size?)
Now the bottle-neck is dose weighting and FFT.
- - - - -
fa2501ec by Takanori Nakane at 2018-02-09T20:42:29+00:00
Fixed a parallelisation bug in Fref calculation. Now results are identical with or without threads.
"schedule(static)" did not improve performance.
- - - - -
8e1161db by Takanori Nakane at 2018-02-10T06:45:59+00:00
Profile FFTW
- - - - -
c9982ff2 by Jasenko Zivanov at 2018-02-10T07:01:14+00:00
repaired and improved program argument logic: all refinement programs can now take an external Cs and kV as inputs; tilt_fit no longer requires an input star-file when a precomputed pixel fit is provided
- - - - -
af7cd43f by Jasenko Zivanov at 2018-02-10T11:00:53+00:00
implemented iterative tilt optimization; further work needed
- - - - -
b184d766 by Dari Kimanius at 2018-02-10T16:30:15+01:00
Merged tcblab/relion-devel into master
- - - - -
eaa13e21 by Takanori Nakane at 2018-02-10T18:59:48+00:00
Parallelise FFTs
- - - - -
a408000c by Takanori Nakane at 2018-02-10T19:40:50+00:00
Parallelised alignPatch().
Now ~ 70 sec for TcdA1 with 6 cores.
- - - - -
727e4c46 by Takanori Nakane at 2018-02-10T22:54:54+00:00
Parallelised patch preparation.
WIP: results not identical. Investigating...
- - - - -
0d3ffbe3 by Takanori Nakane at 2018-02-10T23:29:25+00:00
Make sure patch sizes are even. This fixes the threading bug in the previous commit.
(But leads to slightly worse RMSDs. Is padding better?)
- - - - -
4cf700e9 by Takanori Nakane at 2018-02-10T23:40:54+00:00
Fix timer labels.
- - - - -
1a745e8e by Takanori Nakane at 2018-02-11T10:33:35+00:00
Fix parallelisation of real-space interpolation.
Parallelise reference calculation.
- - - - -
74af370f by Jasenko Zivanov at 2018-02-11T11:00:25+00:00
Anisotropic tilt fit implemented and tested
- - - - -
e7a8c226 by Takanori Nakane at 2018-02-11T15:18:15+00:00
Reduce calls to SINCOS from nfx * nfy to nfx + nfy.
- - - - -
b5b30489 by Jasenko Zivanov at 2018-02-12T00:22:00+00:00
defocus_fit now supports anisotropic coma correction and reestimates per-micrograph astigmatism before per-particle defocus estimation
- - - - -
c8c65913 by Jasenko Zivanov at 2018-02-12T03:16:27+00:00
Support for anisotropic coma added to reconstruct_openmp; activated timing switch in BackProjector
- - - - -
b88fe3b2 by Jasenko Zivanov at 2018-02-12T05:13:42+00:00
BackProjector::reconstruct partly parallelized
- - - - -
2228a677 by Jasenko Zivanov at 2018-02-12T05:17:30+00:00
Removed --jfft parameter from reconstruct_openmp
- - - - -
f5451942 by Jasenko Zivanov at 2018-02-12T05:58:51+00:00
Removed parallelization in BackProjector::reconstruct; replaced by progress indicator
- - - - -
b33d7db8 by Jasenko Zivanov at 2018-02-12T06:31:40+00:00
Fixed bug in anisotropic coma correction
- - - - -
2904229b by Bjoern Forsberg at 2018-02-12T19:08:50+01:00
Remove unsafe de-allocation of host-projector
CPU-side allocations were deallocated to save RAM, but when
multiple deviceBundles (accelerated code path) are used,
this de-allocation creates null-pointers for all but the
first such bundle.
- - - - -
df9b9011 by Bjoern Forsberg at 2018-02-12T19:11:01+01:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
50c112a2 by Bjoern Forsberg at 2018-02-12T18:22:34+00:00
Merged in tcblab/relion-devel-tcblab (pull request #137)
Master
- - - - -
e9f8af24 by Charles Congdon at 2018-02-12T15:46:19-08:00
Fix issues with ALTCPU build
- - - - -
3b6d3ce4 by Jasenko Zivanov at 2018-02-13T04:29:42+00:00
tilt no longer applied twice in defocus estimation
- - - - -
56b59d0a by Jasenko Zivanov at 2018-02-13T07:59:55+00:00
corrected units in anisotropic coma model
- - - - -
37ac67b4 by Jasenko Zivanov at 2018-02-13T08:56:20+00:00
reverted units in anisotropic coma model
- - - - -
1daef6e7 by Sjors Scheres at 2018-02-13T14:32:23+00:00
improved SGD initial model generation and subset handling in general; now subsets are just shorter iterations; particle order gets randomised at every iteration
- - - - -
d5d75df7 by Sjors Scheres at 2018-02-13T14:42:15+00:00
forgot a comman in rlnSgdFinalRsolution
- - - - -
96b011a9 by Takanori Nakane at 2018-02-13T18:08:40+00:00
Make micrograph model C++11 compatible.
- - - - -
cd6721a7 by Sjors Scheres at 2018-02-13T18:31:48+00:00
added GUI options for fast_subsets in 2D and 3D classification
- - - - -
66c54748 by Sjors Scheres at 2018-02-13T18:32:31+00:00
changed I/O of postprocessing, to allow post-processing of multibodies
- - - - -
ebcfee3b by Sjors Scheres at 2018-02-13T18:52:20+00:00
added own_motioncor and multibody jobtype to GUI
- - - - -
f662b48a by Sjors Scheres at 2018-02-13T18:52:45+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
1d9aa420 by Jasenko Zivanov at 2018-02-13T19:21:25+00:00
added tilt test option to tilt_fit
- - - - -
24600927 by Jasenko Zivanov at 2018-02-13T19:49:09+00:00
fixed aniso-tilt bug in reconstruct_openmp
- - - - -
0e9904d6 by Takanori Nakane at 2018-02-13T20:02:54+00:00
Write rlnMicrographMetadata column to corrected_micrographs.star
- - - - -
22ee8144 by Sjors Scheres at 2018-02-14T09:40:55+00:00
attempt to resolve conflicts to sync relion-devel into relion-devel-lmb
- - - - -
171b9a11 by Sjors Scheres at 2018-02-14T09:41:29+00:00
Merged tcblab/relion-devel into master
- - - - -
f77abbe5 by Sjors Scheres at 2018-02-14T10:32:03+00:00
resolving conflicts from merging
- - - - -
44e3f263 by Sjors Scheres at 2018-02-14T11:54:44+00:00
resolved conflicts after merging, still need to test
- - - - -
9daca5c2 by Sjors Scheres at 2018-02-14T11:55:15+00:00
tried to solve cuda_autopicker after merging, but failed to do Mmean calculation. STILL PENDING
- - - - -
967fb1d5 by Takanori Nakane at 2018-02-14T12:04:25+00:00
Save aligned but non-dose weighted micrographs (--save_noDW).
This flag is always on for MOTIONCOR2.
- - - - -
4210f81c by Sjors Scheres at 2018-02-14T12:24:47+00:00
change GUI colors to orange for 3.0 alpha version
- - - - -
9831f683 by Sjors Scheres at 2018-02-14T12:25:00+00:00
hide amyloid picking from the GUI
- - - - -
767c6f72 by Sjors Scheres at 2018-02-14T12:52:08+00:00
now have Mmean working again in cuda_autopicker.cu
- - - - -
58dbf2ac by Jasenko Zivanov at 2018-02-14T13:57:20+00:00
global aberration plot
- - - - -
14428517 by Jasenko Zivanov at 2018-02-14T14:02:46+00:00
global aberration plot added
- - - - -
3fd51ec2 by Takanori Nakane at 2018-02-14T14:06:18+00:00
Merge branch 'master' into tiff
Resolved Conflicts:
src/apps/CMakeLists.txt
src/fftw.h
- - - - -
3c38e62f by Sjors Scheres at 2018-02-14T14:13:00+00:00
Merged in tiff (pull request #7)
Tiff
- - - - -
9c7a97fe by Sjors Scheres at 2018-02-14T14:26:23+00:00
add option to skip padding to multibody refinement
- - - - -
127cf523 by Sjors Scheres at 2018-02-14T14:28:05+00:00
add option to switch padding back on
- - - - -
be9180c9 by Takanori Nakane at 2018-02-14T15:04:29+00:00
Added pixel size, voltage, pre_exposure and dose rate to the micrograph metadata.
Fixed voltage (it was always 200 kV)
- - - - -
32e3def5 by Sjors Scheres at 2018-02-14T15:32:48+00:00
repaired cuda autopicker after merge. now confirmed tests
- - - - -
bb7dd64b by Sjors Scheres at 2018-02-14T15:34:14+00:00
removed debugging statements from own motioncor implementation and reactivated threads in GUI
- - - - -
9b8e5d68 by Jasenko Zivanov at 2018-02-14T15:48:14+00:00
added reference-based aberration plot
- - - - -
1345469b by Sjors Scheres at 2018-02-14T15:54:47+00:00
tidying up multibody jobtype
- - - - -
09bcc903 by Takanori Nakane at 2018-02-14T15:59:36+00:00
Moved shift calculation to MotionModel class
- - - - -
b166369c by Takanori Nakane at 2018-02-14T16:10:46+00:00
Merge branch 'master' into tiff
Resolved Conflicts:
src/motioncorr_runner.cpp
- - - - -
7decc405 by Jasenko Zivanov at 2018-02-14T16:21:29+00:00
bugfix in reference-based aberration plot
- - - - -
644eed76 by Jasenko Zivanov at 2018-02-14T16:52:38+00:00
bugfix 2 in reference-based aberration plot
- - - - -
5a5f7c3c by Takanori Nakane at 2018-02-14T17:28:34+00:00
Write fitting coefficients.
WIP: Reading needs more work.
- - - - -
91094530 by Sjors Scheres at 2018-02-14T18:48:44+00:00
write to stdout that we're reading optimiser.star to prevent apparent hanging of multibody job
- - - - -
61bae3ad by Sjors Scheres at 2018-02-14T18:49:25+00:00
repaired dangerous bug of mixing gold-standard half-sets with new subsets
- - - - -
b9893c19 by Sjors Scheres at 2018-02-14T18:49:58+00:00
added gold-standard test for bugs in acc code
- - - - -
f6425eba by Jasenko Zivanov at 2018-02-14T19:09:56+00:00
Added relion_recombine, a greatly simplified version of relion_recombine_frames. Removed deprecated parameters from relion_motion_fit.
- - - - -
7436f265 by Sjors Scheres at 2018-02-15T08:44:56+00:00
debugged new io of postprocessing
- - - - -
0a86fe74 by Sjors Scheres at 2018-02-15T08:45:53+00:00
changed handling of randomsieOriginalParticles call
- - - - -
a1f3e6ac by Sjors Scheres at 2018-02-15T09:27:17+00:00
removed very old jobfilenames in pipeline
- - - - -
c44ab817 by Sjors Scheres at 2018-02-15T09:28:00+00:00
in chronological order: show job nr first in lists of jobs
- - - - -
16cf61fd by Sjors Scheres at 2018-02-15T09:29:36+00:00
removed coreAllocate and DeAllocate of backprojectors
- - - - -
0d3c8598 by Jasenko Zivanov at 2018-02-15T09:44:33+00:00
hot pixel removal added to movie refinement programs
- - - - -
5ed122c9 by Jasenko Zivanov at 2018-02-15T09:49:01+00:00
hot pixel removal parallel again
- - - - -
50ce4806 by Sjors Scheres at 2018-02-15T11:24:10+00:00
repaired bug in ppipeliner of multibody
- - - - -
02eac8a1 by Sjors Scheres at 2018-02-15T11:32:56+00:00
try to repair bug reported by Takanori about decomposeFileNameSymlink
- - - - -
c80e8ba0 by Sjors Scheres at 2018-02-15T11:34:51+00:00
try to repair bug reported by Takanori about decomposeFileNameSymlink
- - - - -
39234bf7 by Sjors Scheres at 2018-02-15T14:57:57+00:00
store ints as longs and floats as doubles in MetaDataContainers
- - - - -
435b9ced by Sjors Scheres at 2018-02-15T14:58:18+00:00
repaired old debug statement
- - - - -
005fd0a7 by Sjors Scheres at 2018-02-15T14:58:42+00:00
raise errror is not enough particles for PCA
- - - - -
d090c850 by Sjors Scheres at 2018-02-15T14:59:45+00:00
added flex_analyser to MultiBody job type
- - - - -
8b78e788 by Jasenko Zivanov at 2018-02-15T16:22:46+00:00
added ctf_tilt option to reconstruct_openmp
- - - - -
33ddc620 by Sjors Scheres at 2018-02-15T16:56:28+00:00
swicthed tab 3 & 4 of multibody gui
- - - - -
c027ae83 by Sjors Scheres at 2018-02-15T16:56:52+00:00
set minimum step size [progress bar
- - - - -
a36bcd67 by Sjors Scheres at 2018-02-15T16:57:22+00:00
set version to 3.0-alpha-2
- - - - -
2d24f9e5 by Sjors Scheres at 2018-02-15T16:57:49+00:00
repaired outputnodes multibody jobtype
- - - - -
c7c6bef7 by Jasenko Zivanov at 2018-02-15T17:40:30+00:00
bugfix in ctf_tilt code
- - - - -
29f52415 by Dari Kimanius at 2018-02-15T20:23:25+01:00
Merged tcblab/relion-devel into master
- - - - -
ac7aa062 by Dari Kimanius at 2018-02-15T20:40:00+01:00
Reverted expectation memory free-ups. This was never a good idea anyways since max memory utilization is during expSetup and maxim.
- - - - -
c141ea61 by Jasenko Zivanov at 2018-02-15T19:41:47+00:00
ctf_tilt option added to defocus_fit
- - - - -
e2481c27 by Dari Kimanius at 2018-02-15T20:54:46+01:00
Don't raise signal in release mode.
- - - - -
3260b96f by Sjors Scheres at 2018-02-16T09:01:18+00:00
resend data_vs_prior for each body to all nodes in iterate
- - - - -
4b4b9643 by Sjors Scheres at 2018-02-16T09:01:45+00:00
also show job number in current job field
- - - - -
882a962a by Sjors Scheres at 2018-02-16T09:02:34+00:00
set minimum number of movies in slidebar to 0
- - - - -
a4b4bf1d by Sjors Scheres at 2018-02-16T09:04:09+00:00
Merged in scheres/relion-devel-lmb (pull request #138)
3.0-alpha-2
- - - - -
e11326d8 by Sjors Scheres at 2018-02-16T14:59:11+00:00
repaired error in filenames for flex_analyse command line generation
- - - - -
c4c45691 by Bjoern Forsberg at 2018-02-16T18:12:38+01:00
Draft duplicate removal (WIP, does not take effect)
- - - - -
38d4695c by Sjors Scheres at 2018-02-16T17:33:29+00:00
removed forgot button and repaired nasty bug NR_BROWSE_TABS
- - - - -
c8f3b573 by Sjors Scheres at 2018-02-16T17:34:04+00:00
dont move away old nodes from initial model and multibody job types
- - - - -
d56c8a14 by Sjors Scheres at 2018-02-16T17:37:58+00:00
Merged tcblab/relion-devel into master
- - - - -
729e7fe5 by Sjors Scheres at 2018-02-16T17:38:39+00:00
Merged in scheres/relion-devel-lmb (pull request #139)
Master
- - - - -
ae44a614 by Sjors Scheres at 2018-02-16T17:47:23+00:00
deactivate buttons on initial screen
- - - - -
464dc72f by Sjors Scheres at 2018-02-16T18:30:19+00:00
cannot deactivate other buttons: will stay that way
- - - - -
0553c6ac by Jasenko Zivanov at 2018-02-17T16:34:34+00:00
phase shift / amplitude contrast fitting and extensive diagnostic visualization added to defocus_fit; tilt removed from ctf
- - - - -
dcfd27da by Jasenko Zivanov at 2018-02-17T18:27:20+00:00
added experimental Cs-estimation to defocus_fit
- - - - -
9c02ea23 by Sjors Scheres at 2018-02-17T19:28:06+00:00
start to port Jasenko's defocus_fit and tilt_fit into a single ctf_refine program. MPI doesn't work yet. tilt-fit gives an ERROR. work-in-progress
- - - - -
7861f6d7 by Sjors Scheres at 2018-02-17T19:28:59+00:00
Merge branch 'jzivanov' of https://bitbucket.org/scheres/relion-devel-lmb into jzivanov
- - - - -
d3ef6e6f by Jasenko Zivanov at 2018-02-18T05:55:24+00:00
deprecated ctf_tilt functionality removed from ctf_refine
- - - - -
5f0e8e44 by Jasenko Zivanov at 2018-02-20T15:48:51+01:00
new experimental aberration plot
- - - - -
af8ce3c3 by Jasenko Zivanov at 2018-02-20T21:09:14+01:00
improved aberration plot
- - - - -
5523f0b5 by Jasenko Zivanov at 2018-02-20T21:44:19+01:00
fixed sign in aberration plot
- - - - -
98a7eba0 by Sjors Scheres at 2018-02-22T14:00:01+00:00
Merged tcblab/relion-devel into master
- - - - -
35b58c59 by Sjors Scheres at 2018-02-22T14:08:41+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
9df91cb4 by Sjors Scheres at 2018-02-22T14:12:10+00:00
removed TEST VERSION warnings
- - - - -
503eef5b by Sjors Scheres at 2018-02-22T15:00:07+00:00
bug upon convergence of auto-refine: combination of multibodies
- - - - -
231c0b2f by Sjors Scheres at 2018-02-22T15:00:26+00:00
report resolution to one decimal only
- - - - -
247e7d60 by Sjors Scheres at 2018-02-22T15:27:07+00:00
removed error in multidimarray when not necessary. Somehow this suddenly caused crashed in refine with GPU...
- - - - -
6d4e09b2 by Takanori Nakane at 2018-02-23T10:06:17+00:00
Fixed --do_fast_subset sizes
- - - - -
abba3c68 by Dari at 2018-02-24T20:21:36+01:00
Added print-out of version and build-info at the start of some important executables.
- - - - -
48dd2b5a by Dari at 2018-02-24T20:28:42+01:00
Merge remote-tracking branch 'remotes/tcblab/relion-devel/master'
- - - - -
bceb88b4 by Dari at 2018-02-24T20:36:09+01:00
Fixed compile issue for scope of isnan().
- - - - -
4302b959 by Dari at 2018-02-24T20:44:04+01:00
Fixed build with RELION_SINGLE_PRECISION defined.
- - - - -
be29959e by Sjors Scheres at 2018-02-26T12:21:14+00:00
repaired bug FSC higher bodies in last iteration
- - - - -
84cc659d by Sjors Scheres at 2018-02-26T12:21:46+00:00
put amyloid option back onto GUI for autopicking
- - - - -
2d93807e by Sjors Scheres at 2018-02-26T12:22:16+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
9d1d83a5 by Takanori Nakane at 2018-02-26T17:09:58+00:00
OpenMP parallelisation of relion_mask_create (--j option)
- - - - -
469428c7 by Takanori Nakane at 2018-02-26T17:14:02+00:00
Merge branch 'master' into tiff
- - - - -
8bd89d4f by Sjors Scheres at 2018-02-26T17:20:28+00:00
add threads for maskcreate jobtype
- - - - -
6a1f9dae by Sjors Scheres at 2018-02-26T17:23:20+00:00
added --j to relion_mask_create command
- - - - -
603b3921 by Takanori Nakane at 2018-02-26T19:40:57+00:00
own_motioncorr: write a log file for each movie, removed a duplicate definition of --j.
- - - - -
5479ff5f by Takanori Nakane at 2018-02-26T20:21:03+00:00
Read and write local motion model.
- - - - -
9f004e14 by Takanori Nakane at 2018-02-26T20:21:34+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
42fc2eda by Dari at 2018-02-27T09:20:42+01:00
Alternative merging scheme for overlapping multibody mask regions.
- - - - -
2aab21a4 by Sjors Scheres at 2018-02-27T09:11:34+00:00
removed suprious debugging statement
- - - - -
71d75c85 by Bjoern Forsberg at 2018-02-27T11:42:06+01:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
7b9325af by Dari Kimanius at 2018-02-27T11:42:50+01:00
Version print-outs are done on MPI master only.
- - - - -
91af1d95 by Dari Kimanius at 2018-02-27T12:15:04+01:00
Droped the discontinuity warning.
- - - - -
20765a15 by Dari Kimanius at 2018-02-27T12:49:20+01:00
Max box size warning adjusted by padding_factor.
- - - - -
1da2b382 by Takanori Nakane at 2018-02-27T12:18:16+00:00
Skip patch alignment & real-space interpolation when patch_x and patch_y are too small.
- - - - -
ae126d87 by Takanori Nakane at 2018-02-27T12:23:54+00:00
Added warnings for non-implemented features.
- - - - -
be306e7d by Bjoern Forsberg at 2018-02-27T13:34:10+01:00
Fixes to helical on gpu-noise
- - - - -
916a5546 by Bjoern Forsberg at 2018-02-27T13:35:54+01:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
83afe0d6 by Dari Kimanius at 2018-02-27T15:43:37+01:00
Segfault fixed (one-liner)
- - - - -
3f798da6 by Bjoern Forsberg at 2018-02-27T17:59:32+01:00
get/setHost methods passes multidimarrays by ref
These methods are currently only used by the helical code where
GPU-side functions are missing and we copy down to host-side
temporarily.
With this commit, the helical tag-case test works ok. It
became broken in commit 14e38cc.
- - - - -
f1e76f6c by Bjoern Forsberg at 2018-02-27T18:02:07+01:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
16a3dddf by Bjoern Forsberg at 2018-02-27T17:36:05+00:00
Merged in tcblab/relion-devel-tcblab (pull request #140)
Master
- - - - -
dd7e9e6d by Bjoern Forsberg at 2018-02-27T18:57:09+01:00
Added tag_case test targeting helical (zero-mask)
- - - - -
c7debcaa by Sjors Scheres at 2018-02-28T09:34:50+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
6d51661d by Sjors Scheres at 2018-02-28T09:40:07+00:00
resolving conflicts after Bjorn repaired error helical class2d
- - - - -
2fef87dc by Sjors Scheres at 2018-02-28T09:48:19+00:00
removed TEST TEST statements and renamed VERISON into VERSION
- - - - -
ab433564 by Sjors Scheres at 2018-02-28T09:49:19+00:00
Merged in scheres/relion-devel-lmb (pull request #141)
Master
- - - - -
640d80fe by Sjors Scheres at 2018-02-28T10:04:21+00:00
removed do_remove_duplicates suprious entry in extractjob, amd added little spacer on select job gui
- - - - -
ad1dde31 by Sjors Scheres at 2018-02-28T10:04:42+00:00
Merged tcblab/relion-devel into master
- - - - -
cd495de8 by Sjors Scheres at 2018-02-28T10:05:05+00:00
Merged in scheres/relion-devel-lmb (pull request #142)
removed do_remove_duplicates suprious entry in extractjob, amd added little spacer on select job gui
- - - - -
ece17c96 by Sjors Scheres at 2018-02-28T10:05:23+00:00
Merged tcblab/relion-devel into master
- - - - -
336d263e by Sjors Scheres at 2018-03-01T10:01:53+00:00
changed VERISON to VERSION
- - - - -
405a2ee3 by Takanori Nakane at 2018-03-01T15:18:54+00:00
Fixed GUI to support local resolution estimate of body reconstructions.
- - - - -
47311f74 by Takanori Nakane at 2018-03-01T16:26:19+00:00
Started working on frame grouping.
- - - - -
d3c548d2 by Takanori Nakane at 2018-03-01T17:03:22+00:00
Frame grouping in local alignment.
WIP: RMSDs remain rather high and convergence is slower. Need to try another approach.
- - - - -
b14ecbe2 by Takanori Nakane at 2018-03-01T17:53:19+00:00
Another approach to inter/extra-polate.
TODO: test which is better
- - - - -
1a39e7a3 by Jasenko Zivanov at 2018-03-01T19:19:48+00:00
phase fitting implemented in ref_aberration_plot
- - - - -
031cdb2e by Jasenko Zivanov at 2018-03-01T19:21:39+00:00
added missing files
- - - - -
b3de1250 by Jasenko Zivanov at 2018-03-01T21:31:08+00:00
aberration plot/fit improved
- - - - -
910e66ba by Charles Congdon at 2018-03-01T16:36:42-08:00
Attempts to use std::complex for projectors
- - - - -
499088fc by Takanori Nakane at 2018-03-02T13:45:42+00:00
Fixed interpolation. No need for special codes for extrapolation.
- - - - -
4beacfe5 by Takanori Nakane at 2018-03-02T14:54:57+00:00
Tidied up interpolation of shifts.
- - - - -
dc1970ff by Sjors Scheres at 2018-03-02T16:42:01+00:00
removed uninitialised first_x in 1D->2D backprojector
- - - - -
e2bff2cf by Sjors Scheres at 2018-03-02T17:06:41+00:00
repaired bug allAlloc -> deviceAlloc autopicker cuda code
- - - - -
b29a221d by Jasenko Zivanov at 2018-03-05T01:34:39+00:00
additive astigmatism in defocus_fit
- - - - -
5c4c7dd0 by Takanori Nakane at 2018-03-05T13:00:15+00:00
Respect n_threads in mask_create.
- - - - -
d3ca5ddf by Sjors Scheres at 2018-03-05T17:17:06+00:00
repaired dirty bug in helical refinement, reported by Jan Lowe
- - - - -
1ba509b7 by Sjors Scheres at 2018-03-05T17:18:20+00:00
repaired dirty bug in helical refinement, reported by Jan Lowe
- - - - -
cbcb8fe7 by Sjors Scheres at 2018-03-05T17:18:34+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
c2cffe77 by Sjors Scheres at 2018-03-05T17:28:59+00:00
added comment warning about bug reported by Jan Lowe
- - - - -
0c2910c6 by Takanori Nakane at 2018-03-05T21:14:16+00:00
MPI parallelisation of signal substraction in flex_analyser
- - - - -
58434d67 by Takanori Nakane at 2018-03-06T17:53:37+00:00
Fix libtiff linking
- - - - -
7d32d416 by Jasenko Zivanov at 2018-03-06T18:10:43+00:00
merged with tiff branch
- - - - -
3004d150 by Sjors Scheres at 2018-03-06T20:02:53+00:00
Merged in scheres/relion-devel-lmb (pull request #143)
Master
- - - - -
311e8a13 by Jasenko Zivanov at 2018-03-06T20:07:21+00:00
movies are now being loaded frame-by-frame during extraction
- - - - -
52429736 by Jasenko Zivanov at 2018-03-07T02:35:03+00:00
binning parameters replaced by pixel sizes in motion_fit; all binning types besides Fourier cropping removed
- - - - -
3630db6c by Jasenko Zivanov at 2018-03-07T03:35:25+00:00
std::numeric_limits<T>::lowest replaced by -std::numeric_limits<T>::max
- - - - -
0d338ac0 by Sjors Scheres at 2018-03-07T09:38:10+00:00
some more shifts 3 to data_dim changes
- - - - -
64341b3c by Dari Kimanius at 2018-03-07T13:14:36+01:00
Merged tcblab/relion-devel into master
- - - - -
f9fd79ca by Takanori Nakane at 2018-03-07T14:32:34+00:00
Added center of mass to relion_image_handler --stats
- - - - -
7129507f by Takanori Nakane at 2018-03-07T15:46:29+00:00
Merge branch 'master' into tiff
- - - - -
ca0db35c by Jasenko Zivanov at 2018-03-07T16:23:42+00:00
added search-and-replace hack for movie names
- - - - -
405c4bd4 by Jasenko Zivanov at 2018-03-07T16:40:02+00:00
search-and-replace hack more robust now
- - - - -
27012c9f by Takanori Nakane at 2018-03-07T16:44:30+00:00
Fixed frame numbers when --first_frame_sum is not 1.
- - - - -
acd22762 by Jasenko Zivanov at 2018-03-07T16:45:58+00:00
more debug info
- - - - -
f1a2463e by Takanori Nakane at 2018-03-07T16:54:38+00:00
Described why we don't scale x and y in Ewald sphere correction.
- - - - -
e3c99ed8 by Jasenko Zivanov at 2018-03-07T17:30:03+00:00
added support for particle coordinates given at micrograph resolution
- - - - -
3730be52 by Takanori Nakane at 2018-03-07T17:34:05+00:00
Combined map from relion_postprocess should be in the same scale
- - - - -
1155af71 by Takanori Nakane at 2018-03-07T17:35:07+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
6b229f6b by Takanori Nakane at 2018-03-08T11:27:31+00:00
Don't complain about angpix for movie extraction
- - - - -
cb99c5e8 by Jasenko Zivanov at 2018-03-08T15:10:26+00:00
B/k-factors are written out by recombine if --debug
- - - - -
921218df by Takanori Nakane at 2018-03-08T16:52:28+00:00
flex_analyser: re-center to the COM of the mask, combined STAR files from MPI slaves, removed unnecessary codes
- - - - -
0d775c25 by Sjors Scheres at 2018-03-08T17:00:31+00:00
Merged in tiff (pull request #8)
Tiff
- - - - -
5d9ba552 by Sjors Scheres at 2018-03-08T17:01:19+00:00
replaced error of normalisation check with a warning
- - - - -
cce8c990 by Sjors Scheres at 2018-03-08T17:02:14+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
e055959f by Sjors Scheres at 2018-03-08T17:10:39+00:00
replaced float for RFLOAT for many shift vectors
- - - - -
dc68a368 by Jasenko Zivanov at 2018-03-08T17:26:20+00:00
freq. ranges now given in A for recombination
- - - - -
6a7e01a2 by Takanori Nakane at 2018-03-08T19:45:56+00:00
No residual offset in flex_analyser
- - - - -
b750025b by Takanori Nakane at 2018-03-08T19:47:30+00:00
Merge commit '921218df3c366aca3c8860b3cf5bcb61e9982e23'
- - - - -
ce6528d4 by Takanori Nakane at 2018-03-08T19:48:18+00:00
Merge branch 'master' into flex
- - - - -
399b11b6 by Takanori Nakane at 2018-03-08T21:22:45+00:00
flex_analyser: removed one Fourier shift for performance.
- - - - -
325ec705 by Jasenko Zivanov at 2018-03-09T02:14:21+00:00
motion_fit parameters now normalized by pixel size and electron dose
- - - - -
bd3a717f by Jasenko Zivanov at 2018-03-09T03:21:56+00:00
fixed default arguments for motion_fit with normalized parameters
- - - - -
02407086 by Jasenko Zivanov at 2018-03-09T04:32:37+00:00
movie pixel size parameters now separated
- - - - -
de1af8dd by Sjors Scheres at 2018-03-09T11:47:30+00:00
improved error handling when noDW labels are not in input starfile
- - - - -
c994f255 by Takanori Nakane at 2018-03-09T15:39:00+00:00
own_motioncor: improved error & warning messages
- - - - -
14b49d18 by Takanori Nakane at 2018-03-09T19:59:41+00:00
Fixed logic with noDW
- - - - -
a2b007de by Takanori Nakane at 2018-03-09T20:36:44+00:00
Fixed SEGV when save noDW & do binning.
- - - - -
79015470 by Takanori Nakane at 2018-03-09T23:13:38+00:00
Fixed SEGV in binning. (see code comments for FourierTransformer pitfall)
- - - - -
b466bc29 by Takanori Nakane at 2018-03-10T11:09:30+00:00
Fixed binning
- - - - -
ccd600bb by Takanori Nakane at 2018-03-10T19:24:21+00:00
Convert Unblur trajectory from angstrom to pixel
PLot local trajectories from MotionCor2
- - - - -
3d61e256 by Jasenko Zivanov at 2018-03-11T03:54:47+00:00
bugfix in motion_fit; added option for unregularized global motion and distance threshold for divergence term
- - - - -
6fc19690 by Jasenko Zivanov at 2018-03-11T20:09:28+00:00
added support for custom frame ranges; input dose is now per-frame
- - - - -
ad1d3a98 by Jasenko Zivanov at 2018-03-12T01:21:38+00:00
tilt_fit now writes out a *_particles.star file with the estimated tilt values
- - - - -
aa731177 by Jasenko Zivanov at 2018-03-12T02:21:34+00:00
bugfix in recombine
- - - - -
0cd95474 by Jasenko Zivanov at 2018-03-12T03:02:27+00:00
bugfix in reconstruct_openmp
- - - - -
862a4a15 by Sjors Scheres at 2018-03-12T12:28:51+00:00
allow reading/writing backup_selection.star with information for liyis project
- - - - -
f35762f4 by Sjors Scheres at 2018-03-12T12:29:10+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
6b99cf73 by Sjors Scheres at 2018-03-12T12:29:58+00:00
allow rescaling output 3d maps in flex analyser
- - - - -
690ba830 by Takanori Nakane at 2018-03-12T12:45:31+00:00
relion_postprocess: added a hidden option --seed for testing
- - - - -
80068bd8 by Takanori Nakane at 2018-03-12T12:46:20+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
84e6b068 by Takanori Nakane at 2018-03-12T13:53:04+00:00
relion_image_handler: COM calculation moved to --com from --stats
- - - - -
9de2d214 by Takanori Nakane at 2018-03-12T22:48:14+00:00
Removed debug output
- - - - -
53cf0f3e by Jasenko Zivanov at 2018-03-12T23:11:31+00:00
preliminary merge
- - - - -
62f68d73 by Jasenko Zivanov at 2018-03-12T23:11:35+00:00
Merge remote-tracking branch 'origin/tiff' into jzivanov
- - - - -
eb8f8a4f by Jasenko Zivanov at 2018-03-13T01:03:35+00:00
added support for corrected_micrographs.star to refinement_program
- - - - -
4bff2eea by Jasenko Zivanov at 2018-03-13T03:19:34+00:00
bugfix in motion_fit
- - - - -
93054667 by Sjors Scheres at 2018-03-13T11:07:54+00:00
Merged tcblab/relion-devel into master
- - - - -
724d06d4 by Sjors Scheres at 2018-03-13T11:08:23+00:00
Merged in scheres/relion-devel-lmb (pull request #144)
Master
- - - - -
04786569 by Jasenko Zivanov at 2018-03-13T16:03:47+00:00
bugfix in refinement_program
- - - - -
74ad5e63 by Jasenko Zivanov at 2018-03-13T17:47:00+00:00
Cherry-pick Jasenko's Micrograph class refactoring
Conflicts:
src/micrograph_model.cpp
src/micrograph_model.h
- - - - -
01ebe42b by Jasenko Zivanov at 2018-03-13T17:59:27+00:00
Cherry-pick Jasenko's Micrograph class refactoring (2)
- - - - -
b88e1ae9 by Jasenko Zivanov at 2018-03-14T06:43:39+00:00
added (center-weight) prior to initial per-particle offset in motion_fit; this prevents confusing particles with better visible neighbors
- - - - -
0762c909 by Jasenko Zivanov at 2018-03-14T11:35:16+00:00
motion_fit accelerated (initial particle offsets parallelized and made optional); some vtk output replaced by general image output
- - - - -
2404953f by Jasenko Zivanov at 2018-03-14T11:37:12+00:00
missing files added
- - - - -
146acb98 by Jasenko Zivanov at 2018-03-14T12:56:34+00:00
all vtk output in the applications replaced by generalized image logging
- - - - -
dc95a898 by Takanori Nakane at 2018-03-14T17:00:20+00:00
relion_display: Initialize do_remove_duplicates to false
- - - - -
edeb1fd3 by Takanori Nakane at 2018-03-14T17:12:25+00:00
Merge branch 'master' into tiff
- - - - -
e2acd955 by Sjors Scheres at 2018-03-15T17:16:54+00:00
repair bug with MDbackup in displayer
- - - - -
f034a91f by Sjors Scheres at 2018-03-15T17:17:18+00:00
always perform 2D classification with pad 2
- - - - -
cbb85295 by Sjors Scheres at 2018-03-15T17:17:44+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
1693f353 by Jasenko Zivanov at 2018-03-16T08:42:30+00:00
motion_fit heavily refactored and parameter estimation option added
- - - - -
c6d98983 by Jasenko Zivanov at 2018-03-16T09:16:35+00:00
motion_fit parameter estimation improved
- - - - -
8ef592b8 by Jasenko Zivanov at 2018-03-16T13:58:48+00:00
motion_fit parameter estimation: quadratic extrapolation of optimum added
- - - - -
4193be69 by Sjors Scheres at 2018-03-16T14:49:44+00:00
allow multiple color selections in multiviewer display
- - - - -
5fa5ca12 by Sjors Scheres at 2018-03-16T14:50:44+00:00
changed position set selection type in popup menu
- - - - -
8b2e4d2d by Sjors Scheres at 2018-03-16T15:35:34+00:00
dont need to specify fn_imgs or fn_parts for loading of backup_selection.star
- - - - -
76ba73ef by Jasenko Zivanov at 2018-03-16T19:27:46+00:00
motion_fit parameter estimation: bugfix in quadratic extrapolation
- - - - -
3ccfed4d by Takanori Nakane at 2018-03-16T20:37:29+00:00
Merge branch 'master' into tiff
- - - - -
82c86b32 by Takanori Nakane at 2018-03-16T20:48:56+00:00
motioncorr_runner: always treat .mrc files as .mrcs files
- - - - -
a0240cb2 by Takanori Nakane at 2018-03-16T20:58:22+00:00
Fixed file handle leaks for TIFF
- - - - -
1ed74c86 by Jasenko Zivanov at 2018-03-17T15:10:08+00:00
progressive 3x3 grid search implemented for motion parameter estimation
- - - - -
4f46b9e7 by Jasenko Zivanov at 2018-03-17T15:32:42+00:00
motion parameter estimation console output improved
- - - - -
39df208e by Takanori Nakane at 2018-03-17T15:50:50+00:00
Avoid crash in very bad movies. WIP: need to think about more robust way.
- - - - -
ba78eadc by Jasenko Zivanov at 2018-03-17T15:52:33+00:00
bugfix in motion parameter estimation
- - - - -
ce7a345e by Jasenko Zivanov at 2018-03-17T21:20:41+00:00
reconstruct_openmp now survives missing images
- - - - -
6ec75770 by Jasenko Zivanov at 2018-03-18T07:01:22+00:00
motion_stats added
- - - - -
99dbe406 by Jasenko Zivanov at 2018-03-18T07:02:39+00:00
motion_stats really added
- - - - -
e3772262 by Jasenko Zivanov at 2018-03-18T07:17:40+00:00
motion_fit can deal with non-positive parameters now
- - - - -
1ca4201f by Jasenko Zivanov at 2018-03-18T07:33:01+00:00
bugfixes in motion_stats and recombine
- - - - -
c13ad5e7 by Takanori Nakane at 2018-03-18T10:28:17+00:00
--use_own does not save movies.
- - - - -
fc0b523b by Sjors Scheres at 2018-03-19T10:43:54+00:00
Merged tcblab/relion-devel into master
- - - - -
34cc75c2 by Sjors Scheres at 2018-03-19T10:44:38+00:00
Merged in scheres/relion-devel-lmb (pull request #145)
Master
- - - - -
49caf151 by Sjors Scheres at 2018-03-19T14:24:18+00:00
add command-line usage
- - - - -
253353db by Sjors Scheres at 2018-03-19T14:24:41+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
b77940ff by Sjors Scheres at 2018-03-19T16:40:16+00:00
added wait_before to pipeliner and usage function in CL parser
- - - - -
7e53011b by Takanori Nakane at 2018-03-19T17:05:18+00:00
Fix bug in interpreting links in the "browse" button.
Previously, a link from "mask.mrc" to "../somewhere/mask.mrc" was expanded to "somewhere/mask.mrcmask.mrc".
Now the link remains intact.
- - - - -
0a0636a6 by Takanori Nakane at 2018-03-19T17:08:20+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
0c713c9b by Sjors Scheres at 2018-03-19T17:26:05+00:00
debugged new --wait_before option in pipeliner
- - - - -
3a6029fa by Sjors Scheres at 2018-03-19T17:30:14+00:00
added --check_job_completion option to pipeliner
- - - - -
f10f61cc by Sjors Scheres at 2018-03-19T17:32:53+00:00
added --check_job_completion option to pipeliner
- - - - -
b8ecb372 by Sjors Scheres at 2018-03-19T17:33:51+00:00
Merged in scheres/relion-devel-lmb (pull request #146)
Master
- - - - -
e8c231e2 by Sjors Scheres at 2018-03-20T12:46:02+00:00
Merged tcblab/relion-devel into master
- - - - -
9dd67689 by Sjors Scheres at 2018-03-20T12:53:50+00:00
resolved conflicts upon merging in jzivanov branch
- - - - -
99df682f by Sjors Scheres at 2018-03-20T12:54:12+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
a9eebf48 by Sjors Scheres at 2018-03-20T13:02:16+00:00
intoduced underscore in EMDL_IMAGE_SIZE_X Y and Z
- - - - -
0253bfce by Takanori Nakane at 2018-03-20T14:01:36+00:00
Merge branch 'master' into tiff
- - - - -
074fc027 by Takanori Nakane at 2018-03-20T14:34:01+00:00
Downgrade from c++11 to c++0x for compatibility with gcc 4.4.7 (in e.g. Scientific Linux 6.8)
- - - - -
144f8250 by Takanori Nakane at 2018-03-20T14:39:38+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
c463efca by Takanori Nakane at 2018-03-20T14:40:08+00:00
Merge branch 'master' into tiff
- - - - -
567e6a08 by Sjors Scheres at 2018-03-20T16:41:01+00:00
made MPI version of reconstruct and moved code to libraries, also removed debug statement from backprojector
- - - - -
1b1b6a97 by Sjors Scheres at 2018-03-20T16:41:41+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
47e30942 by Takanori Nakane at 2018-03-20T16:55:58+00:00
add_compile_options(-std=c++0x) is not necessary as we use CCFLAGS and CXXFLAGS.
- - - - -
87844352 by Sjors Scheres at 2018-03-21T09:13:44+00:00
removed --j option from reconstruct, now mpi parallelisation only
- - - - -
b8db72c7 by Sjors Scheres at 2018-03-21T09:55:27+00:00
removed annoying version print from all programs, replaced with --version argument in any program using the IOParser
- - - - -
911a6635 by Sjors Scheres at 2018-03-21T10:26:36+00:00
switch --reverse_curvature, now default should be mostly correct
- - - - -
906784a2 by Sjors Scheres at 2018-03-21T10:26:57+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
12407d81 by Takanori Nakane at 2018-03-21T11:45:59+00:00
Change the color of the global trajectory to blue
- - - - -
3c788c56 by Takanori Nakane at 2018-03-21T11:46:24+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
9d5eb869 by Jasenko Zivanov at 2018-03-21T14:53:33+00:00
motion_fit: Gauss process regression implemented
- - - - -
2a90d3a8 by Jasenko Zivanov at 2018-03-21T14:55:19+00:00
svd_helper added
- - - - -
385bcf41 by Jasenko Zivanov at 2018-03-21T16:19:35+00:00
bugfix in refinement_program
- - - - -
bed06d47 by Jasenko Zivanov at 2018-03-21T21:18:20+00:00
exp-kernel option added to GP-motion regression; acceleration term adapted to new model
- - - - -
b60cea3e by Jasenko Zivanov at 2018-03-22T14:15:23+00:00
Tracks are now stored in .star files named after micrographs (including path); FCCs stored per-micrograph and summed up by recombine_frames
- - - - -
1e860e0a by Jasenko Zivanov at 2018-03-22T15:50:10+00:00
Bfactors are written out as .star files by reconstruct_frames and can be loaded from there (instead of computing them from FCCs); reconstruct_frames recreates the directory structure given by particle .star files, like motion fit
- - - - -
8ec5d08a by Jasenko Zivanov at 2018-03-22T17:21:18+00:00
The corrected_micrographs.star file can have any name now; recombine_frames now writes out a star file with exchanged image filenames (particles.star)
- - - - -
1ada12f0 by Sjors Scheres at 2018-03-22T17:23:39+00:00
writing additional info to postprocess.star for handover to Jasenkos programs
- - - - -
aef14c07 by Sjors Scheres at 2018-03-22T17:23:56+00:00
writing additional info to postprocess.star for handover to Jasenkos programs
- - - - -
9b2a6f08 by Sjors Scheres at 2018-03-22T17:24:56+00:00
WIP: initial attempt at jobtypes for Jasenkos programs, not functional yet
- - - - -
26b9d69c by Sjors Scheres at 2018-03-22T17:25:53+00:00
WIP: implemented latest version including Csfit from Jasenkos defocus_fit into ctf_refine, also made --only_do_unfinished and MPI version. testing pending...
- - - - -
ae7c2afd by Sjors Scheres at 2018-03-22T17:26:05+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
4b5e47ae by Sjors Scheres at 2018-03-22T17:33:21+00:00
divide data set by rlnMicrographName, not the StackName of rlnImageName, as the latter may be one for the entire dataset, for example after subtraction
- - - - -
fe5eba0b by Sjors Scheres at 2018-03-22T17:33:37+00:00
make sure output directory exists
- - - - -
6c11fa56 by Jasenko Zivanov at 2018-03-22T21:49:06+00:00
initial trajectories read from metadata file - work in progress
- - - - -
a940d6a3 by Takanori Nakane at 2018-03-22T22:39:48+00:00
Fixed Micrograph::read().
- - - - -
16bf5f8b by Takanori Nakane at 2018-03-22T23:06:42+00:00
Further fix for Micrograph::read().
Thanks Jasenko for finding this.
- - - - -
2b8bb968 by Jasenko Zivanov at 2018-03-23T01:59:08+00:00
head reattached; two files lost
- - - - -
59ea13ee by Jasenko Zivanov at 2018-03-23T02:05:32+00:00
mrc2vtk restored
- - - - -
e6ef3a26 by Takanori Nakane at 2018-03-23T02:22:19+00:00
cherry-picked a940d6a
- - - - -
aaee5e74 by Jasenko Zivanov at 2018-03-23T02:24:36+00:00
simulated cherry-picking of 16bf5f8
- - - - -
f1c4e4e5 by Takanori Nakane at 2018-03-23T16:18:16+00:00
Fixed ThirdOrderPolynomialModel::read(). Thanks again to Jasenko for pointing this out.
- - - - -
dfd27a71 by Jasenko Zivanov at 2018-03-23T18:00:13+00:00
memory-corrupting bug fixed in micrograph_model; derivatives corrected in gp_optimization; slice-by-slice loading of micrographs now optional
- - - - -
867176de by Jasenko Zivanov at 2018-03-23T18:19:03+00:00
parameter optimization no longer crashes on fewer than two particles
- - - - -
2944b99a by Sjors Scheres at 2018-03-26T11:57:54+01:00
debugging and further development of ctf_refine
- - - - -
c5fecf2e by Sjors Scheres at 2018-03-26T12:34:40+01:00
further debugging and improvement of ctf_refine
- - - - -
822461f1 by Sjors Scheres at 2018-03-26T12:35:01+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
21fcb407 by Takanori Nakane at 2018-03-26T12:48:05+01:00
Added "Use exhaustive search" option in the GUI for CTFFIND4.
This is "No" by default because it is > 10 times faster and CTFFIND 4.1.10 fixed a bug in the fast search.
- - - - -
6918df93 by Takanori Nakane at 2018-03-26T12:50:10+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
5d871399 by Jasenko Zivanov at 2018-03-26T13:44:16+01:00
pre-merge commit
- - - - -
44379fdf by Jasenko Zivanov at 2018-03-26T14:18:02+01:00
merged with master
- - - - -
a78ec972 by Sjors Scheres at 2018-03-26T15:54:16+01:00
kmin for both defocus and tilt fit
- - - - -
64b3f542 by Sjors Scheres at 2018-03-26T15:54:32+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
c957bdac by Sjors Scheres at 2018-03-26T14:58:25+00:00
Merged in jzivanov (pull request #10)
Jzivanov
- - - - -
73659912 by Sjors Scheres at 2018-03-26T15:58:39+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
c78c2247 by Sjors Scheres at 2018-03-26T16:08:15+01:00
removed duplicate kmin definition
- - - - -
a15cb5ad by Jasenko Zivanov at 2018-03-26T19:35:50+01:00
motion_fit now allows for the estimation of all 3 parameters; parameter estimation is now recursive
- - - - -
501dd74c by Jasenko Zivanov at 2018-03-26T20:46:16+01:00
motion_fit parameter estimation deals with zero values more gracefully
- - - - -
7ad60760 by Jasenko Zivanov at 2018-03-27T02:06:38+01:00
Single precision support added
- - - - -
0d0b3a64 by Sjors Scheres at 2018-03-27T13:29:48+00:00
Merged in jzivanov (pull request #11)
Jzivanov
- - - - -
8cf7e486 by Takanori Nakane at 2018-03-28T11:44:10+01:00
Merge branch 'master' into tiff
- - - - -
6e71b037 by Takanori Nakane at 2018-03-28T14:20:21+01:00
getShiftAt(): clarified the meaning of x & y
- - - - -
1ea50c17 by Takanori Nakane at 2018-03-28T14:54:47+01:00
Disable local motion correction when the fit is too bad.
- - - - -
3aed2adb by Jasenko Zivanov at 2018-03-28T15:10:00+01:00
Particle origins are now considered during movie extraction
- - - - -
883fc761 by Sjors Scheres at 2018-03-28T15:24:17+01:00
cleaned up ctf_refinement a bit
- - - - -
46bc9922 by Sjors Scheres at 2018-03-28T15:24:47+01:00
changed order grey and black lines in EPS plot
- - - - -
b4f6cb36 by Sjors Scheres at 2018-03-28T15:26:16+01:00
added motion_fitter library and motion_refine (mpi) programs
- - - - -
ccd9fe8c by Takanori Nakane at 2018-03-28T15:59:14+01:00
Make 'preread images' and 'use scratch' mutually exclusive.
- - - - -
780d56e4 by Takanori Nakane at 2018-03-28T15:59:57+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
f726ed31 by Sjors Scheres at 2018-03-28T16:00:58+01:00
repaired bug in handling micrograph filenames inside pipeline
- - - - -
eb27b9d2 by Sjors Scheres at 2018-03-28T16:01:15+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
d2f9158a by Takanori Nakane at 2018-03-28T16:17:04+01:00
Tidy up the Motioncor2 tab in the GUI
- - - - -
cdb03926 by Takanori Nakane at 2018-03-28T16:19:08+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
10371249 by Jasenko Zivanov at 2018-03-28T16:25:13+01:00
Parameter search can now deal with negative initial parameters; offsets removed from extraction again
- - - - -
4c122e70 by Sjors Scheres at 2018-03-28T16:39:26+01:00
small changes to GUI for motioncorr runner
- - - - -
a04de3a3 by Sjors Scheres at 2018-03-28T16:39:42+01:00
write out combined PDF logfile
- - - - -
1c7ad9cb by Takanori Nakane at 2018-03-28T17:11:25+01:00
Merge branch 'tiff'
- - - - -
ac004b06 by Jasenko Zivanov at 2018-03-28T17:11:51+01:00
GP-based motion optimization parallelized
- - - - -
2d4eee64 by Jasenko Zivanov at 2018-03-28T17:56:32+01:00
motion parameter estimation now returns a sensible value even if it aborts because the max. number of iterations has been reached
- - - - -
9ddfedf1 by Jasenko Zivanov at 2018-03-28T17:57:59+01:00
Merge remote-tracking branch 'origin/master' into jzivanov
- - - - -
9afd23b4 by Jasenko Zivanov at 2018-03-29T01:54:09+01:00
Accelerated parameter optimization implemented (not --exact_params, --mot_range); Added option to set max. nr. of eigendeformations (--max_ed)
- - - - -
9c837227 by Jasenko Zivanov at 2018-03-29T01:55:17+01:00
missing files added
- - - - -
699c88e4 by Jasenko Zivanov at 2018-03-29T02:16:50+01:00
number of eigendeformations can be set to the number of particles by setting --max_ed to a negative value
- - - - -
5c6911d1 by Sjors Scheres at 2018-03-29T12:05:42+01:00
added combine_frames to motion fitter
- - - - -
d39fd4d8 by Sjors Scheres at 2018-03-29T12:06:00+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
49215602 by Jasenko Zivanov at 2018-03-29T13:37:31+01:00
bugfix in 3-parameter-estimation; motion optimization further parallelized
- - - - -
96a1b0bc by Jasenko Zivanov at 2018-03-29T14:26:01+01:00
bugfix in 3-parameter-estimation
- - - - -
3543b467 by Sjors Scheres at 2018-03-29T15:54:26+01:00
debugging combination of combine_frames into motion_refine
- - - - -
2924d10f by Sjors Scheres at 2018-03-29T16:27:15+01:00
write RELION_VERSION as comment to STAR files
- - - - -
669eb3e6 by Jasenko Zivanov at 2018-03-30T01:43:26+01:00
bugfixes in motion estimation incl. GP-model derivatives (slower now due to unhindered convergence - 1000 iterations appears sufficient); motion model accelerated (~4x speedup per iteration); optimization interface (and all implementations) changed to accommodate external temporary storage; minor bugfix in strings.cpp
- - - - -
74618c32 by Jasenko Zivanov at 2018-03-30T18:41:05+01:00
libLBFGS added to jaz/lbfgs; optimization code moved to jaz/optimization; motion fit now performed using LBFGS (in motion_fit)
- - - - -
c03de8a3 by Jasenko Zivanov at 2018-03-30T19:33:34+01:00
LBFGS motion_fit parameters exposed and optimized
- - - - -
d8d67332 by Jasenko Zivanov at 2018-03-30T20:09:08+01:00
damage model is now being considered by TSC evaluation
- - - - -
491a57d1 by Jasenko Zivanov at 2018-03-30T20:12:44+01:00
tiny bugfix in parameter estimation printout
- - - - -
1349384c by Jasenko Zivanov at 2018-03-30T20:46:48+01:00
Merge remote-tracking branch 'origin/master' into jzivanov
- - - - -
ddce7da5 by Jasenko Zivanov at 2018-03-30T22:26:02+01:00
motion_refine now works on Relion2 datasets (hasCorrMic was uninitialized, --mov and --mov_replace work now); motion_refine switched to LBFGS; support for a max. number of eigendeformations added; motion_fitter code sanitized
- - - - -
52377805 by Jasenko Zivanov at 2018-03-31T00:44:32+01:00
control flow bug fixed in motion parameter estimation
- - - - -
47e45b95 by Jasenko Zivanov at 2018-03-31T12:15:38+01:00
motion_fitter prepared for parameter estimation
- - - - -
3b4deaeb by Jasenko Zivanov at 2018-03-31T12:45:49+01:00
motion_fitter renamed to motion_refiner, since it does two other things now
- - - - -
0dc245fa by Jasenko Zivanov at 2018-03-31T12:46:19+01:00
missing files added
- - - - -
16a02a0a by Jasenko Zivanov at 2018-03-31T12:57:58+01:00
motion-related code moved to jaz/motion
- - - - -
4329d872 by Jasenko Zivanov at 2018-03-31T14:04:29+01:00
motion_refinement class renamed to motion_helper and reduced to necessary methods; some deprecated code moved jaz/archive (not being compiled)
- - - - -
e6735e5f by Jasenko Zivanov at 2018-03-31T14:05:02+01:00
missing files added
- - - - -
1b33349f by Jasenko Zivanov at 2018-03-31T16:01:59+01:00
motion estimation functionality from MotionRefiner factored out into MotionEstimator
- - - - -
fe8c171e by Jasenko Zivanov at 2018-03-31T18:17:01+01:00
MotionRefiner refactored further
- - - - -
46d6bacd by Jasenko Zivanov at 2018-03-31T20:07:18+01:00
frame recombination functionality from MotionRefiner factored out into FrameRecombiner
- - - - -
59809dae by Jasenko Zivanov at 2018-03-31T20:09:53+01:00
missing files added
- - - - -
0d6b2eb5 by Jasenko Zivanov at 2018-03-31T20:55:43+01:00
preemptive bugfix in FrameRecombiner
- - - - -
de3a310c by Jasenko Zivanov at 2018-03-31T20:57:48+01:00
unsaved files updated
- - - - -
99fdefa3 by Jasenko Zivanov at 2018-03-31T23:50:11+01:00
motion_refinement further refactored and debugged
- - - - -
c5a2178d by Jasenko Zivanov at 2018-03-31T23:54:36+01:00
tiny bugfix in motion_refinement
- - - - -
71168650 by Jasenko Zivanov at 2018-04-01T04:25:29+01:00
bugfixes in motion_refine and motion_fit; results consistent again
- - - - -
ad394c15 by Jasenko Zivanov at 2018-04-01T19:36:30+01:00
preemptive bugfix in obs_model
- - - - -
054acf1f by Jasenko Zivanov at 2018-04-01T21:48:26+01:00
freq. cutoff for motion estimation at .143 FSC now optional; bugfix im motion parameter estimation
- - - - -
63f03583 by Jasenko Zivanov at 2018-04-02T00:19:08+01:00
bugfix im motion parameter estimation completed
- - - - -
e8f8a674 by Jasenko Zivanov at 2018-04-02T15:17:29+01:00
functionality for the loading of micrographs and initial tracks factored out of MotionRefiner into MicrographHandler
- - - - -
167b9e33 by Jasenko Zivanov at 2018-04-02T15:54:33+01:00
MicrographHandler now robust against metadata-star-files missing in corrected_micrographs.star (it throws an exception now instead of returning unallocated memory)
- - - - -
c24f01f2 by Jasenko Zivanov at 2018-04-02T20:33:57+01:00
reference map handling from MotionRefiner factored out into ReferenceMap
- - - - -
b8d11248 by Jasenko Zivanov at 2018-04-03T01:55:07+01:00
motion package further refactored; MotionParamEstimator, MotionEstimator and FrameRecombiner no longer (cyclically) dependent on MotionRefiner; all public fields removed
- - - - -
c35c534e by Jasenko Zivanov at 2018-04-03T02:48:07+01:00
motion_refine now survives one-particle micrographs
- - - - -
fa1f6291 by Jasenko Zivanov at 2018-04-03T20:20:35+01:00
parameter estimation implemented in motion_refine
- - - - -
da855ba2 by Jasenko Zivanov at 2018-04-04T01:20:20+01:00
motion_refine now supports --first/last_frame for relion3 datasets as well (i.e. polynomial initial tracks)
- - - - -
a29b82cc by Jasenko Zivanov at 2018-04-04T02:09:01+01:00
motion_refine: initial tracks are now returned exactly the same if --max_iters is set to zero
- - - - -
990fb5ef by Jasenko Zivanov at 2018-04-04T04:06:27+01:00
experimental Nelder-Mead-based hyperparameter optimization implemented in MotionParamEstimator
- - - - -
8809931e by Jasenko Zivanov at 2018-04-04T16:45:55+01:00
missing files added
- - - - -
3dfaa542 by Jasenko Zivanov at 2018-04-04T19:06:45+01:00
MicrographHandler now determines the max. safe number of frames in the chosen part of the dataset before loading any movies
- - - - -
32ee3340 by Jasenko Zivanov at 2018-04-04T22:01:02+01:00
motion_refine now supports the --find_shortest_movie option. If set, only the frame range available in all movies will be loaded. If not set, all movies with an insufficient number of frames will be determined beforehand and ignored (the user will be warned)
- - - - -
53e509d1 by Jasenko Zivanov at 2018-04-05T03:14:09+01:00
motion hyper-parameter optimization now fully transitioned to the Nelder-Mead uphill-simplex algorithm; definition of sigma_div no longer includes the dose; single-exponential kernel is now the default
- - - - -
8f4459f1 by Jasenko Zivanov at 2018-04-05T03:45:55+01:00
value of ThreeHyperParameterProblem::accThresh tweaked
- - - - -
5ffe8859 by Jasenko Zivanov at 2018-04-05T03:51:51+01:00
output of parameter optimization clarified
- - - - -
d9bd1de9 by Jasenko Zivanov at 2018-04-05T04:05:42+01:00
ParameterEstimator: sig_acc now clamped to 1/eps
- - - - -
eda9c594 by Jasenko Zivanov at 2018-04-05T18:45:21+01:00
ParameterEstimator now supports separate frequency thresholds for alignment and evaluation, allowing for a gap in-between
- - - - -
673ef2be by Jasenko Zivanov at 2018-04-06T01:45:07+01:00
more debugging annotations in motion_refine
- - - - -
85b616ef by Jasenko Zivanov at 2018-04-06T02:52:24+01:00
fixed memory leak in LibLbfgsAdapter
- - - - -
e89d9998 by Jasenko Zivanov at 2018-04-07T04:32:11+01:00
memory leak fixed in MotionModel; memory footprint reduced for ReferenceMap
- - - - -
5a1f3eb3 by Jasenko Zivanov at 2018-04-07T05:02:27+01:00
added --seed option to motion parameter estimation
- - - - -
4bbdbfb2 by Jasenko Zivanov at 2018-04-07T10:32:11+01:00
MotionParamEstimator: optimized allocation order of alignment data to minimize heap fragmentation
- - - - -
b1e2af4e by Takanori Nakane at 2018-04-07T18:39:00+01:00
Removed duplicated 'scratch_dir' in Class3D.
- - - - -
d0101a75 by Jasenko Zivanov at 2018-04-08T10:57:53+01:00
New, experimental FFTW wrapper implemented in jaz/new_ft
- - - - -
57e31b52 by Jasenko Zivanov at 2018-04-09T03:47:18+01:00
AlignmentSet now a template; used with floats in MotionParamEstimator now
- - - - -
16531fbe by Takanori Nakane at 2018-04-10T19:52:18+01:00
Crop cross-correlation map (WIP: still testing)
- - - - -
f4bfcddd by Takanori Nakane at 2018-04-11T12:43:23+01:00
CCF downsampling tested on two datasets. Debug statements removed.
- - - - -
c1909678 by Jasenko Zivanov at 2018-04-11T21:57:33+01:00
potential memory-alignment bug fixed in new fftw-wrapper; --no_glob_off parameter replaced by --glob_off
- - - - -
b3bdd6ef by Takanori Nakane at 2018-04-12T17:27:18+01:00
Tidy up binning
- - - - -
c76414dc by Jasenko Zivanov at 2018-04-13T10:51:19+01:00
Pitfall disarmed in BackProjector::reconstruct
- - - - -
087a11d4 by Jasenko Zivanov at 2018-04-14T15:34:28+01:00
added experimental double-reconstruction program
- - - - -
784b7744 by Jasenko Zivanov at 2018-04-14T16:34:14+01:00
debugging in double-reconstruction program
- - - - -
c1b6f9e6 by Jasenko Zivanov at 2018-04-14T16:41:48+01:00
include statements slightly pruned
- - - - -
d1c24538 by Jasenko Zivanov at 2018-04-14T18:29:48+01:00
double-reconstruction program debugged
- - - - -
497d2894 by Jasenko Zivanov at 2018-04-15T15:50:14+01:00
double_reconstruct now only writes out debugging data if --debug is set to a non-empty string
- - - - -
c910012e by Jasenko Zivanov at 2018-04-15T15:53:28+01:00
double_reconstruct now reads the .star file even if --debug is set
- - - - -
817fb010 by Jasenko Zivanov at 2018-04-15T15:57:32+01:00
progress indicator debugged in double_reconstruct
- - - - -
08eab721 by Jasenko Zivanov at 2018-04-15T16:59:54+01:00
removed resolution-rounding from postprocessing
- - - - -
5b1071f1 by Jasenko Zivanov at 2018-04-15T17:25:13+01:00
added an option to divide the per-voxel average by its weight prior to computing the preliminary FSC (which is needed for regularization)
- - - - -
20657149 by Jasenko Zivanov at 2018-04-15T17:35:38+01:00
double_reconstruct progress indicator further debugged
- - - - -
ee9e19c0 by Jasenko Zivanov at 2018-04-16T13:19:13+01:00
experimental L1-averaging added to double_reconstruct
- - - - -
60c9415b by Jasenko Zivanov at 2018-04-16T13:35:36+01:00
altered weight-formulation for L1-averaging in double_reconstruct
- - - - -
d9b46f01 by Jasenko Zivanov at 2018-04-16T14:20:07+01:00
bugfix in L1-averaging
- - - - -
c2e00645 by Jasenko Zivanov at 2018-04-16T18:30:13+01:00
bugfix in FourierTransformer: in setReal(MultidimArray<RFLOAT>), the plans are now also recomputed if the *complex* array has changed in the meantime (only the location, not the size is being compared)
- - - - -
c02bf7a1 by Sjors Scheres at 2018-04-17T11:23:59+00:00
Merged in jzivanov (pull request #12)
Jzivanov
- - - - -
056ace5d by Sjors Scheres at 2018-04-17T15:10:08+01:00
removed some old program files
- - - - -
f7cb2182 by Sjors Scheres at 2018-04-17T15:13:01+01:00
small error in help text
- - - - -
ce9a4ee8 by Sjors Scheres at 2018-04-17T15:13:20+01:00
added help text for first and last frame
- - - - -
0016402f by Takanori Nakane at 2018-04-17T17:09:23+01:00
Early binning in motion correction to save memory (but not used by default).
- - - - -
76379592 by Sjors Scheres at 2018-04-17T17:41:08+01:00
removed UNBLUR wrapper, movie refinement and particle polishing from GUI, and updated the pipeliner for motion_refine jobs
- - - - -
5762b398 by Dari Kimanius at 2018-04-17T20:27:19+02:00
Merged tcblab/relion-devel into master
- - - - -
c5b8729a by Dari Kimanius at 2018-04-17T21:28:09+02:00
Fix TIFF-related compile issues
- - - - -
734780c4 by Jasenko Zivanov at 2018-04-17T20:32:23+01:00
minor changes made at the LMB
- - - - -
4f92ec5f by Jasenko Zivanov at 2018-04-17T22:11:05+01:00
single precision build compiles and links again
- - - - -
635845b2 by Sjors Scheres at 2018-04-18T09:07:52+01:00
adjusted motion_refine command line defaults as per jazsenko
- - - - -
e179db00 by Jasenko Zivanov at 2018-04-18T15:03:50+01:00
bugfix in NewFFT; NewFFT::XPlan now knows whether it is reusable and it answers isCompatible() accordingly
- - - - -
9a686c7a by Jasenko Zivanov at 2018-04-18T17:16:14+01:00
NewFFT: output array for on-the-fly transforms is rescaled to the correct size (if necessary) prior to the computation of the plan
- - - - -
49a2dd64 by Jasenko Zivanov at 2018-04-18T17:39:20+01:00
NewFFT further refactored: both on-the-fly and reusable-plan methods perform the required size-checks and make an optional copy of the input array, then they both call the same private methods that perform the actual transform and normalization.
- - - - -
bb3dd08f by Jasenko Zivanov at 2018-04-18T18:37:59+01:00
ft_test app added
- - - - -
3c92a255 by Sjors Scheres at 2018-04-19T10:46:52+01:00
added keep_tilt_prior_fixed on GUI
- - - - -
576f02fe by Sjors Scheres at 2018-04-19T10:47:37+01:00
allow keep_tilt_prior_fixed and local averaging in helices, also repaired bug in NrOrientations for skip_align
- - - - -
6338dde1 by Sjors Scheres at 2018-04-19T10:50:55+01:00
setting version to 3.0-alpha-3
- - - - -
f75c7b9b by Sjors Scheres at 2018-04-19T09:51:50+00:00
Merged in scheres/relion-devel-lmb (pull request #147)
relion3.0-alpha-3
- - - - -
fa8dfe12 by Sjors Scheres at 2018-04-19T10:52:57+01:00
Merged tcblab/relion-devel into master
- - - - -
7e3f99cd by Sjors Scheres at 2018-04-19T11:31:14+01:00
dynamic library linking, otherwise ftlk gives errors..
- - - - -
3950b63c by Public at 2018-04-19T11:40:39+01:00
dynamic librray linking
- - - - -
5182d721 by Sjors Scheres at 2018-04-19T10:42:17+00:00
Merged in scheres/relion-devel-lmb (pull request #148)
dynamic librray linking
- - - - -
8511700d by Sjors Scheres at 2018-04-19T11:42:41+01:00
Merged tcblab/relion-devel into master
- - - - -
95856e24 by Jasenko Zivanov at 2018-04-19T14:04:52+01:00
bugfix in newFT: plan creation for rectangular images corrected
- - - - -
522e8e9b by Sjors Scheres at 2018-04-19T16:05:16+01:00
tau2 can be both input or ourpur. repaired bug introduced by jasenko a few days ago
- - - - -
b454472b by Sjors Scheres at 2018-04-19T16:06:37+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
64030743 by Jasenko Zivanov at 2018-04-19T16:10:19+01:00
motion_refine: support for --mov and --mov_end arguments reestablished for Relion 2 datasets
- - - - -
f5741cf7 by Takanori Nakane at 2018-04-19T16:10:49+01:00
Merge branch 'jzivanov' into tiff
- - - - -
ee27f792 by Sjors Scheres at 2018-04-19T15:11:31+00:00
Merged in scheres/relion-devel-lmb (pull request #149)
Master
- - - - -
b7750388 by Sjors Scheres at 2018-04-19T16:11:46+01:00
Merged tcblab/relion-devel into master
- - - - -
7c6a2d4f by Jasenko Zivanov at 2018-04-19T16:14:33+01:00
...now including the initial micrograph
- - - - -
3802506a by Sjors Scheres at 2018-04-19T16:38:44+01:00
added --ctf_multiplied to new reconstruct progra,
- - - - -
3c2a56ec by Jasenko Zivanov at 2018-04-19T16:43:28+01:00
Merge remote-tracking branch 'origin/master' into jzivanov
- - - - -
bef82638 by Jasenko Zivanov at 2018-04-19T17:14:43+01:00
merged with master; movie path parameters and support for preextracted movies removed from motion_refine
- - - - -
18c82d76 by Jasenko Zivanov at 2018-04-19T18:48:09+01:00
tau2 handling in BackProjector::reconstruct() made slightly safer; ft_test emptied
- - - - -
b7d1b715 by Bjoern Forsberg at 2018-04-20T10:37:04+02:00
Merged tcblab/relion-devel into master
- - - - -
12c35300 by Bjoern Forsberg at 2018-04-20T10:37:45+02:00
3D particles can be shown from model.star display
- - - - -
4131686b by Bjoern Forsberg at 2018-04-20T10:39:06+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
d8bb8ea6 by Bjoern Forsberg at 2018-04-20T10:58:23+02:00
Particle displ. uses the "apply orients" tickbox
- - - - -
80d19045 by Takanori Nakane at 2018-04-20T11:03:03+01:00
Always use float in motioncorr to save memory.
Confirmed EMPIAR-10153 gave the same resolution as before.
- - - - -
89f966eb by Takanori Nakane at 2018-04-20T11:04:50+01:00
Merge branch 'tiff'
- - - - -
07e09b0c by Takanori Nakane at 2018-04-20T11:20:36+01:00
Added RELION_TEST cmake flag to prevent test programs from being built.
- - - - -
e09b244d by Takanori Nakane at 2018-04-20T11:24:00+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
7c41c55f by Takanori Nakane at 2018-04-20T11:34:49+01:00
Remove unnecessary transformers.
- - - - -
e0bdd387 by Jasenko Zivanov at 2018-04-20T12:14:57+01:00
Merge remote-tracking branch 'origin/master' into jzivanov
- - - - -
cf053a60 by Jasenko Zivanov at 2018-04-20T12:33:57+01:00
motion_refine now looks for all missing movies at initialization
- - - - -
8bfe8bd2 by Takanori Nakane at 2018-04-20T12:49:27+01:00
relion_stack_create: mkdir when necessary as reported by Max Wilkinson.
- - - - -
14209148 by Jasenko Zivanov at 2018-04-20T12:58:04+01:00
It is no longer necessary to use mrc:mrcs to load stacks stored in .mrc files in motion_refine
- - - - -
936a1515 by Jasenko Zivanov at 2018-04-20T13:34:27+01:00
image stacks are now loaded slice-by-slice in ctf_refine
- - - - -
def78c66 by Jasenko Zivanov at 2018-04-20T14:47:10+01:00
debugging statements added to defocus_refinement
- - - - -
0f7bbb2a by Bjoern Forsberg at 2018-04-20T15:50:20+02:00
Acc. noise-masking runs with 3D input particles
Seems to give erraneous output though....
- - - - -
892d8bc5 by Takanori Nakane at 2018-04-20T15:11:18+01:00
More help in GUI for motion correction.
- - - - -
206c3001 by Takanori Nakane at 2018-04-20T15:44:00+01:00
Fixed motioncor GUI and added "save non-dose weighted as well?"..
- - - - -
5b4eb9c1 by Sjors Scheres at 2018-04-20T15:20:00+00:00
Merged in scheres/relion-devel-lmb (pull request #150)
Master
- - - - -
34a2ef84 by Takanori Nakane at 2018-04-20T17:50:58+01:00
Fix minimum inter-particle distance in autopicker
- - - - -
994d6f25 by Jasenko Zivanov at 2018-04-20T20:55:45+01:00
ctf_refine now uses ReferenceMap; prediction for defocus estimation now considers current tilt assumption; prediction for tilt estimation is now taken from the opposite half-set
- - - - -
35a82d0d by Jasenko Zivanov at 2018-04-20T20:56:26+01:00
missing files added
- - - - -
035cf51d by Jasenko Zivanov at 2018-04-21T00:03:48+01:00
CtfRefiner: tilt refinement factored out into TiltEstimator
- - - - -
9e4fc5bf by Jasenko Zivanov at 2018-04-21T11:55:47+01:00
tilt estimation now considers per-particle defoci
- - - - -
160fc319 by Jasenko Zivanov at 2018-04-21T15:56:32+01:00
defocus estimation factored out of CtfRefiner into DefocusEstimator; defocus and tilt estimation can now have different kmin
- - - - -
d31420e5 by Jasenko Zivanov at 2018-04-21T16:33:37+01:00
bugfix in ctf_refine
- - - - -
c2bdef4a by Jasenko Zivanov at 2018-04-21T16:36:46+01:00
ctf_refine: no longer printing out particle index on --diag
- - - - -
784a9384 by Jasenko Zivanov at 2018-04-21T17:34:34+01:00
bugfix in TiltEstimator
- - - - -
7afac3ae by Jasenko Zivanov at 2018-04-21T18:54:08+01:00
bugfix in defocus estimation; diagnostic output improved
- - - - -
7da5d92b by Takanori Nakane at 2018-04-24T16:32:44+01:00
Fixed a bug in "combine STAR file" and tidied up indentation
- - - - -
c64674d4 by Sjors Scheres at 2018-04-25T09:05:51+01:00
Merged tcblab/relion-devel into master
- - - - -
598ea88f by Jasenko Zivanov at 2018-04-25T10:21:54+01:00
spaces at the end of logfile.pdf filenames removed
- - - - -
92a87fce by Jasenko Zivanov at 2018-04-25T12:59:32+01:00
FrameRecombiner now uses the outer freq. of the reference as the outer limit for B-factor estimation, unless a value is provided through --bfac_maxfreq - this requires loading the reference in that case; CtfRefiner slightly refactored in preparation for anisotropic mag. estimation
- - - - -
83dde660 by Jasenko Zivanov at 2018-04-25T17:12:51+01:00
tentative anisotropic magnification fit added to ctf_refine
- - - - -
1bedbb03 by Takanori Nakane at 2018-04-25T20:04:40+01:00
Removed reconstruct_openmp. Marked double_reconstuct_openmp as testing program.
- - - - -
f326aa40 by Sjors Scheres at 2018-04-26T15:07:17+01:00
write out debug maps in reconstruct
- - - - -
0ebeace9 by Takanori Nakane at 2018-04-26T15:29:37+01:00
Manual picker: treat shift + left-click as middle-click and ctrl + left-click as right-click.
This is useful for touch panel users.
Also tidied up indentations.
- - - - -
adc748b1 by Takanori Nakane at 2018-04-26T15:31:46+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
84ee51c1 by Jasenko Zivanov at 2018-04-27T12:01:19+01:00
ctf_refine now works on centered particles
- - - - -
e1ee1c67 by Jasenko Zivanov at 2018-04-27T13:10:26+01:00
ctf_refine and motion_refine now warn about missing star-file columns that are required for forward-projection
- - - - -
7aecb45a by Takanori Nakane at 2018-04-27T16:55:32+01:00
Made motioncorr more robust.
WIP: We still need better GO/NO-GO decision before fitting.
- - - - -
19f43c92 by Sjors Scheres at 2018-04-30T10:08:58+01:00
rescale originXYZ as these are in pixels
- - - - -
24c927ec by Sjors Scheres at 2018-04-30T10:09:18+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
1e5db961 by Jasenko Zivanov at 2018-05-01T02:42:01+01:00
motion_refine: B/k-factors scaleable, noise-whitening optional
- - - - -
57b1cf49 by Takanori Nakane at 2018-05-01T13:00:44+01:00
Fixed garbage in the qsub template field when the environmental variable in not set.
And tidied up indentations.
- - - - -
de8b0d7b by Takanori Nakane at 2018-05-01T13:02:05+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
39e88dc1 by Sjors Scheres at 2018-05-02T09:05:40+01:00
add unmaksed 2d classes to trainingset
- - - - -
b745f839 by Sjors Scheres at 2018-05-02T09:06:47+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
317c6570 by Dari Kimanius at 2018-05-02T17:34:02+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
de56860e by Jasenko Zivanov at 2018-05-02T21:58:30+01:00
motion_refine: initial per-particle trajectories are now aligned rigidly with the data if --glob_off is set
- - - - -
6373b95e by Jasenko Zivanov at 2018-05-02T22:23:44+01:00
motion_param_estimator: internal scale of sigma_acc increased to 10,000
- - - - -
46e6e35a by Jasenko Zivanov at 2018-05-02T23:53:22+01:00
motion_estimator: scaling of input parameters by dose now optional (--params_by_dose)
- - - - -
29bf55fc by Jasenko Zivanov at 2018-05-03T00:37:58+01:00
motion_estimator: input parameters scaled by dose by default (turned off by --absolute_params)
- - - - -
7a20387a by Jasenko Zivanov at 2018-05-03T01:11:32+01:00
motion_param_estimator: internal inversion of sigma_acc removed
- - - - -
bdcad7b1 by Jasenko Zivanov at 2018-05-03T01:25:35+01:00
motion_param_estimator: internal scale of sigma_acc reduced to 100
- - - - -
902f2c0b by Jasenko Zivanov at 2018-05-03T01:49:06+01:00
motion_param_estimator: internal parameter representation further optimized
- - - - -
f502d8d2 by Jasenko Zivanov at 2018-05-03T04:08:55+01:00
motion_param_estimator: initial step size increased to accelerate convergence
- - - - -
ab46b358 by Takanori Nakane at 2018-05-03T13:36:02+01:00
Stop if RotGain or FlipGain are given to MotionCor2.
WIP: gain rotation & flipping will be implemented in RELION.
- - - - -
14afb1b8 by Takanori Nakane at 2018-05-03T13:36:55+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
27209a41 by Takanori Nakane at 2018-05-03T14:23:50+01:00
relion_image_handler: Fixed typo.
And tidied up indentations.
- - - - -
38c6b93a by Takanori Nakane at 2018-05-03T17:03:00+01:00
Rotate and/or Flip gain within run_motioncorr.
- - - - -
a8f46151 by Jasenko Zivanov at 2018-05-03T17:22:23+01:00
motion_estimator: --unreg_glob option working again (only useful for Relion 3.0 datasets)
- - - - -
8ae32026 by Jasenko Zivanov at 2018-05-03T22:19:03+01:00
motion_estimator: more debugging data written out
- - - - -
b4bd9a7e by Takanori Nakane at 2018-05-04T17:33:25+01:00
Made gain rotation & flip as drop-down lists
- - - - -
55a1fd59 by Sjors Scheres at 2018-05-08T16:12:26+00:00
Merged in scheres/relion-devel-lmb (pull request #151)
Master
- - - - -
96c948b5 by Jasenko Zivanov at 2018-05-08T23:23:22+01:00
cubic interpolation is now numerically more robust
- - - - -
234a08ef by Sjors Scheres at 2018-05-09T12:31:54+01:00
repaired bug autopicker helices
- - - - -
c7b332c2 by Sjors Scheres at 2018-05-09T12:32:32+01:00
call jasenkos motion_refine bayesian polishing
- - - - -
71a1d9f4 by Sjors Scheres at 2018-05-09T12:33:02+01:00
repaired bug mysize when reading debug files
- - - - -
8be1b3a3 by Sjors Scheres at 2018-05-09T12:33:33+01:00
added debug option to keep debug files for reconrtuct
- - - - -
b2c98583 by Sjors Scheres at 2018-05-09T12:39:37+01:00
added do_fsc0999 option in reconstruct to avoid artefacts in some helical reconstructions
- - - - -
5d725c62 by Sjors Scheres at 2018-05-09T12:58:21+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
0c12ab79 by Sjors Scheres at 2018-05-09T17:42:15+01:00
added functionality to select micrographs or particles based on metadata value in the STAR file
- - - - -
fdc4bff6 by Jasenko Zivanov at 2018-05-09T23:41:32+01:00
--k_cut[_A] and --k_eval[_A] arguments in motion_param_estimator replaced by --align_frac and --eval_frac
- - - - -
7a149c30 by Jasenko Zivanov at 2018-05-09T23:42:36+01:00
Merge remote-tracking branch 'origin/master' into jzivanov
- - - - -
9eccc52e by Jasenko Zivanov at 2018-05-10T00:03:01+01:00
merged with master
- - - - -
fd6dc20d by Sjors Scheres at 2018-05-10T08:34:49+00:00
Merged in jzivanov (pull request #13)
Jzivanov
- - - - -
0b69e8f5 by Sjors Scheres at 2018-05-10T10:01:27+01:00
updated command-line interface to reflect changes in ctf_refine by Jasenko
- - - - -
2a109810 by Sjors Scheres at 2018-05-10T10:01:47+01:00
renamed Run now button for run
- - - - -
70880647 by Sjors Scheres at 2018-05-10T10:17:36+01:00
added Jasenkos new eval_frac and align_frac to CL IO of motion_refine GUI
- - - - -
f44fa096 by Sjors Scheres at 2018-05-10T10:19:40+01:00
removed obsolete starsplit from Jasenko
- - - - -
29c4b39d by Sjors Scheres at 2018-05-10T17:17:00+01:00
set version to 3.0-alpha-4
- - - - -
c134d67a by Sjors Scheres at 2018-05-10T16:18:08+00:00
Merged in scheres/relion-devel-lmb (pull request #152)
Master
- - - - -
a4fc3333 by Sjors Scheres at 2018-05-11T14:04:07+01:00
new kmin options for ctf_refine
- - - - -
60fe0dcc by Sjors Scheres at 2018-05-11T17:35:36+01:00
removed spurious output
- - - - -
b3781e37 by Sjors Scheres at 2018-05-11T17:35:55+01:00
added --average_all_movie_frames option
- - - - -
f0bc6652 by Sjors Scheres at 2018-05-11T17:36:36+01:00
moved warning to stderr
- - - - -
5f1970ab by Sjors Scheres at 2018-05-11T17:37:00+01:00
changed output to logfile for nr_repeat=1
- - - - -
33b5b8c5 by Sjors Scheres at 2018-05-11T16:37:41+00:00
Merged in scheres/relion-devel-lmb (pull request #153)
Master
- - - - -
07039aa1 by Sjors Scheres at 2018-05-14T09:36:15+01:00
output directory for bayesian polishing is called Polish again
- - - - -
9da06468 by Sjors Scheres at 2018-05-14T09:36:38+01:00
output to stdout instead of stderr
- - - - -
064435ad by Dari Kimanius at 2018-05-14T17:10:47+02:00
Automatic TBB download and build.
- - - - -
31aa05bf by Charles Congdon at 2018-05-14T10:07:36-07:00
Merge branch 'complex_projector'
- - - - -
b047cf06 by Sjors Scheres at 2018-05-16T09:40:16+01:00
only set ctf and beamtilt values from micrograph star file if particles input star did not contain them
- - - - -
a5678ff7 by Sjors Scheres at 2018-05-16T09:40:39+01:00
explicitly mention MotionCorr input star file on motion_refine GUI
- - - - -
2eb209a9 by Takanori Nakane at 2018-05-16T15:05:54+01:00
flex_analyser: clarified comment. fixed indentations
- - - - -
7a75e79d by Dari Kimanius at 2018-05-16T17:38:49+02:00
Fixed CPU size unallocated random image pointer.
- - - - -
c2a5c7b6 by Charles Congdon at 2018-05-16T09:46:32-07:00
Various bugfixes for the ALTCPU code
- - - - -
169fe619 by Charles Congdon at 2018-05-16T09:46:42-07:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
e576a1f4 by Charles Congdon at 2018-05-16T14:32:19-07:00
Clean up some things pulled back in by merge
- - - - -
5219e97b by Charles Congdon at 2018-05-16T14:54:23-07:00
More cleanup
- - - - -
3bffa751 by Charles Congdon at 2018-05-16T15:09:54-07:00
Fix CUDA build and isolate from ALTCPU
- - - - -
3145e60f by Dari Kimanius at 2018-05-17T17:35:06+02:00
Added FOURCE_OWN_TBB for consistency.
- - - - -
710c901c by Jasenko Zivanov at 2018-05-17T17:47:18+01:00
optimal motion parameters are now written to <out>/opt_params.txt and can be used by motion_refine with the --params_file argument (overrides cmd. line arguments)
- - - - -
77c0f2eb by Sjors Scheres at 2018-05-18T08:26:54+00:00
Merged in jzivanov (pull request #14)
optimal motion parameters are now written to <out>/opt_params.txt and can be used by motion_refine with the --params_file argument (overrides cmd. line arguments)
- - - - -
2d92ce64 by Sjors Scheres at 2018-05-18T16:03:34+01:00
repaired serious bug in preprocessing when extracting rescaled particles
- - - - -
ce845915 by Sjors Scheres at 2018-05-18T16:35:23+01:00
added option to randomise order in STAR file
- - - - -
e9c676be by Sjors Scheres at 2018-05-18T16:36:18+01:00
added relion_star_handler program, modified gui for that, and for opt_params.txt input of motion_refine
- - - - -
0811ad59 by Sjors Scheres at 2018-05-18T17:09:31+01:00
debugging new functionalities
- - - - -
317eda99 by Jasenko Zivanov at 2018-05-18T23:50:23+01:00
cc-padding, cout flushing, pixel size warnings
- - - - -
615c6289 by Sjors Scheres at 2018-05-21T15:27:24+01:00
added --randomize_at_A and unhidden --random_seed
- - - - -
8791cabe by Sjors Scheres at 2018-05-21T15:29:06+01:00
dont give error for nr_split=0 when submitting Select job
- - - - -
e7f7f0e0 by Sjors Scheres at 2018-05-21T15:29:36+01:00
added delete statements for all new CPlot2D statements
- - - - -
08a19f9e by Sjors Scheres at 2018-05-21T16:12:09+01:00
allow to set both split_size and nr_split in star_handler and Select jobtypes
- - - - -
2ae7764c by Sjors Scheres at 2018-05-21T16:42:42+01:00
added in pix to renceter options
- - - - -
160908ed by Sjors Scheres at 2018-05-21T17:14:56+01:00
added options to add and remove columns from STAR files in sta_handler
- - - - -
f3f0579c by Takanori Nakane at 2018-05-21T17:23:27+01:00
flex_analyser: write PCA weights and eigenvectors to files
- - - - -
a948447c by Takanori Nakane at 2018-05-21T17:23:54+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
3480dda3 by Takanori Nakane at 2018-05-21T21:32:32+01:00
motioncorr_runner: save angpix even when dose_weighting is not applied
- - - - -
8d4f6a6f by Takanori Nakane at 2018-05-21T21:33:11+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
7c73c967 by Sjors Scheres at 2018-05-22T11:23:40+01:00
always provide voltage and dose rate to motioncorr_runner
- - - - -
7b96aa9b by Sjors Scheres at 2018-05-22T11:23:57+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
11e9740f by Sjors Scheres at 2018-05-22T12:13:38+01:00
reversed change yesterday in Cplot2D: fn_eps[i] could be a linux wildcard
- - - - -
ef7988f4 by Sjors Scheres at 2018-05-22T12:14:56+01:00
repaired bug combine star_handler
- - - - -
628cbfc6 by Takanori Nakane at 2018-05-22T12:24:30+01:00
motioncorr_runner: Made --angpix mandatory. Always writes angpix, voltage, dose and pre-exposure to STAR files.
GUI: Do not print --gpu when --use_own. Changed default Cs to 2.7 (Titan & Talos).
- - - - -
f0ce4bb1 by Sjors Scheres at 2018-05-22T12:26:54+01:00
added functionality for linux wildcards in test for existence of file in combineallEPSintoPDF
- - - - -
9ceccdb9 by Takanori Nakane at 2018-05-22T12:31:06+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
a41cecdd by Takanori Nakane at 2018-05-22T12:34:41+01:00
motioncorr_runner: removed debug output
- - - - -
36f64577 by Takanori Nakane at 2018-05-22T14:32:39+01:00
motioncorr_runner: save angpix to the image header
- - - - -
3d167887 by Takanori Nakane at 2018-05-22T15:20:29+01:00
mask_create: implemented raised crown mask (including raised cosine) with the --denovo option.
- - - - -
6a8828aa by Takanori Nakane at 2018-05-22T17:29:11+01:00
flex_analyser: initialise rescale_3dmodels. This fixes SEGV in --do_map.
- - - - -
295e94fb by Takanori Nakane at 2018-05-22T17:33:06+01:00
Merge branch 'jzivanov'
- - - - -
606df0be by Charles Congdon at 2018-05-22T10:23:08-07:00
Add installation instructions for accelerated kernels
- - - - -
88fbe212 by Charles Congdon at 2018-05-22T11:28:14-07:00
Add run instructions for ALTCPU version
- - - - -
72b5a18d by Sjors Scheres at 2018-05-23T09:38:09+01:00
small changes to autopicker: threshold variation opposite direction
- - - - -
fc8a1dd2 by Jasenko Zivanov at 2018-05-23T11:21:33+01:00
traditional B-factor estimation implemented; \n replaced by std::endl in package motion
- - - - -
dd37050b by Takanori Nakane at 2018-05-23T12:00:17+01:00
relion_star_handler: Added --copy_column_from. Fixed indentations.
- - - - -
055e6505 by Jasenko Zivanov at 2018-05-23T15:37:19+01:00
missing files added
- - - - -
378ea466 by Takanori Nakane at 2018-05-23T16:14:52+01:00
relion_star_handler: implemented "--stat_column" to show column statistics and histogram
- - - - -
f359fc8d by Takanori Nakane at 2018-05-24T15:05:50+01:00
Flip the Y axis in PDF outputs from old and new Polish, CtfRefine and MotionCorr.
- - - - -
825b5a23 by Takanori Nakane at 2018-05-24T15:20:27+01:00
Remove temporary comment
- - - - -
ffadef71 by Jasenko Zivanov at 2018-05-24T16:59:10+01:00
Merge remote-tracking branch 'origin/master' into jzivanov
- - - - -
3ddc01f7 by Jasenko Zivanov at 2018-05-24T22:43:23+01:00
bugfixes in bfactor_estimator
- - - - -
955a6b9f by Sjors Scheres at 2018-05-25T10:14:01+01:00
added apply_rounded_offsets_only option
- - - - -
ef516524 by Sjors Scheres at 2018-05-25T10:14:58+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
3f92c063 by Sjors Scheres at 2018-05-25T10:17:33+01:00
write out incomplete last split starfile
- - - - -
bd258a48 by Sjors Scheres at 2018-05-25T10:19:44+01:00
removed spurious spaces which GPUs to use and replaced log_diam_min and max
- - - - -
3e607f55 by Sjors Scheres at 2018-05-25T10:20:25+01:00
log_diam min and max, as well as different default threshold
- - - - -
2f7dc9ae by Sjors Scheres at 2018-05-25T10:20:40+01:00
added log_diam min and max on GUI
- - - - -
16eeb665 by Sjors Scheres at 2018-05-25T10:22:27+01:00
switched back off asktraining
- - - - -
dab5a185 by Sjors Scheres at 2018-05-25T12:14:48+01:00
small change in warning
- - - - -
b4a00ea9 by Sjors Scheres at 2018-05-25T12:17:03+01:00
added relion_it.py for onthefly processing
- - - - -
8b40498f by Sjors Scheres at 2018-05-25T14:19:37+01:00
added WaitForJobs, so one can also submit to an external cluster
- - - - -
55b85896 by Sjors Scheres at 2018-05-25T16:37:28+01:00
added 3D classification option
- - - - -
006ba5a0 by Sjors Scheres at 2018-05-25T16:26:07+00:00
Merged in scheres/relion-devel-lmb (pull request #154)
Master
- - - - -
f7e7c07f by Jasenko Zivanov at 2018-05-25T22:59:51+01:00
bfactor-estimation reconstructions written out sequentially
- - - - -
edd80936 by Jasenko Zivanov at 2018-05-27T19:00:26+01:00
ctf_nyquist_test app reintroduced
- - - - -
a11da39f by Jasenko Zivanov at 2018-05-27T20:43:19+01:00
frame_recombiner is now able to handle positive B-factors (supplied via the --bfactors argument)
- - - - -
0d2b0d19 by Jasenko Zivanov at 2018-05-28T05:46:57+01:00
frame recombination now supports altering the box size - as long as the reference pixel size is an integral multiple of the movie pixel size
- - - - -
039358e9 by Sjors Scheres at 2018-05-28T10:56:52+01:00
added option to discard images based on statstitics in Subset selection jobtype and star_handler
- - - - -
4865827f by Sjors Scheres at 2018-05-28T10:57:13+01:00
write where .lock of pipleine file comes from
- - - - -
fd773f47 by Sjors Scheres at 2018-05-28T11:54:25+01:00
write out more informative .lock file for pipeliner#
- - - - -
40efe00e by Sjors Scheres at 2018-05-28T16:25:31+01:00
add option to autopick with a 3D reference
- - - - -
9400d367 by Sjors Scheres at 2018-05-28T16:25:57+01:00
only print info logfile if nr_repeat>1
- - - - -
58421d02 by Jasenko Zivanov at 2018-05-28T16:56:50+01:00
frame_recombine: boxes are now always extracted at optimal per-frame pixel-positions; this preserves the signal near the edges.
- - - - -
ce4722cf by Sjors Scheres at 2018-05-28T19:13:48+01:00
added inimodel and second pass with 3D reference autopicking
- - - - -
dd4e3a15 by Sjors Scheres at 2018-05-28T19:14:17+01:00
small change in IO autopicker
- - - - -
eb776a49 by Sjors Scheres at 2018-05-29T07:31:33+01:00
added option skip annealing in SGD and switch off do_shifts_onthefly whne do_gpu or do_cpu, as per Bjorns recommendation
- - - - -
4c8eafc4 by Sjors Scheres at 2018-05-29T07:32:02+01:00
added label for optimiser skip SGD annealing of multiple classes
- - - - -
8d11bbe4 by Sjors Scheres at 2018-05-29T07:32:27+01:00
updated version to 3.0-alpha-5
- - - - -
940e028c by Sjors Scheres at 2018-05-29T07:59:44+01:00
added better functionality second pass
- - - - -
0e7e4cdb by Sjors Scheres at 2018-05-29T09:44:17+01:00
minor changes in relion_it
- - - - -
222159dd by Sjors Scheres at 2018-05-29T12:27:32+01:00
use MotionCor2 by default, also repaired bug when importing micrographs instead of movies
- - - - -
499d11f8 by Takanori Nakane at 2018-05-29T12:36:31+01:00
The type of CPlot2D::SetFlipY() should have been void.
- - - - -
22747240 by Takanori Nakane at 2018-05-29T12:37:24+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
61a3c95e by Takanori Nakane at 2018-05-30T12:47:35+01:00
relion_project: use the pixel size in the header of the input volume when not specified.
write the pixel size to the output.
- - - - -
4f58ec6f by Takanori Nakane at 2018-05-30T12:55:54+01:00
motioncorr_runner: improved warnings for patch_x/y
- - - - -
c6236096 by Sjors Scheres at 2018-05-30T15:58:18+00:00
Merged in jzivanov (pull request #15)
Jzivanov
- - - - -
2351474d by Sjors Scheres at 2018-05-30T17:21:31+01:00
added functionality to CtfFind and Autopick jobtypes to write out logfile.pdf, also changes star_handler way of doing histograms
- - - - -
1e103dd7 by Sjors Scheres at 2018-05-30T17:36:58+01:00
corrected small error in autopicker logfile
- - - - -
e699791b by Sjors Scheres at 2018-05-30T17:37:22+01:00
repaired print_sym_ops and metadata labels without --i or --o
- - - - -
2bb629e4 by Sjors Scheres at 2018-05-30T18:07:27+01:00
added relion_it script to generate Rosenthal and Henderson B-factor plots automatically
- - - - -
e328eda7 by Sjors Scheres at 2018-05-30T18:08:16+01:00
small repairs in logfile for ctffind runner
- - - - -
c77d59db by Sjors Scheres at 2018-05-30T18:12:06+01:00
repaired bug when importing micrographs instead of movies
- - - - -
9683841c by Sjors Scheres at 2018-05-30T17:25:14+00:00
Merged in scheres/relion-devel-lmb (pull request #156)
Master
- - - - -
2cf72e4b by Takanori Nakane at 2018-05-30T19:54:44+01:00
relion_reconstruct: fixed indentations
- - - - -
4cf11d3e by Takanori Nakane at 2018-05-30T19:56:33+01:00
relion_reconstruct: set the pixel size in the output image header
- - - - -
0843bee3 by Takanori Nakane at 2018-05-30T19:56:56+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
ad4aad93 by Sjors Scheres at 2018-05-31T09:44:56+01:00
dont die if an autopick star file doesnt exist when making logfile
- - - - -
fc2e78c3 by Sjors Scheres at 2018-05-31T09:45:14+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
a4ca5fa1 by Sjors Scheres at 2018-05-31T11:54:16+01:00
rerun logfile.pdf generation autopick
- - - - -
198f38a8 by Sjors Scheres at 2018-05-31T12:40:10+01:00
dont restart 2D classifications/inimodel in first batch upon restarting the script if the number of particles in the split_job hasn't increased meanwhile
- - - - -
8bad6fee by Takanori Nakane at 2018-05-31T12:50:52+01:00
Added relion_align_symmetry. Useful for aligning initial models generated in C1 to the symmetry convention.
TODO: local refinement, try multiple point group simultaneously
- - - - -
2124a288 by Takanori Nakane at 2018-05-31T12:52:37+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
377fdef9 by Sjors Scheres at 2018-05-31T15:00:22+01:00
still debugging this nasty second pass
- - - - -
d1ea66ae by Sjors Scheres at 2018-05-31T15:00:40+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
d93b3060 by Takanori Nakane at 2018-05-31T15:11:34+01:00
relion_align_symmetry: implemented local search
- - - - -
251e6dfe by Takanori Nakane at 2018-05-31T15:12:11+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
0839d840 by Sjors Scheres at 2018-05-31T15:25:04+01:00
switch off debug statement in autopicker
- - - - -
dde634b0 by Takanori Nakane at 2018-05-31T15:47:04+01:00
Mentioned relion_align_symmetry in the help message in the GUI.
- - - - -
6fca449f by Charles Congdon at 2018-06-01T14:21:12-07:00
Try to ensure long types are not accidentally truncated to int
- - - - -
99af1cc6 by Sjors Scheres at 2018-06-04T12:15:13+01:00
add plots of total accumulated motion per micrograph to logfile.pdf
- - - - -
d16e1cbe by Sjors Scheres at 2018-06-04T12:15:29+01:00
more verbose generation of logfile
- - - - -
6e5af92c by Sjors Scheres at 2018-06-04T12:15:39+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
7b555823 by Charles Congdon at 2018-06-04T14:46:30-07:00
README and CMakeLists.txt updates
- - - - -
4acc21f5 by Charles Congdon at 2018-06-04T16:42:32-07:00
More updates to the README.md
- - - - -
fe703332 by Jasenko Zivanov at 2018-06-05T03:16:48+01:00
dot size increased in motion plot
- - - - -
0057edad by Sjors Scheres at 2018-06-05T09:10:13+01:00
repaired small bug in histogram EPS and in autopick logfile generation on gpu
- - - - -
ff152225 by Sjors Scheres at 2018-06-05T09:10:40+01:00
removed option to ctffind on movies from GUI
- - - - -
f896b662 by Sjors Scheres at 2018-06-05T09:11:07+01:00
added option to ignore ctffind parameters and do EPA for gctf
- - - - -
17e8e5fa by Sjors Scheres at 2018-06-05T09:41:52+01:00
repaired bug of too large auto-diameters in sorting
- - - - -
a0ba6466 by Takanori Nakane at 2018-06-05T12:01:08+01:00
relion_image_handler: Implemented --invert_hand. Made --o mandatory for safety.
- - - - -
7ae38930 by Sjors Scheres at 2018-06-05T12:05:50+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
7e5a5f19 by Takanori Nakane at 2018-06-05T15:32:04+01:00
relion_image_handler: some usages do not need --o
- - - - -
aabf2018 by Takanori Nakane at 2018-06-05T17:34:48+01:00
relion_image_handler: added power-spectrum. <cos(delta-phi)> is still work in progress, so hidden.
- - - - -
942f48b8 by Jasenko Zivanov at 2018-06-06T01:28:27+01:00
Old B-factor estimation now works in single precision; Micrograph no longer returns a global component of -9999 for missing frames
- - - - -
ff0a012e by Takanori Nakane at 2018-06-06T11:11:09+01:00
Merge branch 'jzivanov'
- - - - -
706f5c23 by Takanori Nakane at 2018-06-06T11:53:58+01:00
Micrograph Model: When not observed, return the shift of the closest observed frame.
- - - - -
b6563ec0 by Takanori Nakane at 2018-06-06T12:06:39+01:00
motioncorr_runner: use the return value of getShiftAt()
- - - - -
dc9c42ee by Sjors Scheres at 2018-06-06T12:07:55+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
41f53f0c by Sjors Scheres at 2018-06-06T12:09:33+01:00
dont die if cannot getShift for calculation of accumulated motion
- - - - -
391e2240 by Sjors Scheres at 2018-06-06T12:10:00+01:00
set default sigma_offset to 10 pixels instead of 3
- - - - -
34760641 by Sjors Scheres at 2018-06-06T12:10:27+01:00
particles can also come straight from CtfRefine
- - - - -
62187108 by Sjors Scheres at 2018-06-06T12:10:41+01:00
write out warning about .lock file from gui
- - - - -
7f7d96f1 by Takanori Nakane at 2018-06-06T12:39:20+01:00
bfactor_plot: Use proper Angstrom sign in the plot
- - - - -
9170f0b0 by Sjors Scheres at 2018-06-06T12:44:55+01:00
removed dependency on 2 tex libraries
- - - - -
060400e2 by Sjors Scheres at 2018-06-06T12:45:31+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
15a688d2 by Sjors Scheres at 2018-06-06T13:09:56+01:00
added PDF with histogram and all values of Zscores
- - - - -
61734ee8 by Takanori Nakane at 2018-06-06T14:15:30+01:00
bfactor_plot: added maximum_nr_particles. skip failed Refine3D.
- - - - -
c326b0df by Takanori Nakane at 2018-06-06T14:16:41+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
fc78e78d by Sjors Scheres at 2018-06-06T14:57:29+00:00
Merged in scheres/relion-devel-lmb (pull request #157)
Master
- - - - -
8caa74a7 by Charles Congdon at 2018-06-06T09:30:39-07:00
Change #pragma simd statements to OpenMP 4.x standard
- - - - -
fc3c5c2f by Charles Congdon at 2018-06-06T11:25:35-07:00
Work around ICC issue before 2018 update 2
- - - - -
8e0a974e by Takanori Nakane at 2018-06-06T21:39:26+01:00
motioncorr_runner: Fixed a bug introduced this morning. In C, 0 is FALSE, all others are TRUE ...
- - - - -
92f048d1 by Takanori Nakane at 2018-06-07T11:47:17+01:00
relion_it.py: Use resolution instead of class distribution to find the best class.
Fixed undefined "batch_repeat_time" error
- - - - -
723a11f6 by Sjors Scheres at 2018-06-07T13:55:10+01:00
switched ctf refinement and bayesian polishing around
- - - - -
f26617db by Sjors Scheres at 2018-06-07T13:55:31+01:00
wait one minyte before dying on non-existent .lock
- - - - -
88ac883d by Sjors Scheres at 2018-06-07T13:55:49+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
61db8679 by Sjors Scheres at 2018-06-08T09:47:11+01:00
write out star file for the new stack
- - - - -
6613592f by Sjors Scheres at 2018-06-08T09:47:37+01:00
replaced int for long in loop over vector
- - - - -
d9d5f224 by Sjors Scheres at 2018-06-08T09:48:11+01:00
added cleaning for ctfrefine, motionrefine and inimodel jobtypes
- - - - -
4eafda97 by Takanori Nakane at 2018-06-08T12:02:25+01:00
Import: sort files by time so that incremental subset works fine in relion_it.py
- - - - -
26c97f7b by Takanori Nakane at 2018-06-08T12:02:54+01:00
own_motioncorr: get rid of warnings about grouping
- - - - -
e6fedc61 by Sjors Scheres at 2018-06-08T15:20:36+01:00
repaired before for stop_after_ctf_estimation
- - - - -
1f590e27 by Bjoern Forsberg at 2018-06-08T17:58:39+02:00
Merge branch 'master' of https://bitbucket.org/tcblab/relion-devel-tcblab
- - - - -
4ef2968c by Bjoern Forsberg at 2018-06-08T18:00:39+02:00
Simplified README.md
- - - - -
9e486b6e by Jasenko Zivanov at 2018-06-09T09:53:18+01:00
bugfix in FCC computation
- - - - -
c6a82a86 by Takanori Nakane at 2018-06-10T18:05:39+01:00
TIFF: supported TIFF images where the last strip is shorter than others.
- - - - -
59351ce3 by Takanori Nakane at 2018-06-10T18:06:38+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
795d3c97 by Takanori Nakane at 2018-06-10T21:14:18+01:00
motioncorr_runner: Check the presence of meta data star file when --only_do_unfinished
- - - - -
b6d8c06b by Takanori Nakane at 2018-06-10T21:15:44+01:00
relion_it.py: when the resolution is the same, take the bigger class
- - - - -
b3107ed7 by Takanori Nakane at 2018-06-11T12:17:32+01:00
Cplot2D::joinMultipleEPSIntoSinglePDF: make this robust against the limitation of the command line length (ARG_MAX)
- - - - -
0db9e08e by Sjors Scheres at 2018-06-11T15:09:21+01:00
added error for not having used relion3 for running mc2
- - - - -
725dffd5 by Sjors Scheres at 2018-06-11T15:09:32+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
ecef6167 by Sjors Scheres at 2018-06-11T15:14:28+01:00
place error below trace not to confuse users..
- - - - -
6cb20e89 by Charles Congdon at 2018-06-11T16:45:53-07:00
Fix ALTCPU build - typo in helper.cpp
- - - - -
b504f463 by Takanori Nakane at 2018-06-12T11:39:07+01:00
Removed some testing programs from standard build
- - - - -
13add675 by Takanori Nakane at 2018-06-12T11:39:36+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
3ca2547b by Sjors Scheres at 2018-06-12T12:03:16+01:00
remove remove_duplicates from extract gui as it doesnt work yet
- - - - -
f2d62fef by Sjors Scheres at 2018-06-12T12:03:40+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
ba20d8e4 by Sjors Scheres at 2018-06-12T14:45:06+00:00
Merged in scheres/relion-devel-lmb (pull request #158)
Master
- - - - -
bf589ad0 by Takanori Nakane at 2018-06-12T16:21:17+01:00
Expose the queue name and the queue submit command as environment variables. Also the maximum number of MPI and threads.
This is based on the pull request #338 from @JanneM and @jmansour on GitHub (thanks!).
- - - - -
4bfad023 by Takanori Nakane at 2018-06-12T16:24:27+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
6d96013c by Takanori Nakane at 2018-06-12T16:44:48+01:00
Fixed some GUI help messages. Guard against SEGV with RELION_QSUB_EXTRAs.
- - - - -
41d09ce3 by Lanqing Huang at 2018-06-12T16:47:40+01:00
Fixed overflow bug in relion_MPI_Bcast Implement Bcast to handle larger buffer size. The definition for nzyxdim of MultidimArray is a long int type. In some cases, large boxsize makes nzyxdim overflow when broadcasting data.
Signed-off-by: Takanori Nakane <nakane.t at gmail.com>
- - - - -
84515654 by Lanqing Huang at 2018-06-12T16:47:40+01:00
Use static_cast for type conversion
Signed-off-by: Takanori Nakane <nakane.t at gmail.com>
- - - - -
e2de1770 by Sjors Scheres at 2018-06-13T09:20:01+01:00
changed defaults in header
- - - - -
3ed88213 by Sjors Scheres at 2018-06-13T11:38:17+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
ad51ac0d by Takanori Nakane at 2018-06-13T12:01:10+01:00
bfactor_plot.py: Print B factors from postprocess as well
- - - - -
7d8c4c32 by Takanori Nakane at 2018-06-13T12:02:16+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
ef08d8fe by Takanori Nakane at 2018-06-13T12:19:51+01:00
GUI: give an early error when GPU is enabled without image alignments.
Don't deactivate the Run button when the job is not submitted due to an error.
- - - - -
ee4e61ff by Bjoern Forsberg at 2018-06-14T13:06:52+02:00
ALTCPU=ON sets CUDA=OFF, quits if both
- - - - -
1043588e by Takanori Nakane at 2018-06-14T14:55:52+01:00
bfactor_plot: Added double axes. Print the coefficients. Expolate the line.
- - - - -
11bd940d by Takanori Nakane at 2018-06-14T16:02:52+01:00
relion_pipeliner: added --setJobAlias option
- - - - -
0f3a9170 by Takanori Nakane at 2018-06-14T16:04:11+01:00
bfactor_plot.py: expolation of the fit. added opts.prefix
- - - - -
8500ea4e by Takanori Nakane at 2018-06-14T20:45:32+01:00
own_motioncorr: Detect and fix hot pixels
MicrographModel: save defect file (currently not used though)
- - - - -
da5e4ef6 by Takanori Nakane at 2018-06-14T23:14:38+01:00
bfactor_plot.py: Make X11 unnecessary, Tidy up aliases.
- - - - -
2922f18e by Takanori Nakane at 2018-06-14T23:15:18+01:00
autopicker: remove unnecessary space from the file name
- - - - -
1d14d04e by Takanori Nakane at 2018-06-15T09:34:23+01:00
relion_it.py: set job aliases
- - - - -
47babf8b by Takanori Nakane at 2018-06-15T09:48:39+01:00
GUI: Improved help messages.
- - - - -
d68b4f6e by Takanori Nakane at 2018-06-15T12:16:02+01:00
relion_it.py: described how to re-run the second pass when the automatic selection of the reference was wrong.
- - - - -
80d8651f by Sjors Scheres at 2018-06-15T12:25:39+01:00
removed confusing save STAR with selected images from subset selection
- - - - -
912a2d84 by Sjors Scheres at 2018-06-15T12:32:18+01:00
Merged tcblab/relion-devel into master
- - - - -
c5a0cc00 by Sjors Scheres at 2018-06-15T12:34:22+01:00
made new option to reset offsets upon re-extract from data.star
- - - - -
6647ff83 by Sjors Scheres at 2018-06-15T12:34:33+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
d36db560 by Takanori Nakane at 2018-06-15T12:43:13+01:00
Merged in tcblab/relion-devel-tcblab (pull request #155)
Resolved Conflicts:
CMakeLists.txt
- - - - -
b4680e08 by Sjors Scheres at 2018-06-15T12:45:42+01:00
Merged tcblab/relion-devel into master
- - - - -
5c89192b by Takanori Nakane at 2018-06-15T15:08:36+01:00
Workaround for FLTK compilation error. See https://patchwork.openembedded.org/patch/112023/
- - - - -
1aa0e221 by Takanori Nakane at 2018-06-15T15:11:00+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
0a1d3ab4 by Takanori Nakane at 2018-06-15T16:53:32+01:00
Add *.tiff to '2D micrograph movies'
- - - - -
f597285d by Sjors Scheres at 2018-06-18T10:52:22+01:00
multiple classes outputNodes in inimodel
- - - - -
e5a15ada by Sjors Scheres at 2018-06-18T10:53:07+01:00
use unmasked fsc as true if solvent correction makes things worse
- - - - -
3a2d2f59 by Sjors Scheres at 2018-06-18T10:53:33+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
7d2e3eb8 by Takanori Nakane at 2018-06-18T15:10:11+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
bb9ab103 by Sjors Scheres at 2018-06-19T09:38:59+01:00
added modified betaGuide
- - - - -
0e42f33e by Sjors Scheres at 2018-06-20T11:27:33+01:00
had forgotten to set version number to beta
- - - - -
f9087e64 by Sjors Scheres at 2018-06-21T09:08:43+01:00
minor improvements by Colin
- - - - -
5c10379e by Sjors Scheres at 2018-06-21T09:10:25+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-3.0_beta
- - - - -
741d2462 by Takanori Nakane at 2018-06-21T10:23:24+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
2fb81c2c by Takanori Nakane at 2018-06-22T11:39:17+01:00
GUI: Fix crash when opening old project directories
- - - - -
96c6a216 by Takanori Nakane at 2018-06-22T11:39:50+01:00
GUI: Fix SEGV reported by @pemsley on GitHub (#370) when double clicking the stdout/stderr box before selecting job.
- - - - -
25c2ced9 by Takanori Nakane at 2018-06-22T11:41:05+01:00
relion_reconstruct: added --skip_weight and --skip_mask for Ewald sphere correction
- - - - -
6ceed106 by Takanori Nakane at 2018-06-25T10:50:32+01:00
GUI: Fix SEGV in Import
- - - - -
97e8c6b7 by Sjors Scheres at 2018-06-26T09:50:20+01:00
better description of possible movie extensions
- - - - -
b0a73ffd by Sjors Scheres at 2018-06-26T09:50:41+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
ca73a0bd by Takanori Nakane at 2018-06-26T13:59:02+01:00
relion_preprocess: print the extracted pixel size.
- - - - -
c316b108 by Takanori Nakane at 2018-06-26T15:07:01+01:00
GUI: More accurate explanation of "Range for defocus fit"
- - - - -
1ce89431 by Sjors Scheres at 2018-06-27T10:02:00+01:00
fixed bug in labelling axes in C2PLot, now have to do this last
- - - - -
fd44b469 by Sjors Scheres at 2018-06-27T10:02:13+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
706a30f1 by Sjors Scheres at 2018-06-28T13:19:46+01:00
use long for split_size in star_handler
- - - - -
29f642b8 by Sjors Scheres at 2018-06-28T13:49:38+01:00
replaced some backslah-n for std::endl
- - - - -
c10c67f2 by Sjors Scheres at 2018-06-28T13:49:55+01:00
replaced all std::endl for backslah-n
- - - - -
6a692c0a by Sjors Scheres at 2018-06-28T13:50:12+01:00
replaced some backslah-n for std::endl
- - - - -
e7a8a07b by Sjors Scheres at 2018-06-29T12:58:48+01:00
use mkdir instead of touch for lock as the former is atomic and the latter is not
- - - - -
548365f3 by Takanori Nakane at 2018-06-29T15:04:04+01:00
relion_star_handler: added --select_str_label, added --remove_duplicates.
TODO: make GUI, support down-sampled particles
- - - - -
efeb7ea9 by Takanori Nakane at 2018-06-29T15:04:51+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
0c4a50f8 by Takanori Nakane at 2018-06-29T15:59:43+01:00
relion_run_ctffind, relion_run_motioncorr: implemented --do_at_most N for pipeline
- - - - -
a69b7f5a by Takanori Nakane at 2018-07-02T11:28:35+01:00
relion_it: now it can pick up results from continued jobs
- - - - -
df3397bb by Takanori Nakane at 2018-07-03T21:20:25+01:00
CPlot2D: fixed SEGV when there are no point in DataSet. This fixes SEGV on CtfFind and AutoPick.
relion_autopick: Fix average particle numbers when there is no input micrograph
- - - - -
998ec341 by Takanori Nakane at 2018-07-04T12:36:16+01:00
motioncorr_runner: make an empty PDF when there is no input. otherwise, the job never finishes.
- - - - -
cb801521 by Takanori Nakane at 2018-07-04T15:53:40+01:00
star_handler: remove_duplicate now supports down-sampled particles
- - - - -
16d3f460 by Takanori Nakane at 2018-07-04T16:29:45+01:00
GUI: Remove duplicates
- - - - -
c78ae484 by Sjors Scheres at 2018-07-06T12:26:04+01:00
warn about lock directory rather than file
- - - - -
b3271e9a by Sjors Scheres at 2018-07-06T14:09:14+01:00
write out reference projections from 3D model for autopicking
- - - - -
8a039a40 by Sjors Scheres at 2018-07-06T14:09:20+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
909754bb by Sjors Scheres at 2018-07-06T14:48:56+01:00
repaired bug in chaeck for particle diameter
- - - - -
cde6d9a7 by Sjors Scheres at 2018-07-06T15:08:28+01:00
warn if pixel size in reference images is not consistent with angpix_ref
- - - - -
65ccfd03 by Sjors Scheres at 2018-07-09T17:46:51+01:00
only write EPS file once, was done many many times....
- - - - -
d2f0dd7d by Sjors Scheres at 2018-07-10T09:40:43+01:00
first submit queue options then add discard_job
- - - - -
d0d093bc by Sjors Scheres at 2018-07-10T12:57:59+01:00
changed ref-based autopicking threshold default to 0.4
- - - - -
769f3405 by Sjors Scheres at 2018-07-11T15:13:20+01:00
dont delete stuff from the last iteration to still allow continuation or multibody refinement
- - - - -
eeb3f704 by Sjors Scheres at 2018-07-16T09:26:03+01:00
bug in reported pixel size
- - - - -
bc8ed810 by Sjors Scheres at 2018-07-17T12:07:01+01:00
worked towards patch from EMBL to have multiple extra fields on GUI
- - - - -
5e76a75a by Sjors Scheres at 2018-07-17T15:32:43+01:00
repaired bug maxres
- - - - -
5a571bdb by Takanori Nakane at 2018-07-17T20:27:47+01:00
relion_ctf_refine: update STAR file with user provided --angpix, --beamtilt_x, --beamtilt_y
- - - - -
32189ec1 by Takanori Nakane at 2018-07-17T20:28:56+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
4160e579 by Takanori Nakane at 2018-07-18T11:52:50+01:00
Alias: Prohibit use of symbols that might cause problems on (some) file systems.
Thanks to suggestion by @rbehrouzi on GitHub issues #372.
- - - - -
a088d7ac by Takanori Nakane at 2018-07-18T12:00:22+01:00
relion_particle_sort: change the default output filename to sorted.star to avoid overwriting the input.
Thanks to suggestion by @daniel-s-d-larsson in Github issue #374.
- - - - -
3f5ab301 by Sjors Scheres at 2018-07-18T15:15:26+01:00
sleep 60 seconds before checkCompleteion in pipeliner because of leeds
- - - - -
3411e814 by Takanori Nakane at 2018-07-22T22:34:13+01:00
ml_model: Set the pixel size in the header for class averages
- - - - -
ad5e48a9 by Jasenko Zivanov at 2018-07-23T08:45:33+01:00
Merge remote-tracking branch 'origin/master' into jzivanov
- - - - -
55f2eaa1 by Jasenko Zivanov at 2018-07-23T08:49:45+01:00
missing apps added (not for release; good to have them in the repository)
- - - - -
d5a96fad by Takanori Nakane at 2018-07-23T14:13:56+01:00
GUI help for Cs: Most JEOL microscopes have Cs=2.7. Only some have 1.4.
- - - - -
c26cc874 by Takanori Nakane at 2018-07-24T14:14:27+01:00
relion_reconstruct: check the presence of the subset column
- - - - -
a0ad84c3 by Jasenko Zivanov at 2018-07-25T08:43:55+01:00
Beam-tilt handling updated: tilt is always used via MD-table, ObservationModel no longer knows its own tilt values. If tilt values are passed through the cmd. line, they are first written into the table; Motion_fit now supports tilt (again), but only through the star-file.
- - - - -
87476f20 by Jasenko Zivanov at 2018-07-25T08:55:44+01:00
Ctf_refine: beam-tilt source stated on console (cmd. line, star file or neither)
- - - - -
540e5699 by Takanori Nakane at 2018-07-25T10:53:18+01:00
alias: prohibits use of % and & as well
- - - - -
73d0d246 by Takanori Nakane at 2018-07-25T11:16:19+01:00
relion_preprocess: set the pixel size in the image header
- - - - -
be495bcc by Sjors Scheres at 2018-07-25T15:16:54+01:00
dont use strstream in slow writing of star files
- - - - -
e568c208 by Sjors Scheres at 2018-07-25T15:17:17+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
35a81911 by Sjors Scheres at 2018-07-25T15:18:06+01:00
moved to beta-2 version
- - - - -
5f3d17a7 by Takanori Nakane at 2018-07-25T16:49:26+01:00
Included MTF STAR files in the distribution.
- - - - -
8206e2a2 by Takanori Nakane at 2018-07-25T16:50:24+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
64cb30a4 by Jasenko Zivanov at 2018-07-25T16:53:30+01:00
Beamtilt classes implemented in ctf_refine
- - - - -
9911a539 by Takanori Nakane at 2018-07-26T10:21:48+00:00
Merged in scheres/relion-devel-lmb (pull request #159)
sync relion3-beta2 to relion-devel
- - - - -
f46fc135 by Sjors Scheres at 2018-07-26T12:12:01+01:00
only master writes out projected images from 3D reference
- - - - -
e8092e90 by Sjors Scheres at 2018-07-26T12:12:32+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
24f471d4 by Takanori Nakane at 2018-07-30T19:49:08+01:00
relion_motioncorr_runner: added --max_io_threads to limit parallel disc access
- - - - -
bd535a88 by Takanori Nakane at 2018-07-30T19:49:45+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
bf9f35ae by Takanori Nakane at 2018-07-31T11:13:49+01:00
relion_reconstruct: removed debug codes (thanks to @craigyk for pointing this out)
- - - - -
814940d4 by Sjors Scheres at 2018-07-31T11:43:53+00:00
Merged in scheres/relion-devel-lmb/jzivanov (pull request #160)
Jzivanov
- - - - -
a7bb1a51 by Sjors Scheres at 2018-07-31T12:47:06+01:00
Merged tcblab/relion-devel into master
- - - - -
ac6ec4c0 by Sjors Scheres at 2018-07-31T12:49:39+01:00
removed temp programs from jasenko from default build
- - - - -
d2dd1b65 by Takanori Nakane at 2018-08-01T15:34:46+01:00
relion_mask_create: write the correct pixel size to the header even without low-pass filtering.
Thanks to Bonnie Murphy for the bug report.
- - - - -
c2e0eb56 by Takanori Nakane at 2018-08-01T15:36:01+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
ca338c43 by Sjors Scheres at 2018-08-01T17:10:21+01:00
updated betaGuide for general beta
- - - - -
3b9e3871 by Sjors Scheres at 2018-08-01T17:10:33+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
8d2476f2 by Sjors Scheres at 2018-08-02T11:12:14+01:00
default behaviour of EXTRA_COUNT same as before
- - - - -
68ab09c0 by Takanori Nakane at 2018-08-03T11:56:17+01:00
pipeliner: revert the default wait between jobs to 10 seconds and added --sec_wait_after to change this.
- - - - -
616d03bb by Takanori Nakane at 2018-08-03T12:24:17+01:00
relion_it.py: include Colin's suggestions
- - - - -
8c2b7743 by Takanori Nakane at 2018-08-03T14:21:07+01:00
LoG picker: output dummy columns of psi and class to prevent problems in Extraction
- - - - -
cea11c92 by Jasenko Zivanov at 2018-08-04T17:13:44+01:00
irrelevant changes to apps/CMakeLists.txt
- - - - -
527acda2 by Jasenko Zivanov at 2018-08-04T17:15:11+01:00
irrelevant changes to apps/CMakeLists.txt (2)
- - - - -
915dde90 by Jasenko Zivanov at 2018-08-05T03:09:02+01:00
Zernike polynomials implemented and tested
- - - - -
31c59854 by Jasenko Zivanov at 2018-08-05T16:25:12+01:00
optics groups implemented (not yet supported) and conversion program written
- - - - -
cf2a2b96 by Jasenko Zivanov at 2018-08-05T20:32:12+01:00
ctf_refine now uses the new data format; CTF::read renamed to CTF::readLegacy; ObservationModel renamed to LegacyObservationModel
- - - - -
50c91f1f by Jasenko Zivanov at 2018-08-05T23:42:03+01:00
new CTF::read renamed to CTF::readByGroup to avoid confusion during merging
- - - - -
416d1e95 by Takanori Nakane at 2018-08-06T17:26:50+01:00
relion_star_handler: write split001.star the last. Otherwise the pipeliner can prematurely
think the job has been finished and this can kill relion_it.py
- - - - -
63970b56 by Takanori Nakane at 2018-08-06T21:34:53+01:00
relion_it.py: proper parsing of SUBMITTED_JOBS. Respect extract_bg_diamer..
- - - - -
fdd22707 by Takanori Nakane at 2018-08-06T21:42:49+01:00
bfactor_plot.py: same improvement as relion_it.py
- - - - -
abeddb88 by Takanori Nakane at 2018-08-06T22:03:41+01:00
relion_it.py: submit class2d_batch001 if not yet.
- - - - -
8cdf3dd5 by Takanori Nakane at 2018-08-07T12:54:48+01:00
ctf.h: use SINCOS macro instead of sincos itself.
- - - - -
afec33f1 by Takanori Nakane at 2018-08-07T15:08:39+01:00
remove jiffy test programs from distribution
- - - - -
dfa94448 by Jasenko Zivanov at 2018-08-07T18:41:57+01:00
ctf_refine now uses optics groups instead of tilt classes and it writes the results into <out>_optics.star; handling of optics groups is more robust now (vis-à-vis undefined groups or groups given in the wrong order in the optics file)
- - - - -
66503ee3 by Jasenko Zivanov at 2018-08-08T00:09:48+01:00
tiny adaptations to tilt_estimator; anisotropic-tilt functionality removed
- - - - -
e203e56c by Jasenko Zivanov at 2018-08-08T00:32:35+01:00
yet another attempt
- - - - -
b557f825 by Jasenko Zivanov at 2018-08-08T00:35:00+01:00
re-commit with slightly changed files to trick git into uploading the current versions
- - - - -
d914e4e4 by Jasenko Zivanov at 2018-08-08T01:23:48+01:00
CTF::readLegacy renamed back to read to make cuda code run again
- - - - -
f24b5b87 by Takanori Nakane at 2018-08-08T12:48:30+01:00
GUI: give an early error message if an optimised parameter file is not specified.
- - - - -
f76a7a0d by Sjors Scheres at 2018-08-08T15:11:26+01:00
added option to write out projections along eigenvectors of all particles
- - - - -
568a9ffb by Sjors Scheres at 2018-08-08T15:57:09+01:00
modification in intel instructions suggested by Charles from intel
- - - - -
ab9f552d by Jasenko Zivanov at 2018-08-09T18:43:54+01:00
Odd Zernike fit implemented in tilt_estimator; coefficients written into _optics.star; MetaDataTable now supports vector<double>-typed columns
- - - - -
e0190e93 by Jasenko Zivanov at 2018-08-10T02:42:57+01:00
beam-tilt angle extracted from and inserted into Zernike coefficients; phase shift caching implemented in ObsModel; ObsModel loading streamlined
- - - - -
1a83cf1b by Jasenko Zivanov at 2018-08-10T03:35:32+01:00
double_reconstruct adapted to new data format (incl. Zernike phase model)
- - - - -
f974e61f by Jasenko Zivanov at 2018-08-10T03:49:10+01:00
double-template definition in metadata_container.h made CUDA compliant (> > instead of >>)
- - - - -
a382efa2 by Jasenko Zivanov at 2018-08-10T13:56:01+01:00
old reconstruct updated to new format (untested)
- - - - -
3038be3e by Jasenko Zivanov at 2018-08-10T13:57:30+01:00
Merge branch 'jaz_3.1' of https://bitbucket.org/scheres/relion-devel-lmb into jaz_3.1
- - - - -
f9e4ec45 by Takanori Nakane at 2018-08-14T10:31:39+01:00
relion_postprocess: added --force_mask option to use the mask even when the resolution becomes worse
- - - - -
b0be4a0a by Takanori Nakane at 2018-08-14T10:32:52+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
fc393691 by Takanori Nakane at 2018-08-14T10:34:48+01:00
relion_motion_refine: make the warning for truncated movies clearer
- - - - -
e189e906 by Takanori Nakane at 2018-08-14T12:31:04+01:00
MetadataTable: fixed SEGV in histogram
- - - - -
9d657c5b by Takanori Nakane at 2018-08-14T17:44:58+01:00
histogram: set ad-hoc upper limit to the bin number to avoid using too much memory
- - - - -
9cc754e8 by Takanori Nakane at 2018-08-16T14:34:09+01:00
GUI: do not add "--fast_subset" when continuing. It is read from optimiser.star.
- - - - -
390afdb9 by Takanori Nakane at 2018-08-16T14:44:31+01:00
relion_it: retry load_star when the file is truncated.
WIP: This is not a perfect solution, which requires a locking mechanism.
- - - - -
5b41293d by Jasenko Zivanov at 2018-08-16T17:12:55+02:00
non-linear aberration fit implemented in ctf_refine
- - - - -
3e9c76d0 by Jasenko Zivanov at 2018-08-16T19:57:38+02:00
demodulate tool written; can output Relion-3.0 star files
- - - - -
ad298041 by Jasenko Zivanov at 2018-08-16T20:32:18+02:00
bugfixes in demodulate
- - - - -
e65ddbe0 by Jasenko Zivanov at 2018-08-16T20:43:03+02:00
RelionErrors printed on construction
- - - - -
079990f0 by Jasenko Zivanov at 2018-08-16T21:08:36+02:00
another bugfix in demodulate
- - - - -
7bb20dde by Takanori Nakane at 2018-08-17T11:11:15+01:00
motion_refiner: improved a warning message
- - - - -
19810adf by Takanori Nakane at 2018-08-17T12:48:54+01:00
relion_autopicker: Fixed 3D reference (had to set XmippOrigin!). Also improved a warning message.
- - - - -
660d890c by Takanori Nakane at 2018-08-17T16:36:39+01:00
decomposePipelineSymlinkName: don't expand jobXXX
- - - - -
6ec12e5f by Takanori Nakane at 2018-08-17T20:10:47+01:00
decomposePipelineSymlinkName: safer link expansion
- - - - -
922ef473 by Jasenko Zivanov at 2018-08-17T23:42:38+02:00
bugfix in star conversion
- - - - -
47258ddb by Jasenko Zivanov at 2018-08-18T00:24:55+02:00
demodulate no longer leaves an optics group id in a relion 3.0 star file
- - - - -
08cd0cd6 by Jasenko Zivanov at 2018-08-18T00:35:55+02:00
exclusion ring implemented in tilt refinement
- - - - -
db1f511b by Takanori Nakane at 2018-08-18T22:40:52+01:00
RelionCor: repaied --early_binning
- - - - -
4be83f76 by Jasenko Zivanov at 2018-08-20T14:58:37+02:00
star file version is now set to 3.0 if no version tag is present
- - - - -
3a88311e by Jasenko Zivanov at 2018-08-20T17:34:34+02:00
format version now written out correctly by convert_star
- - - - -
9a025627 by Takanori Nakane at 2018-08-23T12:46:27+01:00
Fix ResMap for multi-bodies. Do not replace 'half1' and 'half2' within directory names. Fix memory error in the default ResMap path name.
- - - - -
e9843194 by Jasenko Zivanov at 2018-08-24T17:23:18+02:00
Fitting of symmetrical aberrations implemented in ctf_refine (untested)
- - - - -
4ce5613f by Jasenko Zivanov at 2018-08-24T17:36:26+02:00
Bugfixes in symmetric aberration fit
- - - - -
a4f05ab5 by Jasenko Zivanov at 2018-08-24T23:09:16+01:00
Symmetric aberration fit further developed and debugged
- - - - -
a7204350 by Jasenko Zivanov at 2018-08-24T23:30:30+01:00
another bugfix
- - - - -
6bb94017 by Jasenko Zivanov at 2018-08-25T18:57:58+01:00
effects of symmetric aberrations on the CTF implemented, including caching
- - - - -
79882c3a by Jasenko Zivanov at 2018-08-25T21:17:45+01:00
distorted CTFs now used everywhere in src/jaz
- - - - -
447d8d4e by Jasenko Zivanov at 2018-08-26T00:17:52+01:00
multithreaded caching issues resolved
- - - - -
62801343 by Jasenko Zivanov at 2018-08-26T16:43:24+01:00
bugfix in double-vector metadata handling
- - - - -
e5aada94 by Takanori Nakane at 2018-08-27T10:11:12+01:00
GUI: Fix RELION_QSUB_EXTRA1/2 (Issue #391 on GitHub)
- - - - -
9c728e0f by Jasenko Zivanov at 2018-08-27T23:20:39+01:00
pixel-weights corrected in symmetric aberration fit
- - - - -
8d578a99 by Jasenko Zivanov at 2018-08-28T14:13:58+01:00
bugfix in ObservationModel
- - - - -
4772b013 by Jasenko Zivanov at 2018-08-28T17:50:09+01:00
ml_optimiser tentatively updated to new data model (untested)
- - - - -
e33c1685 by Jasenko Zivanov at 2018-08-28T18:03:30+01:00
optics_group removed from ExpOriginalParticle
- - - - -
4c50458c by Jasenko Zivanov at 2018-08-28T18:17:51+01:00
opt. group in acc_ml_optimiser_impl.h now also obtained from Experiment
- - - - -
d42b6f71 by Jasenko Zivanov at 2018-08-28T19:26:14+01:00
double-template '>>' changed to the antiquated '> >' to make NVCC happy
- - - - -
8ccb43d8 by Jasenko Zivanov at 2018-08-28T19:30:03+01:00
more >> changed to > >
- - - - -
5bdb2d65 by Jasenko Zivanov at 2018-08-29T00:36:43+01:00
CTF and pixel-size handling corrected in ml_optimiser
- - - - -
8dd8ca0c by Jasenko Zivanov at 2018-08-29T03:16:14+01:00
ctf handling redone in ml_optimiser: all ctfs are now created at the beginning on all nodes
- - - - -
25dcb426 by Jasenko Zivanov at 2018-08-29T04:03:10+01:00
remnants of the old approach removed from ml_optimiser
- - - - -
11ccb51a by Takanori Nakane at 2018-08-30T15:04:58+01:00
relion_preprocess and GUI: warn about --set_angpix as this option breaks Bayesian Polising.
- - - - -
6581a2b1 by Takanori Nakane at 2018-08-30T15:09:42+01:00
Changes suggested by @Fravadona (thanks!) on GitHub issues #387 for Mac OS X build.
Note that we do not officially support Mac OS X.
- - - - -
340fd229 by Takanori Nakane at 2018-08-30T15:11:36+01:00
motion_helper: atomic operation is not necessary here (confirmed by Jasenko on 29 Aug 2018).
- - - - -
4fbd8096 by Jasenko Zivanov at 2018-08-30T20:07:53+01:00
CTF::getFftwImage now works with images smaller than the original
- - - - -
88b603a6 by Jasenko Zivanov at 2018-08-30T20:19:06+01:00
Bugfix by Takanori applied
- - - - -
0317826a by Jasenko Zivanov at 2018-08-31T03:44:34+01:00
anisotropic mag written out by mag-fit and read by ObervationModel - only used in prediction so far (untested)
- - - - -
954ce066 by Jasenko Zivanov at 2018-08-31T04:46:54+01:00
bugfix in MagnificationHelper
- - - - -
783b2f6a by Jasenko Zivanov at 2018-08-31T14:48:54+01:00
bugfix in symmetric-aberration handling for cropped CTF-images
- - - - -
ac095850 by Jasenko Zivanov at 2018-08-31T14:53:25+01:00
refine continuation bugfix by Takanori applied
- - - - -
20670b20 by Jasenko Zivanov at 2018-08-31T19:31:04+01:00
mag-matrix now transposed and multiplied from the left to A3D due to the internal transposition in Projector::get2DFourierTransform
- - - - -
e76fd3bc by Jasenko Zivanov at 2018-08-31T22:16:22+01:00
anisotropic mag support added to double_reconstruct
- - - - -
b90a3b7f by Jasenko Zivanov at 2018-08-31T23:51:16+01:00
anisotropic mag can be estimated iteratively now
- - - - -
7ad52893 by Jasenko Zivanov at 2018-09-01T00:20:05+01:00
point-group symmetrisation can be performed in parallel now (and it is in double_reconstruct)
- - - - -
5b658abd by Jasenko Zivanov at 2018-09-01T00:29:26+01:00
bugfix in symmetrisation parallelisation
- - - - -
c4f6f3f9 by Jasenko Zivanov at 2018-09-03T04:02:38+01:00
CTFs are now defined in coordinates undistorted by anisotropic magnification
- - - - -
95fcb400 by Jasenko Zivanov at 2018-09-03T11:14:34+01:00
astigmatism is now expressed as a bilinear form inside the CTF class
- - - - -
0aa0b022 by Jasenko Zivanov at 2018-09-04T00:29:52+01:00
astigmatism is now adapted to the new coordinate system after anisotropic mag estimation
- - - - -
ea019680 by Jasenko Zivanov at 2018-09-04T01:05:07+01:00
bugfix in particle-MDT output handling in ctf_refiner
- - - - -
bd838593 by Jasenko Zivanov at 2018-09-04T01:22:23+01:00
bugfix in per-micrograph constant astigmatism adaptation to anisotropic mag
- - - - -
a01ea12a by Jasenko Zivanov at 2018-09-04T03:35:32+01:00
anisotropic mag tentatively added to CPU-refine
- - - - -
949d97f0 by Jasenko Zivanov at 2018-09-05T08:49:06+01:00
symmetric aberration fit now uses a mag- and tilt-corrected prediction
- - - - -
34ffdc5c by Jasenko Zivanov at 2018-09-05T16:26:31+01:00
backprojector can no longer write outside of the data and weight volumes (relevant in the presence of a magnification matrix)
- - - - -
0d6516a9 by Jasenko Zivanov at 2018-09-05T17:19:22+01:00
forward-projector (NN and Lerp) as well as NN-backprojector now also safe from out-of-bounds access
- - - - -
107f7590 by Jasenko Zivanov at 2018-09-06T07:52:48+01:00
mag_fit: predicted-image gradient is now computed in 3D and projected into the 2D image in order to avoid the systematic underestimation of the gradient magnitude resulting from linear interpolation (which leads to a systematic overestimation of mag. anisotropy)
- - - - -
e6914cde by Jasenko Zivanov at 2018-09-06T10:58:11+01:00
bugfix for new gradients under phase modulation
- - - - -
d57af264 by Jasenko Zivanov at 2018-09-06T13:14:08+01:00
relion_reposition added (untested)
- - - - -
f1f36b4d by Takanori Nakane at 2018-09-06T17:37:28+01:00
ml_optimiser: repaired --reuse_scratch, which was not doing anything before.
- - - - -
c15860a2 by Jasenko Zivanov at 2018-09-06T20:21:05+01:00
bugfix in relion_reposition
- - - - -
3262f347 by Jasenko Zivanov at 2018-09-07T07:29:48+01:00
relion_particle_select added; --cc_pad option added to particle_reposition
- - - - -
1891ab14 by Jasenko Zivanov at 2018-09-07T07:31:01+01:00
relion_particle_select actually added
- - - - -
5747461c by Jasenko Zivanov at 2018-09-07T11:59:55+01:00
added option to do gridding instead of using a Wiener filter to double_reconstruct
- - - - -
f57f57a0 by Takanori Nakane at 2018-09-10T14:06:08+01:00
flex_analyser: --fn_keepmask is mandatory when --subtract
- - - - -
034de963 by Takanori Nakane at 2018-09-10T14:22:49+01:00
mask_create: repaired --and, --or, --and_not, --or_not (Bitbucket issue #24 reported by Lukas Kater)
- - - - -
88101f9d by Jasenko Zivanov at 2018-09-11T08:00:28+01:00
motion_refine: initial MotionCor2 trajectories are now being used even if only a global trajectory is known (untested)
- - - - -
28e69e72 by Takanori Nakane at 2018-09-11T11:52:21+01:00
LoG picker: clarified help messages.
Also tested Jasenko's previous commit 88101f9.
- - - - -
4311d4af by Jasenko Zivanov at 2018-09-11T13:19:14+01:00
ctf_mask_test added, new way of predicting signal-delocalisation by the CTF implemented
- - - - -
a59283b7 by Takanori Nakane at 2018-09-11T17:59:12+01:00
GUI: clarified meanings of hyper-parameters for Bayesian Polish
- - - - -
bc5b5dc3 by Jasenko Zivanov at 2018-09-12T09:03:27+01:00
An optional spherical mask can now be multiplied with both data and weight before the division in double_reconstruct
- - - - -
5d2fe1d1 by Jasenko Zivanov at 2018-09-12T09:30:15+01:00
bugfix in spherical mask: radius no longer doubled
- - - - -
e7f62b86 by Jasenko Zivanov at 2018-09-12T16:46:40+01:00
merged master into jaz-3.1 using auto-merge
- - - - -
47fe9feb by Sjors Scheres at 2018-09-13T10:05:20+01:00
started removing ori_particles
- - - - -
b7c90c7e by Jasenko Zivanov at 2018-09-13T12:45:01+01:00
per-particle B-factor fit implemented; some files moved around
- - - - -
49f14702 by Jasenko Zivanov at 2018-09-13T12:45:25+01:00
missing files added
- - - - -
92e56d9d by Jasenko Zivanov at 2018-09-13T13:12:28+01:00
more missing files added
- - - - -
34814ea8 by Jasenko Zivanov at 2018-09-13T17:22:19+01:00
min_scale, min_B and max_B parameters added to B-factor fit
- - - - -
a2109866 by Jasenko Zivanov at 2018-09-13T18:39:29+01:00
bugfix in spherical masking in double_reconstruct
- - - - -
5bfcd4ce by Takanori Nakane at 2018-09-14T12:52:18+01:00
align_symmetry: added --only_rot (useful for helices)
- - - - -
60b44e06 by Takanori Nakane at 2018-09-14T12:59:34+01:00
align_symmetry: --only_rot respected in local search
- - - - -
c6615829 by Takanori Nakane at 2018-09-14T15:01:04+01:00
motioncorr_runner: clarified help strings for dose weighting
- - - - -
5c42ecc9 by Jasenko Zivanov at 2018-09-14T15:13:16+01:00
Options --window and --scale added to motion_refine and tested
- - - - -
7eb3ad04 by Jasenko Zivanov at 2018-09-14T15:13:59+01:00
Merge remote-tracking branch 'origin/master'
- - - - -
498e6d88 by Jasenko Zivanov at 2018-09-14T15:16:06+01:00
Merge remote-tracking branch 'origin/master' into jaz_3.1
- - - - -
a8223488 by Jasenko Zivanov at 2018-09-14T15:31:32+01:00
motion_refine: new magnification now written into star-file when --scale is used
- - - - -
c8f28cb6 by Jasenko Zivanov at 2018-09-14T15:32:18+01:00
debugging logs removed
- - - - -
2b21ed2e by Jasenko Zivanov at 2018-09-14T15:35:18+01:00
merge with master #1
- - - - -
24a1d72a by Jasenko Zivanov at 2018-09-14T15:35:23+01:00
Merge remote-tracking branch 'origin/master' into jaz_3.1
- - - - -
2adc9f0c by Sjors Scheres at 2018-09-14T16:36:19+01:00
keep removing movie stuff and ori_particles
- - - - -
0c50f116 by Jasenko Zivanov at 2018-09-15T11:02:39+01:00
relion_plot_delocalisation added and made user-interpretable
- - - - -
e2b08835 by Jasenko Zivanov at 2018-09-15T11:17:05+01:00
bugfix in relion_plot_delocalisation
- - - - -
385928aa by Jasenko Zivanov at 2018-09-15T13:00:47+01:00
plot_delocalisation: min_freq option added and per-particle feature implemented
- - - - -
772e8639 by Jasenko Zivanov at 2018-09-15T13:11:59+01:00
plot_delocalisation: pretty-printing in plot improved
- - - - -
5a10f6c5 by Jasenko Zivanov at 2018-09-15T13:54:28+01:00
per-particle B-factor fitting slightly improved
- - - - -
7e5c8d7f by Jasenko Zivanov at 2018-09-15T17:22:18+01:00
B-factor fit debugged and accelerated
- - - - -
7683ed82 by Jasenko Zivanov at 2018-09-15T19:36:11+01:00
B-factor fit better integrated into ctf_refine; B-factor plotting added
- - - - -
c9844ceb by Jasenko Zivanov at 2018-09-16T14:42:20+01:00
B-factor estimation: accelerated version numerically safer now; non-accelerated version just as precise as the accelerated one now
- - - - -
71429aab by Jasenko Zivanov at 2018-09-16T17:11:01+01:00
B-factor estimation: added per-micrograph estimation and diagnostic plots for both estimation types
- - - - -
dbce4dd6 by Jasenko Zivanov at 2018-09-16T17:48:53+01:00
B-factor diagnostic plots more meaningful (excessive y-range avoided)
- - - - -
1dd39fc0 by Jasenko Zivanov at 2018-09-17T10:37:22+01:00
ring-fits in B-factor plots no longer clipped
- - - - -
70e2c8f4 by Sjors Scheres at 2018-09-17T11:57:33+01:00
removed movies from refine_mpi
- - - - -
959fd50f by Jasenko Zivanov at 2018-09-17T14:14:25+01:00
GPL preamble added to all files in src/jaz (and all subdirectories written by JaZ)
- - - - -
72945256 by Jasenko Zivanov at 2018-09-17T15:40:48+01:00
bugfix in B-factor-estimator parallelisation
- - - - -
4efbc46e by Jasenko Zivanov at 2018-09-17T22:11:08+01:00
double_reconstruct can now optionally write out weight volumes
- - - - -
8729fe21 by Jasenko Zivanov at 2018-09-17T22:29:01+01:00
bugfix in weight-volume downsampling
- - - - -
abd833e4 by Jasenko Zivanov at 2018-09-17T23:17:15+01:00
weight-volume value-scale now invariant under changes in padding factor
- - - - -
1a4f5c8d by Sjors Scheres at 2018-09-18T10:06:53+01:00
removed polishing and movies from ml_optimiser
- - - - -
8670539d by Sjors Scheres at 2018-09-18T10:08:16+01:00
removed spurious debugging statement. now 2d_small_tagcase works fine
- - - - -
4dc3e921 by Sjors Scheres at 2018-09-18T12:14:29+01:00
remove old polishig code
- - - - -
4c52af16 by Sjors Scheres at 2018-09-18T12:15:03+01:00
fixed some indentation
- - - - -
872ce3bd by Sjors Scheres at 2018-09-18T12:15:32+01:00
got rid of randomised_particle order vector
- - - - -
95b21ddb by Sjors Scheres at 2018-09-18T15:59:26+01:00
minor change to keep exactly the same
- - - - -
e148d873 by Sjors Scheres at 2018-09-18T15:59:42+01:00
updated version to 3.1-devel-scheres
- - - - -
ba33c88d by Sjors Scheres at 2018-09-18T16:01:29+01:00
removed movie frames from accelerated part of the code
- - - - -
69cc8f67 by Jasenko Zivanov at 2018-09-18T19:54:47+01:00
self-subtraction from same-halfset-prediction implemented (contains a heuristic)
- - - - -
e57c8c5b by Takanori Nakane at 2018-09-20T11:55:08+01:00
GUI: made the meanings of 'last_frame' clear and consistent (closes GitHub issue #400)
- - - - -
57c76544 by Jasenko Zivanov at 2018-09-20T12:00:17+01:00
CTF value-setting updated on line 5156 of MlOptimiser.cpp
- - - - -
ec58ed04 by Sjors Scheres at 2018-09-20T12:14:08+01:00
Merged master into jaz_3.1_scheres
- - - - -
00daf2ef by Sjors Scheres at 2018-09-20T11:15:55+00:00
Merged in origin/jaz_3.1_scheres (pull request #16)
CTF value-setting updated on line 5156 of MlOptimiser.cpp
- - - - -
b530dd37 by Sjors Scheres at 2018-09-20T12:18:48+01:00
no longer need all columns to be present in loadSafely
- - - - -
b0f6fae4 by Sjors Scheres at 2018-09-20T12:19:31+01:00
Merge branch 'jaz_3.1_scheres' of https://bitbucket.org/scheres/relion-devel-lmb into jaz_3.1_scheres
- - - - -
b9680588 by Sjors Scheres at 2018-09-20T12:24:58+01:00
introduced error when failing to read image
- - - - -
a7ecef50 by Sjors Scheres at 2018-09-20T12:25:34+01:00
renamed to ori_part_id when referring to original order of particles in MDimg
- - - - -
e1c62056 by Sjors Scheres at 2018-09-20T12:26:24+01:00
merged with Jasenkos setValuesByGroup and did the same in calculateExpectedAngularError
- - - - -
c40c1129 by Jasenko Zivanov at 2018-09-20T15:06:18+01:00
last commit before excision of self-subtraction
- - - - -
0b24be2d by Jasenko Zivanov at 2018-09-20T15:19:13+01:00
self-subtraction removed
- - - - -
4b57882d by Jasenko Zivanov at 2018-09-20T15:48:37+01:00
merged with jaz_3.1_scheres
- - - - -
f1e9c255 by Jasenko Zivanov at 2018-09-20T20:21:07+01:00
new particle star-file format implemented
- - - - -
91de027f by Jasenko Zivanov at 2018-09-20T21:13:00+01:00
All Jaz 3.1 programs adapted to work with offsets given in Angstrom
- - - - -
781e81ea by Jasenko Zivanov at 2018-09-20T21:38:06+01:00
Missing files added
- - - - -
0c621f74 by Jasenko Zivanov at 2018-09-20T22:31:00+01:00
ctf_refine now writes the output star-file in the new combined format
- - - - -
65f77f5c by Jasenko Zivanov at 2018-09-20T23:02:43+01:00
metadata-table pairs written out by ObservationModel::save() are now guaranteed to have the correct names
- - - - -
480bcb25 by Jasenko Zivanov at 2018-09-21T11:46:20+01:00
Pixel size now stored in EMDL_IMAGE_PIXEL_SIZE instead of EMDL_MLMODEL_PIXEL_SIZE
- - - - -
785ffc8d by Sjors Scheres at 2018-09-21T11:12:59+00:00
Merged in jaz_3.1 (pull request #17)
Jaz 3.1
- - - - -
eb550b0b by Sjors Scheres at 2018-09-21T12:30:58+01:00
put movie frames (images) back into the Experiment after discussion with John. Better organisation than before. Still needs to be tested
- - - - -
be85463b by Takanori Nakane at 2018-09-24T11:18:49+01:00
ctffind_runner: supported GCTF 1.18 B1
- - - - -
dbf36481 by Sjors Scheres at 2018-09-24T11:32:35+01:00
putting multiple images per particle back in. Only old cpu-code done.
- - - - -
a0bd8a63 by Sjors Scheres at 2018-09-24T17:00:53+01:00
put multiple images per particle back into accelerated code
- - - - -
8e457131 by Takanori Nakane at 2018-09-24T17:11:46+01:00
star_handler: repaired duplicate removal of down-sampled particles
- - - - -
f4b77c22 by Takanori Nakane at 2018-09-25T14:50:19+01:00
ml_optimiser: fixed (global) resolution estimate in multibody
- - - - -
70d4e9f5 by Takanori Nakane at 2018-09-25T16:40:49+01:00
decomposePipelineSymlinkName: added a warning for a corner case. This should be re-factored in future.
- - - - -
4c20f1d0 by Sjors Scheres at 2018-09-26T11:59:21+01:00
started transforming all offsets to Angstroms instead of pixels... still todo: various apps
- - - - -
2c52e458 by Sjors Scheres at 2018-09-26T14:49:59+01:00
repaired bug in getImagePixelSize
- - - - -
7d7ef874 by Sjors Scheres at 2018-09-26T14:50:19+01:00
delat with multiple optics groups and AB-matrix precalculation
- - - - -
4b4702ad by Sjors Scheres at 2018-09-26T16:07:00+01:00
offsets are in Angstroms in stdout
- - - - -
5fda7c46 by Sjors Scheres at 2018-09-26T16:07:12+01:00
changed offsets to Angsrtoms
- - - - -
20715193 by Sjors Scheres at 2018-09-26T16:07:29+01:00
offsets were already in Angstroms, so only changed label
- - - - -
9d165893 by Sjors Scheres at 2018-09-26T16:08:52+01:00
removed sort job from the GUI
- - - - -
b8e7560f by Sjors Scheres at 2018-09-27T11:57:29+01:00
repaired bug in maingui
- - - - -
c43456c4 by Sjors Scheres at 2018-09-27T15:21:30+01:00
debugging displayer with Angstrom offsets
- - - - -
95fcf226 by Sjors Scheres at 2018-09-27T15:21:52+01:00
make sure names of optics and particles tables are set
- - - - -
ec8c5b15 by Sjors Scheres at 2018-09-27T15:22:16+01:00
offsets in Angstroms
- - - - -
41404c3e by Sjors Scheres at 2018-09-27T15:22:37+01:00
offsets in Angstroms
- - - - -
2aff7a25 by Sjors Scheres at 2018-09-27T16:57:05+01:00
helical track_length in Angstrom
- - - - -
fd48ad47 by Sjors Scheres at 2018-09-28T11:06:02+01:00
changes ori_particle into part_id
- - - - -
478df5df by Takanori Nakane at 2018-09-28T11:17:19+01:00
Testing lowpass-filtering mask
- - - - -
d2465e7b by Takanori Nakane at 2018-09-28T11:55:10+01:00
Autopicker: The LoG picker does not support GPU, causing crash. Block such usage.
- - - - -
41d41e0a by Takanori Nakane at 2018-09-28T15:44:20+01:00
relion_it: don't use GPU for the LoG picker
- - - - -
fce656ab by Sjors Scheres at 2018-09-28T16:03:31+01:00
angstrom offsets, and first try at different image sizes for each optics_group. Still todo: test latter and adapt preprocessing, image-handler etc etc
- - - - -
b26fc956 by Sjors Scheres at 2018-09-28T18:00:44+01:00
debugging different image sizes, 2d small_tagcase still crashing
- - - - -
2bf41055 by Takanori Nakane at 2018-09-28T20:17:52+01:00
PDF generation: fixed synchronisation issue (thanks to Nicolas Coudray)
- - - - -
36b8f76f by Takanori Nakane at 2018-09-28T21:30:31+01:00
duplicate removal: fixed the sign (!)
- - - - -
e7bfa4bb by Sjors Scheres at 2018-10-01T16:20:04+01:00
now 2d small_tagcase is running OK with different possible box sizes for images
- - - - -
463ac45d by Sjors Scheres at 2018-10-01T16:28:33+01:00
repaired small bug in gpu code and removed debugging statement
- - - - -
d27d42b3 by Sjors Scheres at 2018-10-01T17:38:46+01:00
repaired small bug thr_wsum_sigma2_noise
- - - - -
5aa7d385 by Takanori Nakane at 2018-10-01T21:40:42+01:00
relion_it.py: make it more robust
- - - - -
796d1ed0 by Jasenko Zivanov at 2018-10-03T02:41:26+01:00
optics-group support added to aniso-mag refinement
- - - - -
3e622532 by Sjors Scheres at 2018-10-03T09:41:08+01:00
removed spurious function
- - - - -
1a868959 by Sjors Scheres at 2018-10-03T09:42:37+01:00
be careful with rounding my_image_sizes
- - - - -
aca7c468 by Sjors Scheres at 2018-10-03T09:43:28+01:00
repaired various problems, most importantly use coarse size in getsquareddifferencescoarse
- - - - -
4adfc3b8 by Sjors Scheres at 2018-10-03T11:22:13+01:00
removed debugging statements
- - - - -
ec8d86b1 by Sjors Scheres at 2018-10-03T11:22:26+01:00
changed gui to red/pink for devel version
- - - - -
4139e163 by Sjors Scheres at 2018-10-03T11:35:20+01:00
Merged master into jaz_3.1
- - - - -
29eb6e0c by Sjors Scheres at 2018-10-03T11:42:10+01:00
trying to resolve conflict with merging master
- - - - -
637473f6 by Sjors Scheres at 2018-10-03T11:45:41+01:00
still trying to fix conflict
- - - - -
7dc7e16b by Sjors Scheres at 2018-10-03T14:36:07+01:00
added option to keep CTFs from micrographs, instead of from data.star file
- - - - -
d9da43d2 by Sjors Scheres at 2018-10-03T14:37:31+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
2bd84cd5 by Sjors Scheres at 2018-10-03T15:59:42+01:00
moved reading of image sizes from exp_model to star_converter, as only needed when reading old-style particles
- - - - -
94d1f1d5 by Sjors Scheres at 2018-10-04T11:56:00+01:00
repaired ubg for empty groups
- - - - -
f6fd0a93 by Sjors Scheres at 2018-10-04T13:40:40+01:00
initial noise estimation: change how to handle empty groups
- - - - -
798e283f by Jasenko Zivanov at 2018-10-05T03:17:11+01:00
CTF::getGamma made const and FilterHelper::raisedCosEnvCorner2DFull added
- - - - -
eb631358 by Jasenko Zivanov at 2018-10-05T03:19:32+01:00
Merge branch 'jaz_3.1' of https://bitbucket.org/scheres/relion-devel-lmb into jaz_3.1
- - - - -
9b4a0b21 by Jasenko Zivanov at 2018-10-05T03:20:44+01:00
Merge remote-tracking branch 'origin/jaz_3.1_scheres' into jaz_3.1
- - - - -
99c13f15 by Jasenko Zivanov at 2018-10-05T03:30:34+01:00
most tomography code moved to src/jaz/tomo
- - - - -
57d41584 by Jasenko Zivanov at 2018-10-05T23:26:15+01:00
first tomography test added and tomography classes partially revised
- - - - -
d66bacfe by Jasenko Zivanov at 2018-10-06T15:26:22+01:00
presence of images now optional for conversion (a warning is issued in the case of absence)
- - - - -
84c20594 by Jasenko Zivanov at 2018-10-06T18:16:23+01:00
tomo_test updated and cleaned up
- - - - -
6b2427b6 by Jasenko Zivanov at 2018-10-07T19:32:39+01:00
different pixel and box sizes now supported throughout ObservationModel; cached values inside of it are only accessible through getters now; outside code refactored to support those changes (all untested)
- - - - -
c9dcc30f by Jasenko Zivanov at 2018-10-07T20:49:34+01:00
TiltEstimator supports varying pixel and box sizes now; optics-group handling refactored in TiltEstimator
- - - - -
81f624a5 by Jasenko Zivanov at 2018-10-07T21:18:41+01:00
AberrationEstimator supports varying pixel and box sizes now; optics-group handling refactored in AberrationEstimator
- - - - -
e76c5c77 by Jasenko Zivanov at 2018-10-07T22:39:40+01:00
DefocusEstimator supports varying pixel and box sizes now; optics-group handling refactored in DefocusEstimator; ReferenceMap requires a separate pixel size now
- - - - -
3d39b1ce by Jasenko Zivanov at 2018-10-07T23:23:47+01:00
BfactorRefiner supports varying pixel and box sizes now
- - - - -
7b3a4440 by Jasenko Zivanov at 2018-10-08T00:37:13+01:00
MagnificationEstimator supports varying pixel and box sizes now
- - - - -
05886842 by Jasenko Zivanov at 2018-10-08T01:06:36+01:00
missing files added
- - - - -
98275af2 by Jasenko Zivanov at 2018-10-08T03:30:29+01:00
bugfixes in new ReferenceMap::getHollowWeight
- - - - -
8b3c4893 by Jasenko Zivanov at 2018-10-08T20:28:52+01:00
Predictions made from a 3D reference now consider the differences in pixel and box size between the reference and the 2D image
- - - - -
650f5efc by Jasenko Zivanov at 2018-10-09T00:25:46+01:00
Anisotropic magnification now considered within CTFP and Ewald-weight computation
- - - - -
6f3cec3d by Jasenko Zivanov at 2018-10-09T01:29:21+01:00
Ewald sphere is now still spherical even under anisotropic magnification - it was an Ewald ellipsoid until now (only in double_reconstruct)
- - - - -
17cb31e9 by Jasenko Zivanov at 2018-10-09T08:34:59+01:00
CTFP temporarily reverted to a version without anisotropy; Ewald sphere still spherical
- - - - -
022f9cbc by Jasenko Zivanov at 2018-10-09T19:33:26+01:00
bugfix in pixel size difference handling (in predictGradient)
- - - - -
045f325a by Jasenko Zivanov at 2018-10-10T01:03:54+01:00
another bugfix in pixel size difference handling; bugfix in Ewald weights under anisotropic mag
- - - - -
647879b5 by Jasenko Zivanov at 2018-10-10T09:10:46+01:00
added 2D-padding option to double_reconstruct; iterative L1-reconstruction code removed
- - - - -
1519ab04 by Jasenko Zivanov at 2018-10-10T10:55:18+01:00
backprojector can now deal with padded 2D images
- - - - -
d89f32f8 by Jasenko Zivanov at 2018-10-10T10:57:48+01:00
backprojection again limited to Nyquist circle, but the one in 2D
- - - - -
79463f7b by Jasenko Zivanov at 2018-10-10T11:36:52+01:00
changes inside backprojector reverted, compatibility with padded 2D images solved on the outside (untested)
- - - - -
39939927 by Jasenko Zivanov at 2018-10-10T21:22:30+01:00
bugfix in r_max handling - changes are in backprojector again
- - - - -
48dafb87 by Takanori Nakane at 2018-10-11T17:23:54+01:00
GUI: check the presence of the body star file to prevent crash in getCommandsMultiBodyJob().
- - - - -
e6787e0d by Takanori Nakane at 2018-10-11T17:25:26+01:00
relion_project: fixed Fctf initialisation (Thanks to Dimitry Tegunov)
- - - - -
e32fa3b9 by Takanori Nakane at 2018-10-11T17:33:41+01:00
relion_refine: added --keep_scratch
- - - - -
94aba040 by Takanori Nakane at 2018-10-11T17:41:14+01:00
relion_refine: fixed the final message for multi-body refinement (no run_class001.mrc)
- - - - -
205aff3b by Jasenko Zivanov at 2018-10-12T05:30:34+01:00
delocalisation masking implemented and added to double_reconstruct; optional support for colour images added (using libPNG); vis_delocalisation program added; bugfix in CTF::getGammaGrad()
- - - - -
64ec9fa5 by Jasenko Zivanov at 2018-10-12T10:16:58+01:00
Backprojection now correctly considers exactly those points inside the 2D ellipse that map into the 3D sphere given by r_max
- - - - -
2c508250 by Jasenko Zivanov at 2018-10-13T09:35:48+01:00
new Ewald weights implemented and tested
- - - - -
63d39d6b by Jasenko Zivanov at 2018-10-13T11:26:15+01:00
new Ewald weight now optional in double_reconstruct
- - - - -
d0fa688b by Jasenko Zivanov at 2018-10-13T13:10:24+01:00
Mag anisotropy implemented in both versions of MlOptimiser (untested)
- - - - -
79b36a92 by Jasenko Zivanov at 2018-10-14T02:30:33+01:00
bugfix in double_reconstruct
- - - - -
2ef6b0d8 by Jasenko Zivanov at 2018-10-14T02:48:00+01:00
fix in acc/acc_ml_optimiser_impl.h
- - - - -
0c106b8e by Jasenko Zivanov at 2018-10-14T10:23:01+01:00
added Ewald ellipsoid option to double_reconstruct (for comparison)
- - - - -
91e45ea1 by Takanori Nakane at 2018-10-16T16:47:51+01:00
RelionCor: optimised loop order. real-space interpolation is 2-3 times faster! (thanks to Jasenko)
- - - - -
d9686b98 by Jasenko Zivanov at 2018-10-17T17:31:35+01:00
Infinite B-factor bug fixed: now checking for NaNs when merging FCC tables
- - - - -
0ad988fd by Jasenko Zivanov at 2018-10-17T17:56:23+01:00
motion_refine will now replace NaN coordinates in the estimated tracks by zero and warn the user
- - - - -
511ad437 by Takanori Nakane at 2018-10-17T20:01:01+01:00
RelionIt: Added GUI by Colin Palmer
- - - - -
a170f95b by Jasenko Zivanov at 2018-10-18T08:50:41+01:00
merge_particles app added: merges datasets as different optics groups
- - - - -
03a8d97a by Jasenko Zivanov at 2018-10-18T09:36:43+01:00
bugfix in ObsModel::c'tor, beam tilt insertion into Zernike coefficients
- - - - -
69fce15c by Jasenko Zivanov at 2018-10-18T10:00:48+01:00
Sphere condition moved from 2D to 3D in cuda kernel (lines 253 and 312 in BP.cuh)
- - - - -
9d1cd956 by Takanori Nakane at 2018-10-18T14:25:01+01:00
flowchart & layer line: respect RELION_PDFVIEWER_EXECUTABLE
- - - - -
d18b4824 by Takanori Nakane at 2018-10-18T14:41:56+01:00
MotionCor2 wrapper: added warning for -Mag
- - - - -
294ce0d7 by Jasenko Zivanov at 2018-10-18T19:23:06+01:00
Hypothetical cause of NaNs removed in noise whitening (MotionHelper::noiseNormalize)
- - - - -
78d21616 by Takanori Nakane at 2018-10-19T16:08:00+01:00
ml_optimiser: disable AB matrix calculation for --cpu and --gpu. This solves Bitbucket issue #36 by Dimitry Tegunov. Thanks for reporting.
- - - - -
6f54d9c4 by Takanori Nakane at 2018-10-19T16:11:45+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
63b24318 by Takanori Nakane at 2018-10-19T17:12:06+01:00
GUI: define parent jobs of multi-body refinement (solves BitBucket issue #41 reported by @ncoudray.)
- - - - -
877139d9 by Takanori Nakane at 2018-10-19T17:23:28+01:00
Image: fixed integer overflow
- - - - -
7e0a5317 by Takanori Nakane at 2018-10-22T15:48:23+01:00
RelionIt: Add 'Browse' buttons & Save settings with comments (by Colin Palmer)
- - - - -
f7ade6cf by Takanori Nakane at 2018-10-22T16:16:24+01:00
RelionIt: mask size should be in angstrom, also changed the default box size
- - - - -
846cf828 by Takanori Nakane at 2018-10-23T16:59:27+01:00
relion_preprocess: added --use_ctf_in_mic (suggestion from Max Wilkinson)
- - - - -
ef44c82b by Takanori Nakane at 2018-10-25T16:38:36+01:00
RelionIt: Improved GUI, automatic suggestion of box size by Colin Palmer
- - - - -
4eb9ba70 by Takanori Nakane at 2018-10-26T12:51:46+01:00
relion_reconstruct: --mask_diameter is essential for Ewald sphere correction
- - - - -
d70bb4a5 by Jasenko Zivanov at 2018-10-29T11:26:55+00:00
bugfix in mag estimation
- - - - -
1c17ee56 by Sjors Scheres at 2018-10-29T15:01:26+00:00
rounding when getPixelFromResolution
- - - - -
16357499 by Sjors Scheres at 2018-10-29T15:03:54+00:00
resolved conflict upon merging
- - - - -
9a6b96af by Takanori Nakane at 2018-10-29T16:03:29+00:00
relion_image_handler: use the pixel size in the input map by default
- - - - -
d88bb6e5 by Takanori Nakane at 2018-10-29T16:04:06+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
008f2b14 by Jasenko Zivanov at 2018-10-29T19:56:46+00:00
another bugfix in mag estimation
- - - - -
361b281e by Takanori Nakane at 2018-10-30T11:02:40+00:00
RelionIt: Added the mask diameter to the GUI (contribution from Colin Palmer)
- - - - -
2a6e1002 by Takanori Nakane at 2018-11-01T15:54:16+00:00
RelionIt: original box size can be any even number (by Colin Palmer)
- - - - -
8b76c4df by Jasenko Zivanov at 2018-11-02T15:52:57+00:00
star_handler now supports star files with multiple tables (only for reading: only the selected table is written out)
- - - - -
7203f0e1 by Jasenko Zivanov at 2018-11-04T09:29:37+00:00
Bayesian polishing: duplicate particles no longer produce NaNs in the trajectories
- - - - -
72228459 by Sjors Scheres at 2018-11-06T09:35:29+00:00
Merged master into jaz_3.1
- - - - -
8e2d9909 by Sjors Scheres at 2018-11-07T09:18:05+00:00
Merged master into jaz_3.1_scheres
- - - - -
da3bbd65 by Sjors Scheres at 2018-11-07T09:34:24+00:00
switched off check for all same pixel sizes
- - - - -
d530a92a by Sjors Scheres at 2018-11-07T09:43:24+00:00
Merged in jaz_3.1 (pull request #19)
Jaz 3.1
- - - - -
f0bd8090 by Sjors Scheres at 2018-11-07T09:50:25+00:00
Merge branch 'jaz_3.1_scheres' of https://bitbucket.org/scheres/relion-devel-lmb into jaz_3.1_scheres
- - - - -
bae6e171 by Sjors Scheres at 2018-11-07T09:52:48+00:00
moved loop img_id up in getAllSquaredDifferences
- - - - -
f42c3c8d by Takanori Nakane at 2018-11-07T12:15:07+00:00
CtfRefine: fixed NaN in RGB values when there is only one particle in a micrograph. This caused Ghostscript crash.
- - - - -
94f4ea34 by Takanori Nakane at 2018-11-08T15:00:06+00:00
Class2D: fixed an out-of-bound error in initial class assignment
- - - - -
9e1bc5e0 by Takanori Nakane at 2018-11-08T17:21:44+00:00
Extract: should not crash even when there is no input micrograph (to keep RelionIt running)
- - - - -
cc5a2744 by Jasenko Zivanov at 2018-11-12T15:28:11+00:00
all CTF-refinement programs now generate a colour PNG output incl. per-pixel residuals
- - - - -
757820d9 by Takanori Nakane at 2018-11-12T17:41:20+00:00
GUI: improved the help message for the defect file.
- - - - -
86324b2d by Takanori Nakane at 2018-11-12T17:45:53+00:00
ml_optimiser: further workaround for numerical issue in random class assignment. Don't know why this is needed...
- - - - -
dfe42243 by Takanori Nakane at 2018-11-19T13:01:11+00:00
TIFF: Use default values for missing tags where the TIFF specification allows
- - - - -
341ee382 by Sjors Scheres at 2018-11-19T14:00:40+00:00
debugging combining images of different sizes and pixel sizes
- - - - -
51fc4227 by Takanori Nakane at 2018-11-19T14:38:33+00:00
image_handler: allowed TIFF to MRCS conversion. TIFF to MRC does not work, though.
- - - - -
0dc43eb1 by Jasenko Zivanov at 2018-11-20T03:12:18+00:00
anisotropic pixel-weights implemented in symmetric aberration fit
- - - - -
e59b07e5 by Jasenko Zivanov at 2018-11-20T03:14:31+00:00
signs in aniso-mag fit now the same as in the paper
- - - - -
c785cf33 by Sjors Scheres at 2018-11-20T12:27:07+00:00
repaired bug with do_simulate
- - - - -
df4caee1 by Sjors Scheres at 2018-11-20T12:27:38+00:00
debugging multi-sized image sets refinements. Still unequal results with different parallelisations
- - - - -
07a53fb9 by Takanori Nakane at 2018-11-21T14:51:56+00:00
flex_analyser: fixed an error message (issue #417 in GitHub)
- - - - -
27dd7c7d by Sjors Scheres at 2018-11-22T11:50:02+00:00
master shouldn't check disk space
- - - - -
fa02bf69 by Sjors Scheres at 2018-11-22T11:50:26+00:00
small tweak in array size
- - - - -
277dd959 by Sjors Scheres at 2018-11-22T11:51:01+00:00
dont randomise particle order when random_seed equals 0
- - - - -
28476297 by Sjors Scheres at 2018-11-22T11:53:02+00:00
repaired bug in fouriertransformer with unequal size in initial noise estimation
- - - - -
61c5559a by Sjors Scheres at 2018-11-22T11:54:09+00:00
removed suprious debug statement
- - - - -
41a65143 by Sjors Scheres at 2018-11-22T14:39:46+00:00
need to clear transformer if image size changes!
- - - - -
fb19ca2d by Sjors Scheres at 2018-11-22T16:16:02+00:00
repaired bug in writing optimiser table: accuracy trans in angst, and introduced force_new_plans in fftw
- - - - -
d6d55ec2 by Sjors Scheres at 2018-11-22T17:12:52+00:00
replaced MDopt calls for obsModel ones in exp_model, debugged groups size in readImages of mlmodel and changed ori_size call in ml_optimiser_mpi to check for space on GPU card
- - - - -
bc0697b0 by Sjors Scheres at 2018-11-23T15:45:08+00:00
valgrind complained about zz call to 2D offsets
- - - - -
992a3c57 by Sjors Scheres at 2018-11-23T15:45:31+00:00
repaired small bug avg_noise_spectra
- - - - -
3478bcc6 by Jasenko Zivanov at 2018-11-26T08:55:02+00:00
Merge remote-tracking branch 'origin/jaz_3.1_scheres' into jaz_3.1
- - - - -
d4f23bb4 by Jasenko Zivanov at 2018-11-26T09:57:03+00:00
ObsModel::getOpticsGroup() reverted to previous form because the change was based on a misunderstanding
- - - - -
684948bd by Jasenko Zivanov at 2018-11-26T10:27:59+00:00
Cuda BP-kernel: all square/circle conditions in input space replaced by circle/sphere conditions in output space (2D, 3D and SGD) [untested]
- - - - -
dd0d72bd by Jasenko Zivanov at 2018-11-26T10:53:07+00:00
CPU BP-kernels adapted analogously
- - - - -
a7e4b08a by Takanori Nakane at 2018-11-26T13:45:17+00:00
CPlot2D histogram: improved robustness. This fixes crash of motioncorr_runner when there are no 'early' frames.
- - - - -
00a96df0 by Jasenko Zivanov at 2018-11-26T14:03:14+00:00
Non-accelerated forward projection adapted and refactored
- - - - -
2adaba87 by Jasenko Zivanov at 2018-11-26T15:30:08+00:00
Non-accelerated backprojection refactored analogously to forward projection (2D->3D, 2D->2D, 1D->2D and 3D->3D)
- - - - -
72b814ac by Sjors Scheres at 2018-11-27T15:46:30+00:00
continue to debug
- - - - -
fe3f16b4 by Sjors Scheres at 2018-11-27T15:47:35+00:00
Merged in jaz_3.1 (pull request #20)
Jaz 3.1
- - - - -
2efd5ca4 by Sjors Scheres at 2018-11-29T10:03:53+00:00
removed spectra from XA and AA estimates for scale correction; and started going back to ori_size spectral sizes for all groups
- - - - -
f9661996 by Sjors Scheres at 2018-11-29T11:44:12+00:00
cpu works with mixed groups, gpu does not yet
- - - - -
64df905b by Sjors Scheres at 2018-11-29T15:46:56+00:00
added applyScale Differences to GPU code
- - - - -
537725fb by Takanori Nakane at 2018-11-29T17:32:22+00:00
Tentative support for 4-bit MRC (mode 101) from SerialEM.
- - - - -
d11d43a6 by Sjors Scheres at 2018-11-30T09:34:27+00:00
repaired bug Jasenko's BP.cuh: forgot padding factor on max_r2 (2)
- - - - -
43e41783 by Takanori Nakane at 2018-11-30T11:58:14+00:00
image_handler: now STAR file is processed as before
- - - - -
7e96ed8c by Sjors Scheres at 2018-11-30T12:09:20+00:00
sort randomised particles on optics_group to avoid recalculating plan of cudafft all the time
- - - - -
dbcdab0c by Sjors Scheres at 2018-11-30T12:09:42+00:00
removed superfluous bracket in timing code
- - - - -
8688d1e6 by Takanori Nakane at 2018-11-30T12:15:09+00:00
GCTF wrapper: explicitly requires CUDA
- - - - -
428fc5f7 by Sjors Scheres at 2018-11-30T14:42:58+00:00
enforce that model has smallest pixel size and largest box size of all optics_groups in the input particles star file
- - - - -
18fa37a4 by Sjors Scheres at 2018-11-30T15:06:37+00:00
try pinkfor 3.1 colours
- - - - -
8085f40c by Jasenko Zivanov at 2018-12-02T11:28:33+00:00
motion estimation updated to new format (untested, parameter estimation temporarily removed, old b-factor estimation permanently removed); ObsModel::loadSafely no longer returns the optics mdt
- - - - -
d0e47df1 by Jasenko Zivanov at 2018-12-02T12:34:16+00:00
motion refinement: tracks stored in Angstrom now
- - - - -
6ac29519 by Jasenko Zivanov at 2018-12-02T13:09:25+00:00
motion refinement: star file written out in new format
- - - - -
51b33123 by Sjors Scheres at 2018-12-03T09:23:59+00:00
small improvement stdout
- - - - -
9838de03 by Takanori Nakane at 2018-12-03T15:42:52+00:00
GCTF wrapper: complain when phase shift estimation is enabled through 'Other Gctf options' instead of --do_phaseshift.
- - - - -
3c5fed45 by Takanori Nakane at 2018-12-03T17:30:22+00:00
GUI: improved parameter check and help message for GCTF
- - - - -
11875c26 by Sjors Scheres at 2018-12-03T17:46:26+00:00
Merged in jaz_3.1 (pull request #21)
Jaz 3.1
- - - - -
f7fdd455 by Sjors Scheres at 2018-12-04T09:08:34+00:00
Merged in jaz_3.1_scheres (pull request #22)
Merged in jaz_3.1 (pull request #21)
- - - - -
389a01a9 by Sjors Scheres at 2018-12-04T17:00:35+00:00
now import can handle optics groups for micrographs and movies
- - - - -
e7b40594 by Sjors Scheres at 2018-12-04T17:00:57+00:00
Merge branch 'jaz_3.1_scheres' of https://bitbucket.org/scheres/relion-devel-lmb into jaz_3.1_scheres
- - - - -
f2d9a0ae by Sjors Scheres at 2018-12-04T17:44:43+00:00
set output STAR file version to current, not hard-coded 300001
- - - - -
ef2c316e by Sjors Scheres at 2018-12-05T14:39:27+00:00
fixed bug in passing on refined helical twist and rise of half2 upon convergence
- - - - -
9d9157b9 by Sjors Scheres at 2018-12-05T14:45:37+00:00
fixed bug in passing on refined helical twist and rise of half2 upon convergence
- - - - -
36243a1a by Sjors Scheres at 2018-12-05T14:46:23+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
3a737145 by Sjors Scheres at 2018-12-05T16:31:23+00:00
first attempt at motioncor and ctffind working with optics groups
- - - - -
7c7a8743 by Takanori Nakane at 2018-12-07T15:56:30+00:00
preprocess: copy rlnBeamTiltGroupName from the micrograph STAR file only when absent in the particle STAR file
- - - - -
df62c800 by Takanori Nakane at 2018-12-07T15:57:39+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
7fd21af7 by Dimitry Tegunov at 2018-12-07T16:29:41+00:00
3D CTF patch by Dmitry Tegunov: Adds support for 3D CTF volumes in FFTW format (no redundant half, center at 0).
Fixed Conflicts:
src/apps/project.cpp
- - - - -
b90d156a by Dimitry Tegunov at 2018-12-07T16:30:21+00:00
More robust support for 3D CTF in FFTW format
- - - - -
99d96a3e by Takanori Nakane at 2018-12-11T11:42:58+00:00
relion_preprocess: fixed handling of the beam tilt group
- - - - -
0f20d5d3 by Takanori Nakane at 2018-12-11T11:44:19+00:00
Image: fix indentations. MRC MODE 101 confirmed for both EPU and SerialEM
- - - - -
d9e7be92 by Sjors Scheres at 2018-12-11T15:06:25+00:00
dont need tablename as part of class
- - - - -
29b5623d by Sjors Scheres at 2018-12-11T15:06:56+00:00
magnification might not awlays be there...
- - - - -
58d874f8 by Sjors Scheres at 2018-12-11T15:07:46+00:00
small adjustment to deal with v3.1 star files
- - - - -
7b0c29f2 by Sjors Scheres at 2018-12-11T15:31:56+00:00
default particle is -1 for currentObject
- - - - -
b946618d by Sjors Scheres at 2018-12-11T15:32:51+00:00
use optics group for CTFs in autopicker
- - - - -
a94a71af by Sjors Scheres at 2018-12-12T15:37:50+00:00
make sure to use long ints for loops over particles!
- - - - -
e409c1e9 by Sjors Scheres at 2018-12-12T15:48:43+00:00
use long int for pc
- - - - -
ced57e06 by Sjors Scheres at 2018-12-12T17:12:26+00:00
repair issue #49 on relion3.0-beta
- - - - -
98d90ed9 by Sjors Scheres at 2018-12-12T17:12:54+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
90561b4e by Jasenko Zivanov at 2018-12-12T18:36:42+01:00
astigmatism-adaptation bug for empty matrices (probably) fixed
- - - - -
911f5860 by Jasenko Zivanov at 2018-12-12T18:37:33+01:00
Merge branch 'jaz_3.1' of https://bitbucket.org/scheres/relion-devel-lmb into jaz_3.1
- - - - -
3fd8a058 by Sjors Scheres at 2018-12-13T09:59:12+00:00
copied takanoris new image.h with 4-bit MRC support to avoid merging conflicts
- - - - -
f73705b7 by Sjors Scheres at 2018-12-13T10:01:52+00:00
trying to avoid merging conflict
- - - - -
5dd0ebbc by Sjors Scheres at 2018-12-13T10:08:14+00:00
trying to resolve conflicts in merge with master
- - - - -
8ebd620d by Sjors Scheres at 2018-12-13T10:09:13+00:00
trying to resolve conflicts in merge with master
- - - - -
6f0d208a by Sjors Scheres at 2018-12-13T10:13:34+00:00
trying to resolve merge conflict
- - - - -
edf14e49 by Sjors Scheres at 2018-12-13T10:14:03+00:00
Merged in master (pull request #23)
Master
- - - - -
960a1fe4 by Sjors Scheres at 2018-12-13T10:48:35+00:00
repaired errors after merging
- - - - -
13f09198 by Takanori Nakane at 2018-12-13T11:50:30+00:00
Get rid of rlnBeamTiltGroupName (as string). Use rlnBeamTiltClass (as int) instead.
- - - - -
6a1a78a9 by Takanori Nakane at 2018-12-13T11:56:11+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
3b7de54b by Sjors Scheres at 2018-12-13T15:58:38+00:00
preprocessing now with obsModel
- - - - -
ea30fb5c by Sjors Scheres at 2018-12-14T12:49:22+00:00
added colour display to relion_display, with Jasenkos colour scheme
- - - - -
615a6b1d by Sjors Scheres at 2018-12-14T15:24:05+00:00
added rainbow colours too
- - - - -
5dc61576 by Takanori Nakane at 2018-12-17T12:11:26+00:00
UCSF MotionCor2 wrapper: allow multiple GPUs per process
- - - - -
1b3b6467 by Takanori Nakane at 2018-12-17T16:23:43+00:00
RelionCor: more efficient gain correction
- - - - -
5f152285 by Sjors Scheres at 2018-12-17T17:44:50+00:00
changed black/white to min/max on display GUI
- - - - -
2c027f2b by Takanori Nakane at 2018-12-18T12:00:35+00:00
CTFFIND/GCTF wrapper: Don't read the pixel size from log files to avoid loss of precision
- - - - -
b3041a1b by Takanori Nakane at 2018-12-18T14:11:44+00:00
RelionCor: make hard-coded max_iter to a command line argument, removed timing code
- - - - -
31e4017b by Takanori Nakane at 2018-12-18T16:37:17+00:00
GUI: improved help for duplicate removal. Also added one missing file.
- - - - -
656344e6 by Takanori Nakane at 2018-12-19T11:07:21+00:00
GUI: fixed crash in Particle Extraction (reported by @kehrnal as GitHub issue #423)
- - - - -
97b3a493 by Takanori Nakane at 2018-12-19T12:13:18+00:00
relion_display: save image as PNG
- - - - -
06bd9d59 by Sjors Scheres at 2018-12-19T12:16:39+00:00
Merge branch 'jaz_3.1_scheres' of https://bitbucket.org/scheres/relion-devel-lmb into jaz_3.1_scheres
- - - - -
4810be4b by Sjors Scheres at 2018-12-19T12:47:02+00:00
added scalebar for colours and blackred and bluewhite schemes
- - - - -
9e446249 by Takanori Nakane at 2018-12-19T14:56:41+00:00
Merge branch 'master' into jaz_3.1_scheres
Conflicts:
src/metadata_label.h
src/preprocessing.cpp
- - - - -
d202ffa5 by Takanori Nakane at 2018-12-19T14:57:58+00:00
Merge branch 'jaz_3.1_scheres' of https://bitbucket.org/scheres/relion-devel-lmb into jaz_3.1_scheres
- - - - -
c5e8e3f1 by Sjors Scheres at 2018-12-19T15:07:22+00:00
renamed colours to fire, ice etc and added fire-n-ice
- - - - -
3cd2912c by Sjors Scheres at 2018-12-19T15:08:40+00:00
added discover for tablename, and repaired bug jasenko in getPhaseCorrection and in getGammaOffset
- - - - -
0aca3da7 by Sjors Scheres at 2018-12-19T15:09:49+00:00
Merge branch 'jaz_3.1_scheres' of https://bitbucket.org/scheres/relion-devel-lmb into jaz_3.1_scheres
- - - - -
cfe4819e by Sjors Scheres at 2018-12-19T15:32:47+00:00
allow opening of 3.1 GUI in old directopries
- - - - -
4d05f527 by Takanori Nakane at 2018-12-19T16:07:43+00:00
relion_reconstruct: noise reconstruction
- - - - -
c58bac76 by Takanori Nakane at 2018-12-19T16:19:45+00:00
ctf_toolbox: fix indentations
- - - - -
7c0f9e9b by Takanori Nakane at 2018-12-19T16:30:11+00:00
GUI: fix the number of menu items
- - - - -
1c21c0e8 by Sjors Scheres at 2018-12-19T17:05:31+00:00
greyToRGB as inline, and remove if-statement from loop over all pixels to outside the loop
- - - - -
cb2288cd by Sjors Scheres at 2018-12-19T17:09:39+00:00
resolving conflict
- - - - -
69d06897 by Sjors Scheres at 2018-12-19T17:12:00+00:00
changed argument to --colour_bar
- - - - -
bed1ed00 by Takanori Nakane at 2018-12-19T17:39:09+00:00
Removed test programs from standard build. (Please don't enable RELION_TEST in CMakeFile. Use cmake -DRELION_TEST=ON)
- - - - -
e55fe70d by Takanori Nakane at 2018-12-19T17:39:57+00:00
relion_display: show label (--text_label)
- - - - -
a51d12dc by Takanori Nakane at 2018-12-19T17:40:23+00:00
Merge branch 'jaz_3.1_scheres' of https://bitbucket.org/scheres/relion-devel-lmb into jaz_3.1_scheres
- - - - -
93b5a5c4 by Sjors Scheres at 2018-12-19T17:44:26+00:00
carry through colour scheme to show original image
- - - - -
8ee1bd4d by Sjors Scheres at 2018-12-19T17:44:44+00:00
Merge branch 'jaz_3.1_scheres' of https://bitbucket.org/scheres/relion-devel-lmb into jaz_3.1_scheres
- - - - -
12d99575 by Sjors Scheres at 2018-12-20T08:36:35+00:00
change button on display window to pink
- - - - -
db3060b3 by Sjors Scheres at 2018-12-20T09:39:17+00:00
make orange background image with PHF
- - - - -
4a64f498 by Sjors Scheres at 2018-12-20T09:39:42+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
60eca62a by Sjors Scheres at 2018-12-20T11:02:13+00:00
background 3.0
- - - - -
5a69b484 by Sjors Scheres at 2018-12-20T11:42:16+00:00
repaired bug downscaled multicolor images
- - - - -
498047ff by Sjors Scheres at 2018-12-20T11:42:33+00:00
new background image position
- - - - -
3a635268 by Sjors Scheres at 2018-12-20T11:42:52+00:00
small bug necessary pixel sizes
- - - - -
5756ffdb by Sjors Scheres at 2018-12-20T11:50:42+00:00
repaired bug to get pixel size of micrographs from obsModel
- - - - -
da32e4a3 by Sjors Scheres at 2018-12-20T11:59:18+00:00
use loadSafely of input STAR file
- - - - -
bbe72b22 by Sjors Scheres at 2018-12-20T12:00:29+00:00
use obsModel.save for output STAR file
- - - - -
742a444e by Sjors Scheres at 2018-12-20T12:10:37+00:00
repaired bug box size of extracted particles upon rescaling or windowing
- - - - -
14a376d0 by Takanori Nakane at 2018-12-20T12:25:25+00:00
Merge branch 'refine-lowpass' into jaz_3.1_scheres
- - - - -
030428f1 by Sjors Scheres at 2018-12-20T12:26:19+00:00
repair bug getting CTF info when prereading images
- - - - -
bf2c1597 by Takanori Nakane at 2018-12-20T12:31:34+00:00
relion_refine: fn_something.contains("None") is dangerous.
- - - - -
80428b9c by Sjors Scheres at 2018-12-20T12:32:02+00:00
repair bug image_full_size not exist yet in calculate power spectra
- - - - -
77b5b9a7 by Takanori Nakane at 2018-12-20T12:32:58+00:00
Merge branch 'jaz_3.1_scheres' of https://bitbucket.org/scheres/relion-devel-lmb into jaz_3.1_scheres
- - - - -
3483585f by Sjors Scheres at 2018-12-20T12:35:40+00:00
started to change star handler for handling optics groups. not finished yet: duplicate removal
- - - - -
faea0ae7 by Sjors Scheres at 2018-12-20T12:51:03+00:00
typo in option rainbow color original image
- - - - -
1aff028e by Takanori Nakane at 2018-12-23T20:52:23+00:00
Fixed compilation on GCC 6.4.0 without -fpermissive (reported by Rafael)
- - - - -
7251d08a by Takanori Nakane at 2018-12-24T12:01:45+00:00
relion_import: added
- - - - -
ecc42d9c by Takanori Nakane at 2018-12-25T08:20:59+00:00
CtfRefine: repaired PDF output
- - - - -
814a673c by Takanori Nakane at 2018-12-25T08:29:25+00:00
ReferenceMap: Use the pixel size in the reference map by default
- - - - -
7005ddb9 by Takanori Nakane at 2018-12-25T11:53:16+00:00
WIP: Repairing particle re-extraction. Need careful testing!
- - - - -
f00119a8 by Takanori Nakane at 2018-12-25T17:39:34+00:00
relion_preprocess: confirmed re-extraction works. TODO: optics groups from run_data.star should be prioritized.
- - - - -
560bddc0 by Takanori Nakane at 2019-01-02T15:10:28+00:00
Give an error when the gain reference is in Digital Micrograph format
- - - - -
ce40fda2 by Takanori Nakane at 2019-01-03T15:37:43+00:00
auto-refine: fixed final messages (thanks to Oliver Clarke for reporting). Removed a dead code.
- - - - -
de24e1d1 by Takanori Nakane at 2019-01-03T17:22:40+00:00
RelionCor: WIP: calculate power-spectrum for CTF estimation
- - - - -
df5a0cf2 by Takanori Nakane at 2019-01-03T17:26:37+00:00
Merge branch 'master' into power-spectrum
Conflicts:
src/ml_optimiser.cpp
- - - - -
4af43bbb by Takanori Nakane at 2019-01-04T12:20:33+00:00
Cherry-picked "relion_reconstruct: noise reconstruction" to fix issue #53.
- - - - -
ca355e1f by Takanori Nakane at 2019-01-04T14:11:31+00:00
motion correction: removed UNBLUR support. fixed --only_do_unfinished
- - - - -
407a9541 by Takanori Nakane at 2019-01-04T14:14:55+00:00
Merge branch 'master' into power-spectrum
- - - - -
cc0025d6 by Sjors Scheres at 2019-01-04T14:31:03+00:00
resolve issue #16 Refine3D do unnecessary rewindowing
- - - - -
8b7884cd by Sjors Scheres at 2019-01-04T14:31:42+00:00
Merge branch 'jaz_3.1_scheres' of https://bitbucket.org/scheres/relion-devel-lmb into jaz_3.1_scheres
- - - - -
1f105469 by Sjors Scheres at 2019-01-04T14:45:47+00:00
resolve issue #15 Refine3D should have --ref_angpix
- - - - -
b10110fc by Takanori Nakane at 2019-01-04T17:51:08+00:00
motion correction writes power spectra (--grouping_for_ps > 0) and CTFFIND uses it (--use_given_ps).
- - - - -
9ae9b9e4 by Sjors Scheres at 2019-01-07T12:10:28+00:00
start implementation of obsmodel in manual picker, not done yet
- - - - -
bc082bbf by Sjors Scheres at 2019-01-07T12:11:21+00:00
dont warn when reading pixel size from header of reference
- - - - -
193664b1 by Takanori Nakane at 2019-01-07T13:56:46+00:00
CtfFind: fixed the pixel size of power spectrum. Now this is working, but need testing.
- - - - -
a3fc626b by Sjors Scheres at 2019-01-07T14:03:54+00:00
add stochastic EM AND repair bug in maximiseOtherParam
- - - - -
cd8a96d4 by Sjors Scheres at 2019-01-07T14:04:14+00:00
Merge branch 'jaz_3.1_scheres' of https://bitbucket.org/scheres/relion-devel-lmb into jaz_3.1_scheres
- - - - -
230a8e7b by Sjors Scheres at 2019-01-08T11:59:51+00:00
add obsModel to manual picker
- - - - -
e1e995f5 by Sjors Scheres at 2019-01-08T12:00:32+00:00
add stochastic EM to GUI, and have non-negative gradients
- - - - -
2c77a665 by Takanori Nakane at 2019-01-08T12:10:13+00:00
relion_display: fixed 'Show Metadata'
- - - - -
d9d2d207 by Sjors Scheres at 2019-01-08T14:05:54+00:00
added obsModel to display program
- - - - -
b0dbd97a by Sjors Scheres at 2019-01-08T14:06:13+00:00
Merge branch 'jaz_3.1_scheres' of https://bitbucket.org/scheres/relion-devel-lmb into jaz_3.1_scheres
- - - - -
0ad5bd01 by Takanori Nakane at 2019-01-08T16:47:27+00:00
relion_pipeliner: added RELION_MPIRUN environmental variable to change hard-coded 'mpirun'. For qsub jobs, the script template is used (as before).
- - - - -
bfab4105 by Takanori Nakane at 2019-01-09T11:33:55+00:00
GUI: fixed the output node name for flex_analyse. Solved issue #55 on BitBucket reported by Matt I.
- - - - -
8f0fdd82 by Takanori Nakane at 2019-01-09T17:27:56+00:00
Patch by Colin Palmer (CCPEM) to put RELION's GUI background image in-line, since otherwise it tries
to look for the pixmap file using the absolute paths of the source and install directories from the build machine, making the binary non-portable.
- - - - -
8229ffed by Takanori Nakane at 2019-01-10T13:47:52+00:00
Made sure RelionError are handled correctly.
- - - - -
78817717 by Takanori Nakane at 2019-01-10T14:02:04+00:00
Merge branch 'master' into jaz_3.1_scheres.
Conflicts:
src/apps/cs_fit.cpp
src/apps/ctf_toolbox.cpp
src/apps/free_aberration_plot.cpp
src/apps/helix_toolbox.cpp
src/apps/motion_diff.cpp
src/apps/motion_stats.cpp
src/apps/particle_polish.cpp
src/apps/particle_polish_mpi.cpp
src/apps/particle_sort.cpp
src/apps/particle_sort_mpi.cpp
src/apps/ref_aberration_plot.cpp
src/apps/refine_mpi.cpp
src/apps/split_stack.cpp
- - - - -
46d525db by Takanori Nakane at 2019-01-10T14:19:42+00:00
clean clutter (in my previous commit, my editor replaced ALL spaces into tabs ...)
- - - - -
c3e92324 by Takanori Nakane at 2019-01-10T14:30:41+00:00
Merge branch 'master' into jaz_3.1_scheres
Conflicts:
src/apps/free_aberration_plot.cpp
src/apps/helix_toolbox.cpp
- - - - -
1de1d206 by Sjors Scheres at 2019-01-10T16:55:07+00:00
add aberrations to the GUI
- - - - -
b4c823ea by Takanori Nakane at 2019-01-14T11:14:01+00:00
ml_optimiser: repaired continuation
- - - - -
5a7ee156 by Takanori Nakane at 2019-01-14T11:57:07+00:00
fixed a warning message
- - - - -
943a876c by Takanori Nakane at 2019-01-15T10:53:56+00:00
RelionIt: fixed crash when "images_are_movies = False" (reported by Yongchan Lee)
- - - - -
da007b97 by Takanori Nakane at 2019-01-15T12:49:52+00:00
relion_run_ctffind: Use the RELION_SHELL environmental variable to specify a shell other than csh
- - - - -
f7450afd by Sjors Scheres at 2019-01-16T17:21:01+00:00
remove polish and movierefine windows from gui library
- - - - -
d9141f7a by Sjors Scheres at 2019-01-16T17:21:59+00:00
remove polish and movierefine jobs from pipeliner
- - - - -
33dd79c4 by Sjors Scheres at 2019-01-18T11:04:52+00:00
removed polish and movierefine
- - - - -
77651855 by Sjors Scheres at 2019-01-21T17:16:39+00:00
started working on GUI
- - - - -
2b5fd268 by Charles Congdon at 2019-01-25T11:21:18+00:00
Corrections/simplifications to README.md
- - - - -
48109d53 by Charles Congdon at 2019-01-25T11:21:18+00:00
Fix a few more cases of data size truncation
- - - - -
97348063 by Dari Kimanius at 2019-01-25T11:21:18+00:00
Class/body-specific symmetry
- - - - -
6eb57f6a by Dari Kimanius at 2019-01-25T11:21:19+00:00
Issue in cmake variable assignment order
- - - - -
c9c7c9b6 by Dari Kimanius at 2019-01-25T11:21:19+00:00
MPI related depricated cmake variable replaced
- - - - -
ac81250c by Takanori Nakane at 2019-01-25T11:21:19+00:00
Fixed build on cmake 2.8
- - - - -
916a4928 by Takanori Nakane at 2019-01-25T11:40:45+00:00
Added explanation of FORCE_OWN_FFTW and FORCE_OWN_FLTK.
- - - - -
855201a6 by Sjors Scheres at 2019-01-29T09:06:00+00:00
swapped tail and awk order run.out
- - - - -
bb2d5a10 by Sjors Scheres at 2019-01-29T09:07:10+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
709b4b3c by Takanori Nakane at 2019-01-29T12:08:42+00:00
MicrographHandler: use ObsModel to read corrected_micrographs.star
- - - - -
3f89d54e by Takanori Nakane at 2019-01-29T17:39:04+00:00
duplicate removal: migrated to ObsModel
- - - - -
f170ab5b by Sjors Scheres at 2019-01-30T11:11:18+00:00
new menu options handling for GUI, write run.job.star files
- - - - -
551350ba by Sjors Scheres at 2019-01-30T11:20:58+00:00
resolved conflict
- - - - -
c7c58559 by Sjors Scheres at 2019-01-30T11:25:28+00:00
resolved conflict 2
- - - - -
6a2d7905 by Takanori Nakane at 2019-01-30T17:28:26+00:00
bfactor_plot.py: Migrated to 3.1 STAR files. Increased the number of digits.
- - - - -
b8179ae6 by Takanori Nakane at 2019-01-30T17:29:06+00:00
Merge branch 'jaz_3.1_scheres' of https://bitbucket.org/scheres/relion-devel-lmb into jaz_3.1_scheres
- - - - -
950c7012 by Sjors Scheres at 2019-01-31T09:30:25+00:00
set menu to first item if cannot find loaded value
- - - - -
963fa84c by Sjors Scheres at 2019-01-31T09:30:51+00:00
remove slash from menu, as has special meaning in fltk
- - - - -
9e0c3290 by Sjors Scheres at 2019-01-31T09:31:31+00:00
allow variable names in pipeliner command-line options
- - - - -
56877d84 by Sjors Scheres at 2019-01-31T09:31:52+00:00
Merge branch 'jaz_3.1_scheres' of https://bitbucket.org/scheres/relion-devel-lmb into jaz_3.1_scheres
- - - - -
8603a15e by Sjors Scheres at 2019-01-31T09:38:17+00:00
negative twist in helical symmetry expand
- - - - -
06a0c8fb by Sjors Scheres at 2019-01-31T09:39:26+00:00
negative twist in helical symmetry expand
- - - - -
024443c1 by Sjors Scheres at 2019-01-31T17:22:04+00:00
removed comment didnt work yet
- - - - -
694bac6e by Sjors Scheres at 2019-01-31T17:22:28+00:00
allow overwriting existing jobs, without continuation
- - - - -
2553d34d by Sjors Scheres at 2019-02-01T09:26:22+00:00
write star files with .tmp filename first, then rename
- - - - -
46b5baca by Sjors Scheres at 2019-02-01T09:27:10+00:00
die for combining: 3.1 star files not done yet
- - - - -
cb0c0fb7 by Sjors Scheres at 2019-02-01T09:27:26+00:00
warn for combining: 3.1 star files not done yet
- - - - -
784536bd by Sjors Scheres at 2019-02-01T09:27:52+00:00
added anisotropic magnification on CTF-refine GUI
- - - - -
42c5e941 by Sjors Scheres at 2019-02-01T09:35:15+00:00
changed overwrite button and ask for confirmation to overwrite
- - - - -
e1251460 by Sjors Scheres at 2019-02-06T13:55:13+00:00
added option to modulate instead of demodulatePhase, this is useful to apply to reference projections before subtraction from experimental images
- - - - -
af0f387f by Sjors Scheres at 2019-02-06T17:35:17+00:00
check for half-FFTW CTFs
- - - - -
5fe721af by Takanori Nakane at 2019-02-06T20:21:36+00:00
relion_image_handler: renumber output stack IDs when reading and writing a STAR file
- - - - -
81ce34f1 by Takanori Nakane at 2019-02-11T19:33:10+00:00
ml_optimiser: fixed crashes in MlModel::pack
- - - - -
6bc8f5e5 by Takanori Nakane at 2019-02-11T19:33:51+00:00
Merge branch 'jaz_3.1_scheres' of https://bitbucket.org/scheres/relion-devel-lmb into jaz_3.1_scheres
- - - - -
dc0810e7 by Sjors Scheres at 2019-02-13T09:43:10+00:00
Merge branch 'jaz_3.1_scheres' of https://bitbucket.org/scheres/relion-devel-lmb into jaz_3.1_scheres
- - - - -
b85ae812 by Takanori Nakane at 2019-02-13T11:26:08+00:00
CtfRefine: warn when refining beam tilt in a Cs corrected scope.
- - - - -
decd5719 by Takanori Nakane at 2019-02-13T11:43:04+00:00
Merge branch 'beta-repo-3.0' into 3.0
- - - - -
6ddca6af by Takanori Nakane at 2019-02-13T12:06:06+00:00
Call this 3.0; Added documentation about libtiff.
- - - - -
ff5413de by Sjors Scheres at 2019-02-13T15:04:42+00:00
added libLBFGS to AUTHORS file
- - - - -
64ae9b2c by Sjors Scheres at 2019-02-13T15:05:02+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
21d17742 by Sjors Scheres at 2019-02-13T16:36:20+00:00
extract possibly in bigger box when doing (per-particle) CTF premultiplication
- - - - -
78dfb812 by Sjors Scheres at 2019-02-13T16:37:39+00:00
option to prevent preread of images, useful for reading optimiser.star for other programs, and backwards compatibility with chnages in translations in Angstroms
- - - - -
35719706 by Sjors Scheres at 2019-02-13T16:38:08+00:00
prepare fr v3.1
- - - - -
cda848bd by Sjors Scheres at 2019-02-13T16:44:34+00:00
now works with obsModel
- - - - -
984c9158 by Takanori Nakane at 2019-02-13T17:32:26+00:00
Added acknowledgements to AUTHORS.
- - - - -
600499f3 by Takanori Nakane at 2019-02-13T17:34:50+00:00
Merge branch 'beta-repo-3.0' into 3.0
- - - - -
4df5e597 by Takanori Nakane at 2019-02-14T14:19:35+00:00
MotionCor2 wrapper: remove *_DWS.mrc produced in >= 1.2.0
- - - - -
4629b844 by Takanori Nakane at 2019-02-14T15:06:31+00:00
CtfRefine & BayesianPolish: get file names of half maps and the mask from a postprocess STAR file. This enables scheduling of these jobs. Solves issue #435 on GitHub.
- - - - -
ace90405 by Takanori Nakane at 2019-02-14T15:10:19+00:00
Updated README
- - - - -
5d395e76 by Takanori Nakane at 2019-02-14T15:18:00+00:00
Merge branch 'bitbucket-3.0'
- - - - -
3cfa14b8 by Takanori Nakane at 2019-02-14T17:18:43+00:00
moved cropInFourierSpace to fftw.h
- - - - -
11ec3bbb by Sjors Scheres at 2019-02-15T11:34:17+00:00
used long long int for nr_directions
- - - - -
c750f816 by Takanori Nakane at 2019-02-15T11:34:58+00:00
Merge branch 'jaz_3.1_scheres' of https://bitbucket.org/scheres/relion-devel-lmb into jaz_3.1_scheres
- - - - -
3f853be8 by Sjors Scheres at 2019-02-15T12:31:01+00:00
added obsModel to project
- - - - -
64348a26 by Sjors Scheres at 2019-02-15T12:31:51+00:00
update pixel size and image size in the optics group table if new_box or rescale_angpix
- - - - -
d0705fa0 by Sjors Scheres at 2019-02-15T12:37:49+00:00
added new particle_subtract program, both for multibodies and normal partial signal subtraction
- - - - -
906beb09 by Sjors Scheres at 2019-02-15T12:38:37+00:00
removed subtraction functionality from flex_analyser
- - - - -
87025089 by Sjors Scheres at 2019-02-15T16:59:39+00:00
added relion_particle_subtract to the GUI
- - - - -
d9e20d9f by Takanori Nakane at 2019-02-15T17:05:38+00:00
Merge branch 'jaz_3.1_scheres' of https://bitbucket.org/scheres/relion-devel-lmb into jaz_3.1_scheres
- - - - -
d7451af9 by Takanori Nakane at 2019-02-15T17:43:36+00:00
GUI: fixed typo
- - - - -
059e659d by Takanori Nakane at 2019-02-15T20:19:46+00:00
Merge branch 'master' into jaz_3.1_scheres
Conflicts:
src/acc/acc_helper_functions_impl.h
src/apps/image_handler.cpp
src/jaz/ctf/ctf_refiner.cpp
src/jaz/motion/motion_refiner.cpp
src/jaz/reference_map.cpp
src/macros.h
src/pipeline_jobs.cpp
src/pipeline_jobs.h
- - - - -
cc3c11ea by Jasenko Zivanov at 2019-02-18T03:12:01+01:00
motion parameter estimation restored: all optics groups can be considered simultaneously for parameter hyper-optimisation if they have the same pixel and box size. Otherwise, a single group can be selected using --par_group. The output filename is now called 'opt_params_group_N.txt' or 'opt_params_all_groups.txt', respectively. For the alignment itself, --only_group can be used to only align micrographs containing particles from the indicated group.
- - - - -
dcd265e5 by Jasenko Zivanov at 2019-02-19T01:37:18+01:00
motion_refine: images are now by default recombined at the pixel and box size indicated in the star file
- - - - -
afe297be by Jasenko Zivanov at 2019-02-21T00:43:50+01:00
ctf_refine: symmetrical and antisymmetrical aberrations are now estimated after the magnification
- - - - -
1ed6bd9a by Takanori Nakane at 2019-02-21T11:59:09+00:00
Merge branch 'jaz_3.1'
Conflicts:
src/jaz/obs_model.h
- - - - -
a39cf449 by Takanori Nakane at 2019-02-21T12:06:45+00:00
Polish: remove addition of an uninitialised magnification column
- - - - -
cbcd62ac by Jasenko Zivanov at 2019-02-21T17:56:52+01:00
the aberration fits now correctly consider the most up-to-date magnification matrix estimate
- - - - -
a20aba6a by Takanori Nakane at 2019-02-21T17:37:49+00:00
Merge branch 'jaz_3.1'
- - - - -
d5360bcc by Jasenko Zivanov at 2019-02-22T04:39:15+01:00
bugfix: magnification estimation now also updates the vector of mag. matrices in the obs. model, not only the entries in the metadata table
- - - - -
a41e9683 by Takanori Nakane at 2019-02-22T07:34:59+00:00
Merge branch 'jaz_3.1'
- - - - -
3ffa62dc by Sjors Scheres at 2019-02-25T10:44:05+00:00
trying out prevention of zero CTF in ctf.h to improve code stability on GPU
- - - - -
ddaac75e by Takanori Nakane at 2019-02-25T10:52:24+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
84c428b1 by Sjors Scheres at 2019-02-25T11:28:34+00:00
adapted for obsModel and repaired bug in transformations
- - - - -
a715f64a by Sjors Scheres at 2019-02-25T11:30:41+00:00
repaired bug in apply_transformation in stack_create
- - - - -
06b59a7e by Takanori Nakane at 2019-02-25T11:30:48+00:00
Pipeliner: Forbid {, }, $ for alias. Escape special characters in flowchart TeX files. Fix the return value from flowchart generator.
- - - - -
fdce358f by Sjors Scheres at 2019-02-25T11:52:58+00:00
repaired bug in apply_transformation in stack_create
- - - - -
632d7c83 by Takanori Nakane at 2019-02-25T11:59:38+00:00
relion_stack_create: add warning about --apply_transformation
- - - - -
6d96df8e by Takanori Nakane at 2019-02-25T12:11:14+00:00
Merge branch 'bitbucket-3.0'
- - - - -
151f181a by Takanori Nakane at 2019-02-25T12:19:12+00:00
Merge branch 'ver3.0'
- - - - -
f73ec6ea by Takanori Nakane at 2019-02-25T12:48:53+00:00
Repaied GPU double precision build (GitHub issue #399; thanks to @miesav)
- - - - -
87d87fd1 by Takanori Nakane at 2019-02-25T12:52:31+00:00
GUI: never submit to the queue for Import/SubsetSelection/ManualPick. (Github issue #445; thanks to @andschenk)
- - - - -
6062e4b2 by Takanori Nakane at 2019-02-25T12:55:07+00:00
Merge branch 'bitbucket-3.0'
- - - - -
ccd88db7 by Takanori Nakane at 2019-02-25T15:03:59+00:00
Polish: removed a confusing message (with MPI, it wrote the IDs of micrographs processed by the rank 0 only)
- - - - -
09b22d06 by Takanori Nakane at 2019-02-25T15:09:17+00:00
run_motioncorr: removed a debug message
- - - - -
b819fb57 by Takanori Nakane at 2019-02-25T17:46:19+00:00
Merge branch 'ver3.0'
Conflicts:
src/gui_jobwindow.cpp
- - - - -
032aac13 by Takanori Nakane at 2019-02-25T17:47:29+00:00
pipeliner: fixed continuation
- - - - -
5e8335bc by Sjors Scheres at 2019-02-26T12:32:04+00:00
repaired bug reported by Claudio Alfiero and Takanori Nakane about multibody refinement having displaced bodies after body 0 converged
- - - - -
54145561 by Takanori Nakane at 2019-02-26T12:37:00+00:00
Merge branch 'bitbucket-3.0'
- - - - -
5aeaca45 by Sjors Scheres at 2019-02-26T15:54:13+00:00
Merge branch 'ver3.0'
Conflicts:
src/gui_jobwindow.cpp
src/pipeliner.cpp
- - - - -
1a479629 by Sjors Scheres at 2019-02-26T15:55:55+00:00
added joinSTAR with optics groups for version 3.1
- - - - -
e3056abc by Sjors Scheres at 2019-02-26T15:56:24+00:00
added joinSTAR with optics groups for version 3.1
- - - - -
92fc0295 by Sjors Scheres at 2019-02-26T15:57:17+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
Conflicts:
src/gui_jobwindow.cpp
- - - - -
3168399f by Takanori Nakane at 2019-02-27T07:35:07+00:00
cache: removed the assumption about the number of digits in optics group IDs
- - - - -
28adbfab by Sjors Scheres at 2019-02-27T12:16:42+00:00
added unique to nr AUs for helical refinement
- - - - -
8f37d8e7 by Sjors Scheres at 2019-02-27T12:17:15+00:00
determine dimensionality upon initiliase and repair bug optics group
- - - - -
05017135 by Sjors Scheres at 2019-02-27T12:17:40+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
7ea4dee6 by Sjors Scheres at 2019-02-28T16:09:14+00:00
origin shifts in Angstrom for 3.1 version of do_center
- - - - -
33209f67 by Sjors Scheres at 2019-02-28T16:09:46+00:00
implemented carrying on of unrecognised metadata labels
- - - - -
058edf74 by Sjors Scheres at 2019-02-28T16:10:11+00:00
rescale and window masks during initialiseGeneral
- - - - -
6d48edde by Sjors Scheres at 2019-02-28T16:15:46+00:00
repaired bug in fn_mask new name for second mask
- - - - -
0cb37fc2 by Takanori Nakane at 2019-02-28T21:54:07+00:00
repair checkMask()
- - - - -
ed35a331 by Sjors Scheres at 2019-03-01T13:53:32+00:00
add option to import old particles.star and to rename optics groups
- - - - -
be838e2f by Sjors Scheres at 2019-03-01T13:55:53+00:00
forgot to add ../src/ml_optimiser.h
- - - - -
d1bc3841 by Sjors Scheres at 2019-03-01T13:56:16+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
440ac10a by Sjors Scheres at 2019-03-01T13:56:39+00:00
removed warning
- - - - -
3ddeade2 by Sjors Scheres at 2019-03-01T14:00:34+00:00
repaired bug import
- - - - -
1c1197d0 by Sjors Scheres at 2019-03-01T15:41:41+00:00
small bug in checkMask
- - - - -
42c623b2 by Sjors Scheres at 2019-03-04T13:48:19+00:00
add ignore optics to displayer
- - - - -
07449e41 by Sjors Scheres at 2019-03-04T13:48:42+00:00
split groups per optics_group when combining
- - - - -
75ebf486 by Sjors Scheres at 2019-03-04T13:50:35+00:00
repaired bugs in clauclate powerSpectra
- - - - -
89e23707 by Takanori Nakane at 2019-03-04T14:55:15+00:00
GUI: Added RELION_QSUB_EXTRA(n)_HELP (GitHub issue #446, thanks to @Fravadona)
- - - - -
89af8fa1 by Takanori Nakane at 2019-03-04T14:59:42+00:00
Merge branch 'ver3.0'
- - - - -
17c64c7b by Sjors Scheres at 2019-03-06T12:15:59+00:00
adjustments to deal with 3.1 star files in nr_uniform and others
- - - - -
b6486afe by Sjors Scheres at 2019-03-06T13:35:59+00:00
added option to ignore optics groups to stack_create
- - - - -
8a9c84f3 by Sjors Scheres at 2019-03-06T17:40:10+00:00
removed 3 bugs with anisotropic magnification: 1. projector and backprojector took transpose instead of inverse (in cpu code; gpu code was correct with inverse); 2. in obs_model.cpp applyAnisoMagTransp should take inverse instead of transpose, renamed it to applyAnisoMag; 3. in generateEulerMatrices of acc_helper_functios_impl.h the MAG matrix should be applied before, not after, the inverse operation. Also removed boolean for IS_NOT_INV from all set2DFourierTransform and get2DFourierTransform calls, as IS_INV was never used anyway
- - - - -
7105600d by Takanori Nakane at 2019-03-07T11:45:25+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
31611d6c by Sjors Scheres at 2019-03-07T14:30:47+00:00
removed comment from jasenko about anisotropic mag and inverse
- - - - -
0c71f303 by Takanori Nakane at 2019-03-07T14:42:00+00:00
AutoPicker: make an error message consistent with GUI
- - - - -
7714b88c by Takanori Nakane at 2019-03-07T15:04:21+00:00
cmake: automatically add the Git commit ID to the version string (available as g_RELION_VERSION variable)
- - - - -
ba4b0498 by Sjors Scheres at 2019-03-07T15:42:11+00:00
generateEulerMatrices is only called for Fine squared differences. Coarse squared differences uses cuda_kernel_make_eulers_3D on the GPU. This wasn't doing the anisotropic magnifications. Fixed this. Now test case with only anisotropic magnification (no CTF) goes fine both on CPU and on GPU
- - - - -
4f65ebd7 by Sjors Scheres at 2019-03-07T15:44:34+00:00
take inverse instead of transpose in generate euler matrices in CPU acc code
- - - - -
1346e18f by Sjors Scheres at 2019-03-07T15:44:50+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
54000142 by Sjors Scheres at 2019-03-08T14:54:48+00:00
repaired bug new boxsize in rescale and new_box
- - - - -
7826971b by Sjors Scheres at 2019-03-11T09:42:38+00:00
repaired bug in project ctf
- - - - -
a877bdfe by Sjors Scheres at 2019-03-11T09:47:47+00:00
repaired bug in image_handler
- - - - -
ddfff781 by Sjors Scheres at 2019-03-11T12:45:59+00:00
repaired stupid typo bug in applyAnisoMag
- - - - -
98171da9 by Takanori Nakane at 2019-03-11T15:48:22+00:00
Call this 3.0.1 because we changed (fixed) relion_refine.
- - - - -
b966fbbe by Sjors Scheres at 2019-03-12T16:27:33+00:00
do either aniso mag or CTF refinement
- - - - -
03df4176 by Sjors Scheres at 2019-03-13T10:19:18+00:00
repaired bug in not setting opticsGroup in CTF::setValuesByGroup, affected symmetrical aberration correction when having mroe than 1 optics group
- - - - -
f8735436 by Takanori Nakane at 2019-03-13T12:55:16+00:00
relion_refine_mpi: Fix buffer overflow during initialisation (solves issue #449; thanks to @Fravadona for finding this and suggesting the fix)
Call this 3.0.2 because we touched relion_refine.
- - - - -
555486c7 by Takanori Nakane at 2019-03-13T15:16:10+00:00
Merge branch 'ver3.0' (MPI initialisation fix)
- - - - -
055da427 by Sjors Scheres at 2019-03-14T16:22:33+00:00
first attempt at allowing external system calls to do the reconstruction
- - - - -
9bc6e9d1 by Sjors Scheres at 2019-03-14T16:22:59+00:00
test program for external reconstruct, does exactly the same as normal reconstruct
- - - - -
8a256f2b by Sjors Scheres at 2019-03-14T17:46:28+00:00
report an error if external reconstruction return an error
- - - - -
8b988a71 by Takanori Nakane at 2019-03-15T16:47:55+00:00
RelionIt: updated load_star for 3.1 STAR files (RelionIt itself does not support 3.1 yet)
- - - - -
38dbb3a7 by Takanori Nakane at 2019-03-18T12:25:59+00:00
GUI: Validate the optics group name in the Import job. Also repaired 'particle optics group name'.
- - - - -
3330d563 by Takanori Nakane at 2019-03-19T10:16:31+00:00
MetadataTable: fix writing of an empty DoubleArray
- - - - -
271ed1c7 by Takanori Nakane at 2019-03-19T10:16:51+00:00
Antisymmetric aberration: need at least 5 elements for beam tilts
- - - - -
7e07dc85 by Sjors Scheres at 2019-03-19T14:03:51+00:00
removed spurious function
- - - - -
dbb1c9e8 by Sjors Scheres at 2019-03-19T14:04:17+00:00
improved cerr and cout output
- - - - -
171553e3 by Sjors Scheres at 2019-03-19T16:39:20+00:00
repaired bug in precalculateInvSigma2
- - - - -
92eadedc by Sjors Scheres at 2019-03-19T16:40:39+00:00
setXmippOrigin external reconstruction AND invtau2=0 if both invw and tau2 are zero, which may happen with different optics groups
- - - - -
8c385ead by Takanori Nakane at 2019-03-19T17:40:00+00:00
Aberration and Magnification fit: OpenMP parallelisation
- - - - -
9d725855 by Takanori Nakane at 2019-03-19T17:40:29+00:00
Magnification fit: removed debug message
- - - - -
6172d60d by Takanori Nakane at 2019-03-19T17:40:46+00:00
B/k fitting: clarified that fitting is done using FCCs from all particles
- - - - -
b2e797b8 by Takanori Nakane at 2019-03-20T11:15:18+00:00
relion_refine: repaired '--reuse_scratch' when there is empty optics groups. Made printed messages 1-indexed.
- - - - -
f1079110 by Sjors Scheres at 2019-03-20T11:55:19+00:00
ObservationModel::applyScaleDifference has to be the other way aroundgit add ../src/jaz/obs_model.cpp
- - - - -
8bfab681 by Sjors Scheres at 2019-03-20T12:13:02+00:00
repaired bug in current and coarse sizes of all optics groups
- - - - -
6de0842b by Sjors Scheres at 2019-03-20T12:14:24+00:00
Merged in learning (pull request #24)
Learning
- - - - -
d557a38f by Takanori Nakane at 2019-03-20T12:41:04+00:00
relion_estimate_gain: a simple gain estimation program
- - - - -
39fb1ce2 by Sjors Scheres at 2019-03-21T12:58:24+00:00
repaired bug applyScaleDiff in angular accuracy estimation
- - - - -
65035519 by Sjors Scheres at 2019-03-21T14:20:12+00:00
import takes MTF file and adds to optics table. postprocess takes original pixel size of micrograph for MTF correction
- - - - -
718438e0 by Sjors Scheres at 2019-03-21T14:51:48+00:00
remapped sigma2_noise onto the optics group in calculateAngularError
- - - - -
c9b234b6 by Takanori Nakane at 2019-03-22T12:14:29+00:00
AutoPick MPI: wait for all workers to finish
- - - - -
8a2fc308 by Sjors Scheres at 2019-03-22T12:36:43+00:00
repaired bug in estimateAngularError
- - - - -
d1ca8cfd by Takanori Nakane at 2019-03-22T14:27:47+00:00
Merge branch 'ver3.0' for MPI autopick fix
- - - - -
15a89764 by Sjors Scheres at 2019-03-22T15:11:07+00:00
implement division by MTF during refine/reconstruct
- - - - -
edb78c0b by Takanori Nakane at 2019-03-22T16:26:53+00:00
ml_optimiser: fixed integer overflow at very fine angular sampling. Call this 3.0.3.
- - - - -
a1157858 by Sjors Scheres at 2019-03-26T09:04:17+00:00
have option to not die upon loadSafely of obsModel, also some tweeks for divideMtf
- - - - -
1898d109 by Sjors Scheres at 2019-03-26T17:21:19+00:00
dont die if cannot convert for 3.1
- - - - -
104376f8 by Sjors Scheres at 2019-03-26T17:22:12+00:00
added writing EPS and logfile.pdf for ctf_refine with the diagnostic images
- - - - -
284f3574 by Sjors Scheres at 2019-03-27T11:37:50+00:00
get real group name
- - - - -
8b428396 by Sjors Scheres at 2019-03-27T11:38:04+00:00
now allow different pixel sizes
- - - - -
71432e0a by Sjors Scheres at 2019-03-27T11:52:31+00:00
added functionality to simulate a CTF image
- - - - -
8e83af3d by Sjors Scheres at 2019-03-27T11:53:01+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
555405e3 by Sjors Scheres at 2019-03-27T15:24:27+00:00
added options to skip gridding and MAP in external reconstruct
- - - - -
a3f9ab6c by Sjors Scheres at 2019-03-27T15:24:53+00:00
added option to ignore optics groups
- - - - -
9cf18c17 by Sjors Scheres at 2019-03-27T15:25:23+00:00
reorganised reconstruct and new implementation skip_gridding
- - - - -
35a7a7e7 by Sjors Scheres at 2019-03-28T11:50:07+00:00
added option to downscale simulated CTF from larger box to simulate too-small box effect of aliasing
- - - - -
448773d7 by Sjors Scheres at 2019-03-28T11:50:50+00:00
made acc kernels for premultiplied CTFs
- - - - -
895ed5fd by Sjors Scheres at 2019-03-28T12:04:47+00:00
option to ignore optics groups
- - - - -
c47811d7 by Sjors Scheres at 2019-03-28T12:29:37+00:00
repaired bug in dLL calculation
- - - - -
11cf2455 by Takanori Nakane at 2019-03-28T15:11:09+00:00
ALTCPU BackProjection kernels: Fixed a template instantiation problem with CTF_PREMULTIPLIED.
- - - - -
717e0cb7 by Sjors Scheres at 2019-03-28T17:32:52+00:00
dont provide --ctf_premultiplied in refine/reconstruct: get from opticsTable instead
- - - - -
bee2671e by Sjors Scheres at 2019-03-28T17:48:04+00:00
simplified ctf_toolbox. Also, it now sets ctfPremultiplied in the obsModel
- - - - -
4ef84a79 by Sjors Scheres at 2019-03-28T17:48:27+00:00
made an option to use ctf-downscaling in getFftwImage
- - - - -
30136dcd by Sjors Scheres at 2019-03-28T17:48:53+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
ca213057 by Sjors Scheres at 2019-03-29T10:19:18+00:00
added option to use ctf padding to relion_refine
- - - - -
47a8c393 by Sjors Scheres at 2019-03-29T10:19:55+00:00
removed phase flipping options from GUI, and don't die if an undefined option is encountered upon reading run.jobstar file
- - - - -
0a37a0ab by Sjors Scheres at 2019-03-29T10:20:37+00:00
added option for ctf padding to ctf_refine program. Also check and die/warn for ctfPremultiplied data
- - - - -
6201e419 by Takanori Nakane at 2019-03-29T14:21:48+00:00
Polish: added --suffix to allow re-use of trajectories
- - - - -
63eb56ca by Takanori Nakane at 2019-03-29T15:52:39+00:00
GUI: made "Do dose-weighting?: Yes" to the default
- - - - -
04bfe039 by Sjors Scheres at 2019-03-29T16:22:13+00:00
implemented ctf_padding in ctf_refine
- - - - -
068a19d9 by Sjors Scheres at 2019-04-01T09:49:31+01:00
repaired bug in new implementation skip_gridding
- - - - -
249fa89b by Takanori Nakane at 2019-04-01T17:03:14+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
b963b4f5 by Takanori Nakane at 2019-04-01T19:29:52+01:00
ObsModel: added setBoxSize to invalidate caches.
- - - - -
8af03234 by Takanori Nakane at 2019-04-01T19:40:13+01:00
BayesianPolish: added CTF premultiplication and cropping after that. The order of operation is --window, --scale, CTF, --crop. Confirmed on apoferritin but NEEDS MORE ERROR CHECKS!
- - - - -
dc18ce2e by Sjors Scheres at 2019-04-02T13:29:21+01:00
dont reconstruct when doing CTFs in reconstructor, just take data/weight
- - - - -
ef684dbe by Sjors Scheres at 2019-04-02T13:29:49+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
5206665c by Takanori Nakane at 2019-04-03T11:28:18+01:00
relion_flex_analyse: preserve the original image names in the rlnImageOriginalName column (reported by Sean Connell on CCPEM)
- - - - -
1a1d5a53 by Charles Congdon at 2019-04-03T11:37:52+01:00
Fix build when Intel Compiler is used
- - - - -
4c10a974 by Takanori Nakane at 2019-04-03T11:38:39+01:00
Merge branch 'ver3.0' of https://bitbucket.org/scheres/relion-devel-lmb into ver3.0
- - - - -
a7b63f34 by Takanori Nakane at 2019-04-03T11:40:47+01:00
Merge branch 'ver3.0' of https://bitbucket.org/scheres/relion-devel-lmb into bitbucket-3.0
- - - - -
e9afc9a1 by Takanori Nakane at 2019-04-03T11:47:22+01:00
Multibody: fix crash during initialisation (reported by Rafael Fernández Leiro). Call this 3.0.4.
- - - - -
a71285c5 by Sjors Scheres at 2019-04-08T08:59:29+01:00
added openMP parallelisation and mask_diameter to helix_inimodel2d
- - - - -
7f46b9f9 by Sjors Scheres at 2019-04-08T09:00:41+01:00
added --auto_iter_max to stop auto-refinement at given nr_iter for Gustav and Jonas
- - - - -
d4abe349 by Charles Congdon at 2019-04-08T11:29:39+01:00
Fix build when Intel Compiler is used
- - - - -
4d5c00fb by Takanori Nakane at 2019-04-08T11:38:05+01:00
Extract: set rlnCtfDataAreCtfPremultiplied when --premultiply_ctf
- - - - -
7a3c7d19 by Takanori Nakane at 2019-04-10T15:29:04+01:00
AUTHORS: Included CPlot2D (reported in GitHub issues #460; thanks for pointing this out)
- - - - -
a5f6579c by Takanori Nakane at 2019-04-10T15:58:09+01:00
Merge branch 'ver3.0'
Conflicts:
AUTHORS
src/flex_analyser.cpp
src/macros.h
src/ml_optimiser.cpp
- - - - -
b8964268 by Takanori Nakane at 2019-04-11T11:36:38+01:00
RelionCor: fixed the path to non-dose weighted sums in a log file
- - - - -
c8f9fe1c by Takanori Nakane at 2019-04-11T11:40:35+01:00
Merge branch 'ver3.0'
- - - - -
6a68032a by Sjors Scheres at 2019-04-11T20:09:02+01:00
repaired bug in writing star file
- - - - -
1be7c133 by Sjors Scheres at 2019-04-11T20:09:46+01:00
write out optimal y-off and rotations, so output STAR files can be used as input, which allows ever finer local angular and translational searches
- - - - -
2c7cd922 by Sjors Scheres at 2019-04-11T20:10:19+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
28cb2db7 by Takanori Nakane at 2019-04-15T12:23:27+01:00
Improved an error message when the directory is not writable.
- - - - -
564237db by Takanori Nakane at 2019-04-16T11:51:00+01:00
Merge branch 'ver3.0'
- - - - -
381f2936 by Takanori Nakane at 2019-04-16T12:02:12+01:00
AutoPick: Number of particles should be an integer.
- - - - -
eef01a82 by Takanori Nakane at 2019-04-23T14:33:47+01:00
CTF: implemented ignore_ctf_after_peak
- - - - -
67c1ed32 by Sjors Scheres at 2019-04-24T14:32:04+01:00
check for not too large value of ires in noise-generation outside particle mask, which went wrong with 3D data with Kino
- - - - -
bd5b1443 by Sjors Scheres at 2019-04-24T14:32:45+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
0a1471b5 by Takanori Nakane at 2019-04-26T11:29:02+01:00
Non-void functions should always explicitly return something (GitHub issue #453 reported by @riverseow and @kmlyn).
- - - - -
5d8fee7c by Takanori Nakane at 2019-04-26T11:37:28+01:00
TIFF: Report the file name upon error. Unfortunately this does not catch all errors. TIFFSetErrorHandler() cannot get the file name without using a global variable, which is not useful for multi-threaded programs ...
- - - - -
9caf2da0 by Takanori Nakane at 2019-04-26T11:54:58+01:00
Merge branch 'ver3.0'
Conflicts:
src/displayer.cpp
src/displayer.h
src/jaz/motion/motion_param_estimator.cpp
src/metadata_table.cpp
- - - - -
1887f6dc by Jasenko Zivanov at 2019-04-29T02:28:47+02:00
modular CTF refinement added; not yet functional
- - - - -
73bdb8ee by Takanori Nakane at 2019-04-30T14:41:41+01:00
AutoPicker: removed debug output (reported by David Hoover on CCPEM)
- - - - -
4346e550 by Jasenko Zivanov at 2019-05-02T01:14:55+02:00
new modular ctf-refinement now functional
- - - - -
290308b7 by Takanori Nakane at 2019-05-03T11:19:51+01:00
Scratch: do not copy same particles (useful after symmetry expansion; suggested by Oli on Github #462)
- - - - -
bbd1e25c by Takanori Nakane at 2019-05-03T11:43:39+01:00
relion_stack_create: added --one_by_one option to save memory
- - - - -
6835b92f by Sjors Scheres at 2019-05-03T11:46:50+01:00
repaired bug in backrotate3D
- - - - -
1fc49bf2 by Takanori Nakane at 2019-05-03T11:48:59+01:00
Merge branch 'jaz_3.1' for modular CTF optimisation.
- - - - -
64c60350 by Sjors Scheres at 2019-05-03T14:54:04+01:00
repair bug backrotate3d int to rfloat
- - - - -
a53aa567 by Sjors Scheres at 2019-05-03T14:54:26+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
15a9c813 by Sjors Scheres at 2019-05-03T15:52:05+01:00
add cl_angpix to star handler
- - - - -
cf628050 by Sjors Scheres at 2019-05-03T15:55:16+01:00
implementation Kent Thurbers rotational prior for helices
- - - - -
6d6901cb by Takanori Nakane at 2019-05-04T11:45:09+01:00
Polish: fixed a bug introduced with --crop
- - - - -
b9fee094 by Takanori Nakane at 2019-05-07T11:26:58+01:00
Scratch: repaired crash when 'parallel disc I/O' is no and running on multiple nodes. Call this 3.0.5.
- - - - -
2084dd9e by Sjors Scheres at 2019-05-07T11:51:45+01:00
adjusted GUI to work with Jasenko's new ctf_mode
- - - - -
c3b33646 by Takanori Nakane at 2019-05-07T11:53:25+01:00
This is 3.0.5.
- - - - -
f1e1dc23 by Takanori Nakane at 2019-05-07T12:11:04+01:00
Merge 3.0.5 for bug fixes
Conflicts:
src/macros.h
src/ml_optimiser.cpp
- - - - -
b17ffa13 by Takanori Nakane at 2019-05-07T14:24:10+01:00
ml_optimiser: trying to implement --strict_lowres_exp (but not working yet ...)
- - - - -
add22da0 by Takanori Nakane at 2019-05-07T14:56:11+01:00
estimate_gain: added a new line to the message
- - - - -
2c4a49fe by Sjors Scheres at 2019-05-07T16:56:13+01:00
strict_lowres_exp only in projector, not in images. Also store in optimiser.star for restarts
- - - - -
08d6c487 by Sjors Scheres at 2019-05-08T11:57:55+01:00
repaired small bug in subset selection with split_size AND nr_split
- - - - -
abd84a5a by Takanori Nakane at 2019-05-08T14:05:16+01:00
ctf_toolbox: --ctf_intact_after_first_peak
- - - - -
31feb245 by Takanori Nakane at 2019-05-08T14:10:37+01:00
AutoPicker: --LoG_upper_threshold
- - - - -
f5f7b309 by Sjors Scheres at 2019-05-09T08:02:43+00:00
Merged in E_step_max_res (pull request #25)
E step max res
- - - - -
917c9804 by Sjors Scheres at 2019-05-09T15:20:11+01:00
remove run.out and run.err when overwriting
- - - - -
e574c19c by Takanori Nakane at 2019-05-10T11:25:44+01:00
ObsModel: Make sure rlnMicrographOriginalPixelSize is present when applying MTF correction
- - - - -
ae9b8e47 by Takanori Nakane at 2019-05-10T12:54:07+01:00
motioncorr_runner: keep EMDL_MICROGRAPH_ORIGINAL_PIXEL_SIZE for MTF correction
- - - - -
4d15fde6 by Takanori Nakane at 2019-05-13T14:46:45+01:00
Improve stability of histogram generation.
- - - - -
a9dff2e3 by Takanori Nakane at 2019-05-13T14:55:22+01:00
Merge branch 'ver3.0' for a histogram bug fix.
Conflicts:
src/autopicker.cpp
- - - - -
81e269df by Takanori Nakane at 2019-05-14T10:33:36+01:00
relion_star_handler --select: show the number of selected items (GitHub issue #466 suggested by @daniel-s-d-larsson)
- - - - -
e781ac67 by Takanori Nakane at 2019-05-14T10:34:15+01:00
scratch: avoid use of wildcards to prevent too long arguments (GitHub issue #467 reported by @EMBL-Structures)
- - - - -
9f23f178 by Takanori Nakane at 2019-05-14T11:49:15+01:00
scratch: avoid use of wildcards to prevent too long arguments (GitHub issue #467 reported by @EMBL-Structures)
- - - - -
3057752e by Sjors Scheres at 2019-05-15T10:04:23+01:00
helical offsets restricted along x in Class2D, rot_sigma on GUI, and Fourier-mask to remove projections along helical axis from Projector
- - - - -
4c0b9363 by Takanori Nakane at 2019-05-16T13:52:00+01:00
RelionCor: use a defect file (same format as UCSF MotionCor2), write local trajectories and hot pixels to metadata STAR files.
- - - - -
e9082074 by Takanori Nakane at 2019-05-16T15:01:49+01:00
MotioncorrRunner::fillDefectMask as a static function (for use in Polishing)
- - - - -
d63796f3 by Takanori Nakane at 2019-05-16T15:02:24+01:00
PostProcess: remove --auto_mask option (one should use 'relion_mask_create' beforehand)
- - - - -
8d193acb by Sjors Scheres at 2019-05-16T15:40:48+01:00
hopefully repaired bug strict_highres_exp in GPU code, reported by Kino
- - - - -
edf4e2e7 by Sjors Scheres at 2019-05-17T09:36:18+01:00
modified apps with relion-specific exit signal to catch error or normal exit
- - - - -
002151f8 by Sjors Scheres at 2019-05-17T09:37:08+01:00
started with a pipeline_control system (thru disk) that allows aborting jobs and catching their errors
- - - - -
da1dccbf by Sjors Scheres at 2019-05-17T09:54:32+01:00
resolving conflicts in setFouriertransforms in projector when merging rotprior into master
- - - - -
fdc69a1f by Sjors Scheres at 2019-05-17T08:58:10+00:00
Merged in rotprior (pull request #26)
Rotprior is a modification suggested by Kent Thurber from the Tycko lab at NIH to allow priors on the rot angle in helical refinement. The branch has also accumulated some other changes, e.g. a fourier mask for helical refinement; a bug in backrotate-3d, the removal of run.out and run.err in overwriting jobs.
- - - - -
ce8b7e7f by Sjors Scheres at 2019-05-17T08:59:40+00:00
Merged in master (pull request #27)
Master
- - - - -
7526d602 by Sjors Scheres at 2019-05-17T11:28:25+01:00
mpi_abort with relion_exit_failure
- - - - -
86ce765c by Sjors Scheres at 2019-05-17T11:29:20+01:00
implemented check for presence of RELION_JOB_ABORT_NOW file in output directory for pipeline_control
- - - - -
0cb79060 by Sjors Scheres at 2019-05-17T11:30:36+01:00
no longer move output nodes to .old copies, as pipeline_control now takes take of checking when a job has finished. Also multiple tweaks to GUI and pipeliner to make this work.
- - - - -
b76584fb by Takanori Nakane at 2019-05-20T10:46:09+01:00
RelionCor: supported 100 kV
- - - - -
8d34c632 by Takanori Nakane at 2019-05-20T10:48:30+01:00
Merge branch 'ver3.0' for 100 kV support
- - - - -
f0259c52 by Takanori Nakane at 2019-05-21T14:38:10+01:00
LoG picker: fixed inter-particle distance cutoff
- - - - -
34dc552d by Takanori Nakane at 2019-05-21T15:20:41+01:00
Polishing: removed an old code path that worked without micrograph metadata STAR files to reduce code complexity.
- - - - -
bdfa183a by Takanori Nakane at 2019-05-21T17:12:09+01:00
Polish: use a mask of defects and hot pixels from RelionCor. WIP: need testing!!!
- - - - -
1a38da59 by Joaquin Oton at 2019-05-22T14:01:37+01:00
Removing DEBUG messages
- - - - -
0f9bbaf5 by Takanori Nakane at 2019-05-23T13:54:30+01:00
relion_star_handler --split: removed an assumption on the number of digits
- - - - -
5d879c25 by Sjors Scheres at 2019-05-26T13:51:46+01:00
added option to average helical asymmetric units in 2D in helix_toolbox
- - - - -
2852439c by Sjors Scheres at 2019-05-26T20:30:45+01:00
first version of the new scheduler
- - - - -
b0989420 by Sjors Scheres at 2019-05-26T20:57:16+01:00
fixed some minor nuisances with sending emails
- - - - -
a894be8b by Sjors Scheres at 2019-05-27T14:46:32+01:00
now scheduler works with spawning new jobs and with setting joboptions with scheduler variables
- - - - -
8520d9c1 by Sjors Scheres at 2019-05-27T15:58:58+01:00
added jobtype to launch external jobs from the relion GUI
- - - - -
1056a884 by Sjors Scheres at 2019-05-28T20:31:18+01:00
now with very limited version of scheduler working: can add jobs to a Scheduler, but nothing else
- - - - -
6aee08b1 by Takanori Nakane at 2019-05-29T11:56:26+01:00
LoG picker: changed the default inter-particle cutoff to 100 % (pick 'touching' neighbours).
- - - - -
57a516c4 by Takanori Nakane at 2019-05-29T12:23:08+01:00
ml_model: remove debug output for --strict_lowres_exp
- - - - -
b0b621a1 by Brian Bockelman at 2019-05-29T14:18:54+01:00
Register build byproducts for FFTW external.
If an external project is used for FFTW / FFTWF, Relion links libraries
directly against the resulting libraries (such as libfftw3f.so).
Some build generators (specifically, Ninja) require that the full dependency
graph is calculated at the beginning of the build process. Since the
file `libfftw3f.so` does not exist at the beginning of the build (the
`ExternalProject` command has not been run yet) -- and there is no
rule to generate that file -- the build fails.
With this change, we introduce a fake target that relies on the
`ExternalProject` and has as a build byproduct the resulting library.
This allows `ninja` to correctly resolve the dependency and build Relion.
Signed-off-by: Takanori Nakane <nakane.t at gmail.com>
- - - - -
387089f5 by Sjors Scheres at 2019-05-29T14:41:42+01:00
write out RELION_EXIT_SUCCESS/FAILURE
- - - - -
e65dbe58 by Takanori Nakane at 2019-05-29T15:15:01+01:00
Bayesian Polish: Improve continuation. Also print out the new pixel size.
- - - - -
5b30b93e by Takanori Nakane at 2019-05-29T15:49:40+01:00
Bayesian Polishing: removed debug code
- - - - -
f4ab755c by Takanori Nakane at 2019-05-29T15:53:57+01:00
Polish: use a mask of defects and hot pixels from RelionCor. WIP: need testing!!!
- - - - -
df0701f8 by Sjors Scheres at 2019-05-29T21:34:19+01:00
major change in scheduler by removing all pointers and replace by std::map, some nasty extern variables needed now...
- - - - -
d6f1faf9 by Sjors Scheres at 2019-05-30T09:52:59+01:00
keep scope of global variables within scheduler.cpp and remove concept of SchedulerNode, now Edge just go from Operator/Jobs
- - - - -
c145c32c by Takanori Nakane at 2019-05-30T13:57:46+01:00
Polish: Patch defects frame-wise. WIP: need testing
- - - - -
6551b322 by Takanori Nakane at 2019-05-30T14:00:11+01:00
Merge branch 'defect-mask-in-Polish' of https://bitbucket.org/scheres/relion-devel-lmb into defect-mask-in-Polish
- - - - -
cf6a3d59 by Sjors Scheres at 2019-05-30T17:12:10+01:00
simplified operators and added operators for reading variables from star files
- - - - -
9273ebf6 by Sjors Scheres at 2019-05-30T18:02:37+01:00
now can add/delete operators thru the GUI
- - - - -
db678b23 by Sjors Scheres at 2019-06-01T17:20:24+01:00
scheduler gui almost finished (delete jobs pending), now testing abort
- - - - -
f81e6b56 by Takanori Nakane at 2019-06-03T12:51:24+01:00
AutoPicker: --LoG_upper_threshold
- - - - -
d9bfb69d by Takanori Nakane at 2019-06-03T12:51:24+01:00
LoG picker: fixed inter-particle distance cutoff
- - - - -
418a9eb1 by Takanori Nakane at 2019-06-03T12:51:24+01:00
LoG picker: changed the default inter-particle cutoff to 100 % (pick 'touching' neighbours).
- - - - -
aa6f9d38 by Takanori Nakane at 2019-06-03T13:01:20+01:00
--LoG_use_ctf is not available in 3.0
- - - - -
73f767f9 by Takanori Nakane at 2019-06-03T13:03:26+01:00
Merge branch 'ver3.0'
- - - - -
198e6641 by Takanori Nakane at 2019-06-03T15:31:02+01:00
ObsModel: don't try to convert empty STAR file.
- - - - -
0898273e by Takanori Nakane at 2019-06-03T17:02:13+01:00
CtfRefine: don't print defocus fitting modes unless actually refining them
- - - - -
7668335f by Sjors Scheres at 2019-06-03T17:48:36+01:00
now with previous node button, delete jobs and complete deactivation of scheduler_run_grp
- - - - -
152e4206 by Takanori Nakane at 2019-06-04T12:24:27+01:00
AutoPicker: reject 3D references for filaments
- - - - -
298e7df5 by Sjors Scheres at 2019-06-04T16:24:52+01:00
gui now sets jobs, operators etc when pressing next/prev/set etc
- - - - -
9019d5f7 by Sjors Scheres at 2019-06-04T16:55:52+01:00
start with executing the current_node, then also no longer need to gotopreviousnode upon abort/error
- - - - -
10fb5b54 by Sjors Scheres at 2019-06-04T17:32:32+01:00
repaired bug initialise pipeline read GUI
- - - - -
c01bdb77 by Takanori Nakane at 2019-06-05T15:07:17+01:00
MotionCorr: Remove --movie. Fix crashes when the movie path contains "." or "_movie".
- - - - -
42017f88 by Takanori Nakane at 2019-06-05T15:54:26+01:00
In-memory TIFF decompression to reduce random disk I/O. Unfortunately, this turned out to be slower on distributed file system. So decided not to use by default.
- - - - -
6c5f32f1 by Sjors Scheres at 2019-06-05T17:11:55+01:00
now can continue with has_not_started being inherited
- - - - -
4f920900 by Sjors Scheres at 2019-06-05T17:30:34+01:00
minor tweaks to scheduler
- - - - -
21b956ed by Sjors Scheres at 2019-06-06T11:47:55+01:00
minor tweaks
- - - - -
ecdfc8a3 by Takanori Nakane at 2019-06-06T12:38:32+01:00
PostProcess: fix a bug where postprocess_masked.mrc is overwritten with --half_maps.
- - - - -
01ef97b2 by Takanori Nakane at 2019-06-06T12:40:52+01:00
Merge branch 'ver3.0' to fix the --half_maps bug in postprocess
- - - - -
67f40229 by Sjors Scheres at 2019-06-06T12:57:31+01:00
repaired bug reported by Kino in strict_highres_exp
- - - - -
9e0c8a47 by Sjors Scheres at 2019-06-06T12:50:31+00:00
Merged in pipeliner (pull request #28)
Pipeliner
- - - - -
8a677f29 by Takanori Nakane at 2019-06-07T10:51:18+01:00
This is 3.0.6.
- - - - -
5f72ed91 by Takanori Nakane at 2019-06-07T20:41:35+01:00
RelionCor: don't die but warn when the number of frames is too small.
- - - - -
cc692836 by Takanori Nakane at 2019-06-10T12:31:28+01:00
image.h: silence compiler warnings
- - - - -
e507479c by Takanori Nakane at 2019-06-10T12:31:28+01:00
MlModel: fix crash with empty groups
- - - - -
0bdf72c8 by Takanori Nakane at 2019-06-10T12:44:32+01:00
Scheduler: non-void functions must ALWAYS explictly return something
- - - - -
fd1d4636 by Takanori Nakane at 2019-06-11T14:39:08+01:00
BackProjector: clean up
- - - - -
30612978 by Sjors Scheres at 2019-06-18T10:05:01+01:00
now have option to copy a Schedule
- - - - -
00ee584b by Takanori Nakane at 2019-06-18T17:10:28+01:00
relion_MPI_Bcast: fixed a bug where the MPI communicator (comm) was ignored when calculating ranks. This fixes the GitHub issue #473 and problems reported in CCPEM by Wim Hagen, Dieter Blaas and others.
- - - - -
cc1cde71 by Takanori Nakane at 2019-06-18T17:13:01+01:00
relion_star_handler --split: The GUI cannot reliably detect a finished job. A fundamental fix requires 3.1, so put a friendly message instead.
- - - - -
213f0322 by Takanori Nakane at 2019-06-18T17:20:51+01:00
Merge branch 'ver3.0'
Conflicts:
src/macros.h
- - - - -
570f1fbf by Takanori Nakane at 2019-06-19T14:22:00+01:00
GUI: make it clear re-centering coordinate is in pixels and the origin is the center of the box.
- - - - -
299bb360 by Takanori Nakane at 2019-06-20T15:27:26+01:00
RelionCor: fixed crash in power-spectrum generation when nx > ny.
- - - - -
ed7bd233 by Sjors Scheres at 2019-06-20T15:46:30+01:00
now output nodes can also be read from a star file
- - - - -
aafd53c0 by Sjors Scheres at 2019-06-21T16:39:03+01:00
mpi_barrier before relion_exit_success
- - - - -
f2054121 by Sjors Scheres at 2019-06-21T16:39:26+01:00
removed spurious cerr
- - - - -
3aad2670 by Sjors Scheres at 2019-06-21T16:40:01+01:00
debugging scheduler...
- - - - -
ae6ce1a0 by Sjors Scheres at 2019-06-21T16:40:20+01:00
button to unlock scheduler
- - - - -
483efe41 by Sjors Scheres at 2019-06-21T16:46:35+01:00
resolving conflicts
- - - - -
42f95076 by Takanori Nakane at 2019-06-24T10:56:30+01:00
CtfRefine: don't add unnecessary columns (e.g. Cs)
- - - - -
5786a012 by Takanori Nakane at 2019-06-24T11:08:51+01:00
Subtract: don't make 'run' folder (reported by Madhan)
- - - - -
f16a92a4 by Takanori Nakane at 2019-06-24T12:16:09+01:00
GUI: add "Use power spectra from MotionCorr job" option
- - - - -
e09d0b2a by Takanori Nakane at 2019-06-24T12:53:18+01:00
GUI: add "Save sum of power spectra" and "Sum power spectra every e/A2"
- - - - -
0cfb4778 by Takanori Nakane at 2019-06-24T16:15:09+01:00
GUI: remove ^J (newline) from 'Check command'
- - - - -
8bdd13d5 by Takanori Nakane at 2019-06-24T16:50:59+01:00
GUI: color failed or aborted jobs in the job list
- - - - -
d61b1b8c by Takanori Nakane at 2019-06-24T17:49:30+01:00
GUI: repair 'overwrite' when submitting to a queue
- - - - -
ed942fce by Takanori Nakane at 2019-06-25T11:50:58+01:00
repair ObservationModel::setBoxSize
- - - - -
da486d63 by Takanori Nakane at 2019-06-25T12:51:33+01:00
JoinStar: don't duplicate optics group tables
- - - - -
2e401de1 by Takanori Nakane at 2019-06-25T16:37:18+01:00
removed deprecated programs (star_combine, star_compare, cs_fit)
- - - - -
1c393152 by Takanori Nakane at 2019-06-25T16:54:24+01:00
relion_convert_star: make command line arguments RELION style. WIP: this application should support conversion of non-particle STAR files.
- - - - -
7c478b01 by Takanori Nakane at 2019-06-26T14:13:16+01:00
Repaired premature call to MPI_Finalize() in relion_particle_subtract_mpi and relion_postprocess_mpi.
- - - - -
f780dc4d by Takanori Nakane at 2019-06-27T13:43:37+01:00
Merged CMakeFile patches from Colin Palmer.
- Ensure relion_lib is linked to libtiff and libomp, otherwise Mac (clang?) build fails
- Optionally allow shared-library build
- - - - -
9b7ca04e by Sjors Scheres at 2019-06-27T14:46:51+01:00
repaired bad conflict resolve
- - - - -
106bd500 by Sjors Scheres at 2019-06-27T14:50:14+01:00
n_digits in split
- - - - -
7d050ca4 by Sjors Scheres at 2019-06-27T17:53:50+01:00
continue to debug scheduler
- - - - -
2749b327 by Sjors Scheres at 2019-06-28T15:00:32+01:00
now gui can toggle between schedules and pipeline
- - - - -
bc5eebb7 by Sjors Scheres at 2019-06-28T15:21:09+01:00
minor tweaks
- - - - -
472d236f by Sjors Scheres at 2019-06-28T15:42:42+01:00
now mark as failed on gui, and more debugging scheduler
- - - - -
b4130730 by Sjors Scheres at 2019-06-28T16:31:42+01:00
further smoothened abort and GUI deactivation
- - - - -
f0cc2317 by Sjors Scheres at 2019-06-28T16:36:57+01:00
stopping message to out, not err
- - - - -
ca4acef9 by Takanori Nakane at 2019-06-28T17:15:01+01:00
estimate_gain: add random sampling of movies; add normalisation
- - - - -
c58438c1 by Takanori Nakane at 2019-06-29T08:42:28+01:00
bfactor.py, relion_it.py: repaired split file names (bug reported by Shaun Rawson)
- - - - -
096f9c45 by Takanori Nakane at 2019-07-02T10:57:24+01:00
Call this 3.0.7
- - - - -
5f12461a by Takanori Nakane at 2019-07-02T17:02:15+01:00
split: always use at least three digits; this is a very dirty workaround but necessary for RelionIt...
- - - - -
a1ffdc0b by Sjors Scheres at 2019-07-03T09:14:53+01:00
keep polishing the schedules
- - - - -
1a320743 by Sjors Scheres at 2019-07-03T09:58:43+01:00
separate show_scheduler on maingui and create_cheduler_gui in gui_netries
- - - - -
7ca3f33b by Sjors Scheres at 2019-07-03T11:51:19+01:00
added option to create 3d model from 2d images straight away
- - - - -
a28ff5fb by Sjors Scheres at 2019-07-03T11:51:42+01:00
small change in format of run.out
- - - - -
44ae9b4e by Sjors Scheres at 2019-07-03T11:56:36+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
20ae8a08 by Sjors Scheres at 2019-07-03T12:48:38+01:00
change only_extract_unfinished to only_do_unfinished
- - - - -
5b563c83 by Sjors Scheres at 2019-07-03T12:48:56+01:00
repaired bug in finding logfiles when using given_PS
- - - - -
a4afa8db by Sjors Scheres at 2019-07-03T12:49:22+01:00
better handling of choice menus in scheduler mode
- - - - -
ec12d791 by Takanori Nakane at 2019-07-03T14:20:53+01:00
ManualPicker: display the end part of file names
- - - - -
c455e8bf by Sjors Scheres at 2019-07-03T14:36:44+01:00
no make new schedule on GUI menu
- - - - -
cb40264e by Sjors Scheres at 2019-07-03T14:37:05+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
4f652902 by Sjors Scheres at 2019-07-04T09:07:26+01:00
small things scheduler
- - - - -
732396e4 by Sjors Scheres at 2019-07-04T09:08:17+01:00
make inimodel2d part of 3.1 release
- - - - -
888e710c by Sjors Scheres at 2019-07-04T11:31:07+01:00
added compilation flag USE_WITHOUT_STRUCTURA_LICENSE
- - - - -
96faf44f by Sjors Scheres at 2019-07-04T11:31:30+01:00
Merge branch 'ver3.0' of https://bitbucket.org/scheres/relion-devel-lmb into ver3.0
- - - - -
1a4fce67 by Sjors Scheres at 2019-07-04T11:34:27+01:00
removed stochastic em option from GUI
- - - - -
815c9010 by Sjors Scheres at 2019-07-04T11:37:39+01:00
resolve conflict merging in 3.0
- - - - -
3e7953d3 by Sjors Scheres at 2019-07-04T15:37:05+01:00
updated messages for CTF-modulation in SGD after discussion with Francois
- - - - -
ccfe5d33 by Sjors Scheres at 2019-07-04T15:41:48+01:00
after merge new of ver3.0
- - - - -
f9212f6e by Sjors Scheres at 2019-07-05T09:04:24+01:00
added external reconstruct option to reconstructor, useful for debugging
- - - - -
f47827ac by Sjors Scheres at 2019-07-05T13:44:58+01:00
repaired bug translations star file helix_inimodel2d
- - - - -
84b06a05 by Sjors Scheres at 2019-07-05T15:36:31+01:00
repaired bug in filename external reocnstruct
- - - - -
3aae2324 by Sjors Scheres at 2019-07-09T15:56:43+01:00
smaller letteres variables, edges etc on scheduler gui to prevent problems on osx
- - - - -
2990ebeb by Sjors Scheres at 2019-07-10T09:48:20+01:00
repaired bug in not initialising do_ctf_padding
- - - - -
4dd6ebed by Takanori Nakane at 2019-07-10T11:17:22+01:00
GUI: don't crash when an output file is not present.
- - - - -
cea2f7d2 by Takanori Nakane at 2019-07-10T11:17:56+01:00
Subtract job: fixed the output file name
- - - - -
5ffd460f by Sjors Scheres at 2019-07-10T15:54:44+01:00
keep checking run.out in scheduler view GUI
- - - - -
42515bb5 by Sjors Scheres at 2019-07-10T15:57:29+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
db08e96c by Sjors Scheres at 2019-07-10T16:13:40+01:00
improved RELION_EXIT__SUCCESS behavior with runGUI
- - - - -
0b273d3c by Sjors Scheres at 2019-07-11T12:21:04+01:00
removed spurious hello statement
- - - - -
86f15141 by Takanori Nakane at 2019-07-12T11:54:30+01:00
GUI: don't set --angpix to CtfFind jobs
- - - - -
7b2f3edc by Takanori Nakane at 2019-07-15T10:58:47+01:00
GUI: improved check for 'save power spectra'
- - - - -
8fd7b74a by Takanori Nakane at 2019-07-16T12:34:16+01:00
MetadataTable: Removed unused 'ignoreLabels'. Check the number of columns in a loop.
- - - - -
23c5a542 by Takanori Nakane at 2019-07-16T12:41:14+01:00
GCTF wrapper: repaired --use_noDW
- - - - -
ea9425e7 by Takanori Nakane at 2019-07-17T15:26:17+01:00
GUI: don't add --fast_subset to Class3D when continuing
- - - - -
bc2ec1d1 by Takanori Nakane at 2019-07-17T15:59:59+01:00
MetaDataTable: escape empty string as "".
- - - - -
01848bec by Sjors Scheres at 2019-07-19T15:19:01+01:00
fixed bug reported by jan in relion_display when used for micrographs instead of particles
- - - - -
78291e5d by Sjors Scheres at 2019-07-19T15:19:20+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
7955397e by Sjors Scheres at 2019-07-19T15:19:52+01:00
now allow 2D with helices in reading model.star
- - - - -
7b3452f1 by Sjors Scheres at 2019-07-19T15:20:21+01:00
kobsmodel loadsafely for reposition
- - - - -
bcc8ab38 by Sjors Scheres at 2019-07-19T15:21:02+01:00
small bnug
- - - - -
65b0acd7 by Sjors Scheres at 2019-07-19T15:21:22+01:00
return int for system call
- - - - -
455b05cf by Sjors Scheres at 2019-07-22T09:38:22+01:00
repaired bug helices 2d
- - - - -
7fcb17f7 by Sjors Scheres at 2019-07-24T14:09:54+01:00
make RELION_EXIT file for manulapicker
- - - - -
78ada05e by Takanori Nakane at 2019-07-24T14:43:58+01:00
CtfRefine: just added a comment
- - - - -
c7f60860 by Takanori Nakane at 2019-07-24T14:47:13+01:00
Merge branch 'defect-mask-in-Polish'
- - - - -
87cc40c3 by Sjors Scheres at 2019-07-24T14:50:32+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
3161ab50 by Sjors Scheres at 2019-07-25T09:21:36+01:00
CTF-correction of references wasn't using obsModel for autopicking on GPU
- - - - -
dbacee9a by Takanori Nakane at 2019-07-25T17:02:02+01:00
MPI: use MPI_C_COMPLEX and MPI_C_DOUBLE_COMPLEX (GitHub issue 234)
- - - - -
6bcdde6c by Sjors Scheres at 2019-07-26T11:44:14+01:00
changed GUI and ml_optimiser CL parsing: offset range and steps are in pixels, not angstroms
- - - - -
145c25ac by Takanori Nakane at 2019-07-29T08:49:32+01:00
Pipeliner: removed NULL characters
- - - - -
d4b8877f by Takanori Nakane at 2019-07-29T12:09:52+01:00
IOparser: When the same argument is specified more than once, use the last one with warning
- - - - -
3cb813a1 by Takanori Nakane at 2019-07-29T12:24:32+01:00
Import: Repaired continuation. Print the number of items.
- - - - -
a9504ae5 by Takanori Nakane at 2019-07-30T11:46:22+01:00
RelionCor: make --early_binning ON by default
- - - - -
6f11c59d by Takanori Nakane at 2019-07-30T12:52:43+01:00
MotionCorr: supported defect maps (for both UCSF MotionCor2 and RelionCor)
- - - - -
43359768 by Takanori Nakane at 2019-07-30T14:17:34+01:00
Detect SerialEM's defect file, which is not supported
- - - - -
4826462c by Takanori Nakane at 2019-08-01T15:33:46+01:00
relion_image_handler: fixed "ISO C++ forbids initialization of member" error
- - - - -
25c3f9e6 by Takanori Nakane at 2019-08-02T12:59:20+01:00
RelionCor: don't "detect" hot pixels already marked in the defect map
- - - - -
23f22c51 by Takanori Nakane at 2019-08-02T15:51:50+01:00
Extract: use gamma offset in CTF premultiplication
- - - - -
a19b6ebc by Takanori Nakane at 2019-08-02T15:57:06+01:00
ml_optimiser: --pad_ctf is probably broken now. Disable it.
- - - - -
cd3294ba by Takanori Nakane at 2019-08-02T16:00:47+01:00
Polish: Don't die with MotionCorr STAR files from 3.0
- - - - -
4bff28a6 by Takanori Nakane at 2019-08-02T17:16:52+01:00
relion_convert_start: supported conversion of micrograph STAR files
- - - - -
61b275fc by Takanori Nakane at 2019-08-05T14:40:53+01:00
Removed ref_aberration_plot and RefinementProgram class.
- - - - -
9e67d506 by Takanori Nakane at 2019-08-05T15:04:31+01:00
Polish: allow polishing of only particles belonging to movies in the input MotionCorr STAR file
- - - - -
29689a93 by Takanori Nakane at 2019-08-05T17:32:45+01:00
GUI: Added --window and --scale for Polishing
- - - - -
9e2bf4a0 by Takanori Nakane at 2019-08-06T11:00:51+01:00
Subset selection: judge datablock type properly
- - - - -
93638b67 by Takanori Nakane at 2019-08-06T11:54:39+01:00
Import: fixed import of coordinates. WIP: the content of suffix STAR file is still not right.
- - - - -
0be9ba50 by Dari Kimanius at 2019-08-06T13:40:24+01:00
FSC added to external reconstruct star-file
- - - - -
16add052 by Dari Kimanius at 2019-08-06T14:33:13+01:00
Subsets fixed for half sets
- - - - -
cc070237 by Takanori Nakane at 2019-08-07T20:29:41+01:00
Refine3D: don't allow Refine3D on even number of MPI processes.
This caused 'loss' of some particles in half1 and resulted in empty groups
due to a bug in "if ((node->size - 1)/nr_halfsets > 1)" logic.
Don't know how to fix the logic yet, so disable such operation.
- - - - -
853ca84b by Takanori Nakane at 2019-08-07T20:38:00+01:00
Refine3D: don't allow Refine3D on even number of MPI processes.
This caused 'loss' of some particles in half1 and resulted in empty groups
due to a bug in "if ((node->size - 1)/nr_halfsets > 1)" logic.
Don't know how to fix the logic yet, so disable such operation.
- - - - -
451976b5 by Takanori Nakane at 2019-08-08T10:34:40+01:00
relion_mask_create: fixed --and, --or, --and_not, --or_not. (GitHub issue #495)
- - - - -
51121a05 by Takanori Nakane at 2019-08-08T10:38:10+01:00
relion_mask_create: fixed --and, --or, --and_not, --or_not. (GitHub issue #495)
- - - - -
ebc3ad60 by Takanori Nakane at 2019-08-08T16:29:57+01:00
Extract: make sure rlnMicrographPixelSize is present
- - - - -
ecdbfa5c by Takanori Nakane at 2019-08-12T16:45:08+01:00
pipeliner: make sure rlnPipeLineJobCounter to be positive
(some tempted to make it negative to avoid JobID overlaps in RelionIt...)
- - - - -
c5a3d3a0 by Takanori Nakane at 2019-08-13T15:20:45+01:00
Refactored histogram
- - - - -
e944e174 by Takanori Nakane at 2019-08-13T15:21:05+01:00
LocalRes: histogram of local resolution
- - - - -
5432f7c8 by Takanori Nakane at 2019-08-14T17:26:16+01:00
TIFF: Detect SerialEM's 4-bit packed TIFFs and raise an error.
- - - - -
76c64d31 by Takanori Nakane at 2019-08-14T20:12:57+01:00
TIFF: Support SerialEM's 4-bit packed TIFFs
- - - - -
6bb61d1d by Takanori Nakane at 2019-08-14T20:48:57+01:00
relion_pipeliner: Add --gentle_clean (Yang Lee's request)
- - - - -
abcb9850 by Takanori Nakane at 2019-08-19T20:45:26+01:00
relion_preprocess: removed unused --fn_list and repaired continuation in the MPI version
- - - - -
8ec3d101 by Takanori Nakane at 2019-08-19T21:54:41+01:00
GUI: Fixed a crash in scheduler
- - - - -
b53114b6 by Takanori Nakane at 2019-08-19T23:04:31+01:00
PostProcess: Make sure the pixel size is positive.
- - - - -
f4b0bf0d by Takanori Nakane at 2019-08-20T10:27:46+01:00
PostProcess: Rather than stopping, use the pixel size in the header when --angpix is not specified.
- - - - -
cd39b15b by Takanori Nakane at 2019-08-20T10:52:41+01:00
PostProcess: don't low pass filter by default when FSC weighting is enabled. (rationale: local resolutions can be higher)
- - - - -
4f2012de by Takanori Nakane at 2019-08-20T14:36:52+01:00
GUI: Fixed a GUI crach (the patch 8ec3d101 caused side effects)
- - - - -
f14183da by Takanori Nakane at 2019-08-20T17:19:01+01:00
CtfFind: don't die on an empty STAR file
- - - - -
250dbbab by Takanori Nakane at 2019-08-25T20:43:48+01:00
GUI: change the default angpix for PostProcess to -1 (use the header value)
- - - - -
3f3dc213 by Takanori Nakane at 2019-08-25T20:44:26+01:00
Polish: repair polishing only micrographs present in the specified MotionCorr job
- - - - -
945cff5b by Takanori Nakane at 2019-08-26T10:00:42+01:00
GUI: fix typo in GCTF's PMID
- - - - -
7996c59c by Takanori Nakane at 2019-08-26T10:04:54+01:00
GUI: respect "reconstruct subtracted bodies" in continuation of MultiBody (reported by Sean Connell)
- - - - -
e9d8fd7a by Vahid Abrishami at 2019-08-27T11:34:10+03:00
Relax sym option has been added
- - - - -
8c763e28 by Sjors Scheres at 2019-09-02T09:25:35+01:00
when importing micrographs, use micrograph_pixel_size and not original_micrograph_pixel_size
- - - - -
98d46edf by Sjors Scheres at 2019-09-02T09:28:59+01:00
added log_upper_thr
- - - - -
8efc93f0 by Sjors Scheres at 2019-09-02T09:51:32+01:00
added log_upper_thr
- - - - -
ad7b9141 by Sjors Scheres at 2019-09-02T09:52:21+01:00
small changes?
- - - - -
8213780a by Sjors Scheres at 2019-09-02T12:15:03+01:00
for micrographs set both orginal and micrograph pixel size
- - - - -
27e6a69e by Vahid Abrishami at 2019-09-02T15:40:01+03:00
prints have been removed
- - - - -
fb64c907 by Vahid Abrishami at 2019-09-04T10:00:32+03:00
some prints has been removed
- - - - -
966c00d7 by Vahid Abrishami at 2019-09-04T10:24:57+03:00
relax_sym now can be continued
- - - - -
1ea66624 by Takanori Nakane at 2019-09-04T12:53:40+01:00
Update the start up screen for 3.1.
- - - - -
713e6653 by Sjors Scheres at 2019-09-05T11:41:23+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
7074d771 by Takanori Nakane at 2019-09-05T12:34:03+01:00
Polish: remove un-processed particles and optics groups (for easier joining)
- - - - -
3696d0e6 by Sjors Scheres at 2019-09-05T12:47:02+01:00
deactived --mics and set interpolator to 2
- - - - -
7d0f09ad by Sjors Scheres at 2019-09-05T12:47:19+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
447b1582 by Takanori Nakane at 2019-09-05T14:49:56+01:00
GUI: improved explanation
- - - - -
dcff6239 by Takanori Nakane at 2019-09-05T14:49:56+01:00
Split: respect BOTH subset size and number of subsets
- - - - -
542ee9c1 by Takanori Nakane at 2019-09-05T15:53:57+01:00
GUI: Repaired access to member functions on a NULL pointer
(this might have been working but very dangerous: https://stackoverflow.com/questions/2533476/what-will-happen-when-i-call-a-member-function-on-a-null-object-pointer)
- - - - -
532bc3c2 by Takanori Nakane at 2019-09-05T16:21:14+01:00
bfactor_plot.py: updated for RELION 3.1 and supported Python 3.
- - - - -
1486db3b by Takanori Nakane at 2019-09-05T17:21:43+01:00
RelionIt: supported Python 3 (WIP: only syntax change; RELION 3.1 not supported yet)
- - - - -
3f106b38 by Sjors Scheres at 2019-09-06T11:25:13+01:00
dont let the GUI die when textToInteger/Float goes wrong
- - - - -
0b2d9df9 by Sjors Scheres at 2019-09-06T11:25:38+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
51bb4d34 by Takanori Nakane at 2019-09-06T12:15:26+01:00
ml_optimiser: add help messages to --maxsig and --skip_griding, and make --maxsig's default to -1, not 0, for consistency.
- - - - -
848092ff by Takanori Nakane at 2019-09-06T14:19:27+01:00
GUI: added "Skip gridding" except for Class2D.
- - - - -
2b19554b by Takanori Nakane at 2019-09-06T15:36:57+01:00
AutoPick; change the default shrink factor to 0.
- - - - -
d164502f by Takanori Nakane at 2019-09-06T17:38:24+01:00
RelionIt: support RELION 3.1 up to CTFFIND.
- - - - -
e8083f67 by Takanori Nakane at 2019-09-10T19:53:59+01:00
Escape strings in a STAR file.
- - - - -
f83a4246 by Takanori Nakane at 2019-09-10T19:54:07+01:00
STAR: fixed an offset bug
- - - - -
e0485a10 by Takanori Nakane at 2019-09-10T20:32:02+01:00
STAR: repair quoting condition
- - - - -
24de06cc by Vahid Abrishami at 2019-09-11T10:44:28+03:00
Removing redundant symmetry in relax mode has been removed
- - - - -
3258bf3b by Takanori Nakane at 2019-09-11T17:01:45+01:00
RelionIt: support RELION 3.1 up to Class2D (i.e. first pass is now OK)
- - - - -
27b99e52 by Takanori Nakane at 2019-09-11T17:29:17+01:00
RelionIt: fixed IniModel task
- - - - -
fce0ac37 by Takanori Nakane at 2019-09-12T10:24:46+01:00
GUI: Print --version before asking for a new project directory. Solves GitHub issue #505 from @rubyFeedback and @delarosatrevin.
- - - - -
4d097909 by Takanori Nakane at 2019-09-12T10:28:17+01:00
Escape strings in a STAR file.
- - - - -
2799df34 by Takanori Nakane at 2019-09-12T10:28:17+01:00
STAR: fixed an offset bug
- - - - -
3604d6fc by Takanori Nakane at 2019-09-12T10:28:17+01:00
STAR: repair quoting condition
- - - - -
11d61457 by Takanori Nakane at 2019-09-12T10:29:23+01:00
Merge branch 'escape_string_in_STAR' of https://bitbucket.org/scheres/relion-devel-lmb into escape_string_in_STAR
- - - - -
408bccef by Takanori Nakane at 2019-09-12T12:42:02+01:00
Remove unused beamtilt functions from StackHelper
- - - - -
ed22a456 by Takanori Nakane at 2019-09-12T12:43:51+01:00
Move FrameMerge class to jaz/archive
- - - - -
911c62e8 by Takanori Nakane at 2019-09-12T14:37:33+01:00
RelionIt: 2nd pass works now
- - - - -
dfc4de86 by Sjors Scheres at 2019-09-12T14:42:22+01:00
order for displayer of images: micrographname above power spectrum
- - - - -
33574c47 by Takanori Nakane at 2019-09-12T15:09:42+01:00
Use FLTK 1.3.5. (GitHub issue #500)
- - - - -
c6e16251 by Sjors Scheres at 2019-09-12T15:32:38+01:00
hardcore Schedules/ directory for output schedules
- - - - -
a1cf4ab7 by Sjors Scheres at 2019-09-12T15:33:04+01:00
removed scroll bars from GUI window in Scheduler mode
- - - - -
24eb7309 by Takanori Nakane at 2019-09-12T17:41:39+01:00
Polish: add recentering
- - - - -
a67cbc50 by Sjors Scheres at 2019-09-13T17:28:33+01:00
be able to work with empty string variables
- - - - -
0785e1af by Sjors Scheres at 2019-09-16T12:28:09+01:00
introduced i/o and jobs view on maingui
- - - - -
0b68c3c3 by Sjors Scheres at 2019-09-16T12:30:04+01:00
Merge branch 'master' into HEAD
- - - - -
17edd5e4 by Sjors Scheres at 2019-09-16T14:31:24+01:00
repaired bug readStarList Takanori
- - - - -
022f33fe by Sjors Scheres at 2019-09-16T14:34:01+01:00
prevent reuse of global variable name in read
- - - - -
2bccba42 by Sjors Scheres at 2019-09-16T15:45:26+01:00
no longer use __ and && for spaces and EOLs, still use restore for __ for backwards compatibility
- - - - -
7d89bca9 by Sjors Scheres at 2019-09-16T16:29:19+01:00
moving buttons around for scheduler/pipeliner1
- - - - -
66b55baa by Sjors Scheres at 2019-09-16T16:31:51+01:00
repaired bug with non-integer paddings in cuda BP kernels
- - - - -
ad170cc1 by Sjors Scheres at 2019-09-16T16:35:57+01:00
repaired bug with non-integer paddings in cuda projector kernels
- - - - -
6ba38b83 by Sjors Scheres at 2019-09-16T16:37:48+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
52c45536 by Takanori Nakane at 2019-09-16T17:14:07+01:00
Fixed --skip_gridding with non integer padding
- - - - -
9c3379ed by Sjors Scheres at 2019-09-16T17:33:25+01:00
dont restoreString with __ for input options as gainref may have __. Now always write job.star, but read run.job to maintain backwards compatibility
- - - - -
bb98cd4e by Vahid Abrishami at 2019-09-17T10:13:01+03:00
Problem with searching in a sigma has been solved
- - - - -
d86263db by Sjors Scheres at 2019-09-17T09:50:11+01:00
return false if no run.job or job.star file is present
- - - - -
a13b48df by Sjors Scheres at 2019-09-18T09:17:19+01:00
better error message for half-sets
- - - - -
5097d0ae by Takanori Nakane at 2019-09-18T12:25:12+01:00
Prevent divide by zero in skip gridding.
- - - - -
3ac296e1 by Sjors Scheres at 2019-09-18T16:17:47+01:00
keep tilt prior fixed debugged
- - - - -
a59c3f42 by Sjors Scheres at 2019-09-18T16:18:04+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
2297c520 by Takanori Nakane at 2019-09-18T17:48:53+01:00
Repaired single precision build (WARNING: accuracy of CtfRefine in single precision is not throughfully tested!)
- - - - -
8fd3a52d by Sjors Scheres at 2019-09-19T14:38:32+01:00
repaired bug in event handler, now shortcut keys work again
- - - - -
3360c244 by Sjors Scheres at 2019-09-19T14:38:42+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
27f8a301 by Sjors Scheres at 2019-09-23T17:25:40+01:00
repaired bug reprted by Tj from Leicester about manualpickrun.job, now being job.star
- - - - -
396cab45 by Sjors Scheres at 2019-09-23T17:25:57+01:00
port for 3.1 star files
- - - - -
acb90f63 by Vahid Abrishami at 2019-09-26T15:05:11+03:00
Merge branch 'master' of https://github.com/3dem/relion into HEAD
- - - - -
dcf451ad by Sjors Scheres at 2019-09-27T16:41:42+01:00
remove spaces from name of new job in scheduler
- - - - -
695b2d5c by Sjors Scheres at 2019-09-27T17:00:26+01:00
repaired bug reported by TJ from Leicester when RELION_QSUB_EXTRA_COUNT < 4
- - - - -
9a52cccd by Sjors Scheres at 2019-09-27T17:04:28+01:00
added option to harsh clean a job
- - - - -
fab92a61 by Takanori Nakane at 2019-09-30T21:08:18+01:00
relion_project: improve the help message (suggestion from @evanseitz on GitHub issue #510)
- - - - -
693b93be by Takanori Nakane at 2019-09-30T21:08:18+01:00
external_reconstruct: repair a critical bug that ignored padding_factor
- - - - -
64c94c36 by Takanori Nakane at 2019-10-01T11:59:51+01:00
Extract: repair uninitialised variables
- - - - -
7290d2cd by Takanori Nakane at 2019-10-01T12:29:16+01:00
GUI: Add local resolution histogram
- - - - -
a28e92e9 by Sjors Scheres at 2019-10-08T17:00:54+01:00
print warnings to stderr if the reference in on different scale/box size as the data
- - - - -
bd925fec by Sjors Scheres at 2019-10-09T10:08:12+01:00
spell out inside SGD
- - - - -
34a9ee6d by Sjors Scheres at 2019-10-09T10:20:31+01:00
set default CTF_IN_SGD to OFF and explain how to switch on on the GUI
- - - - -
6ba935b4 by Takanori Nakane at 2019-10-09T12:54:37+01:00
CTFRefine: report anisotropy
- - - - -
646db982 by Sjors Scheres at 2019-10-09T14:56:01+01:00
tweaked cmake message ctf i sgd
- - - - -
79162cb4 by Sjors Scheres at 2019-10-09T14:56:21+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
d1464cb8 by Sjors Scheres at 2019-10-09T15:13:08+01:00
changed warnings of pixel and box size
- - - - -
92aaf5a3 by Takanori Nakane at 2019-10-09T16:13:40+01:00
relion_reconstruct: add --class as requested by Oliver Clarke (GitHub issue #513)
- - - - -
dab3bd41 by biochem_fan at 2019-10-10T11:40:24+01:00
Create a bug-report template
- - - - -
d0a30123 by Sjors Scheres at 2019-10-14T15:24:39+01:00
made search helical symmetry netter organised, less push_backs
- - - - -
60f263ac by Sjors Scheres at 2019-10-14T15:24:51+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
5a8f8278 by Takanori Nakane at 2019-10-14T17:04:19+01:00
GUI: use "--reconstruct_subtracted_bodies" in continuation as well (as reported by Dr. Yanagisawa)
Call this 3.0.8 (hopefully final version in 3.0.x)
- - - - -
8b251028 by Takanori Nakane at 2019-10-14T17:15:03+01:00
Final preparation for 3.1 public beta
- - - - -
5d341435 by biochem_fan at 2019-10-14T17:26:22+01:00
Create a bug-report template
- - - - -
d5a160bb by Takanori Nakane at 2019-10-14T20:36:48+01:00
GUI: use formatted strings for safety
- - - - -
a6aaa580 by Takanori Nakane at 2019-10-14T21:41:25+01:00
README: fix installation instructions as pointed out by Rui Zhang
- - - - -
2ebd8861 by Sjors Scheres at 2019-10-16T09:33:20+01:00
added anisomag and scaling to particle_subtractor and to multibody inside ml_optimiser
- - - - -
8f040699 by Sjors Scheres at 2019-10-16T09:33:34+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
263d5954 by Takanori Nakane at 2019-10-16T11:02:13+01:00
relion_particle_symmetry_expand: use EMDL_ORIENT_ORIGIN_X/Y_ANGSTROM not EMDL_ORIENT_ORIGIN_X/Y.
- - - - -
8bdbc749 by Takanori Nakane at 2019-10-16T11:35:01+01:00
README: further update on installation
- - - - -
6110eaca by Vahid Abrishami at 2019-10-16T13:43:53+03:00
Merge branch 'master' of https://github.com/3dem/relion into HEAD
- - - - -
9b4ce5e3 by Takanori Nakane at 2019-10-16T11:44:09+01:00
README: more updates
- - - - -
ce28d249 by Takanori Nakane at 2019-10-16T11:46:19+01:00
README: make it clear this is 3.0 repository
- - - - -
767dc597 by Vahid Abrishami at 2019-10-16T13:52:06+03:00
Merge branch 'master' of https://github.com/3dem/relion into HEAD
- - - - -
8fe742e8 by Vahid Abrishami at 2019-10-16T14:18:08+03:00
Code cleaning
- - - - -
ff6c2066 by Takanori Nakane at 2019-10-16T16:20:16+01:00
Extract: Remove particles not present in the input micrographs
- - - - -
e09586fd by Sjors Scheres at 2019-10-16T16:23:43+01:00
starting to take out operate_on star files
- - - - -
0d069339 by Sjors Scheres at 2019-10-16T16:23:50+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
b86482c4 by Sjors Scheres at 2019-10-16T16:26:05+01:00
removed STAR file option for operate_on in preprocess
- - - - -
de05e4d6 by Joaquin Oton at 2019-10-18T12:16:38+01:00
Bug fixed. Pixelsize in reference map is checked
- - - - -
6d496657 by Joaquin Oton at 2019-10-18T12:17:39+01:00
Merge branch 'master' of bitbucket.org:scheres/relion-devel-lmb
- - - - -
0ac9f4f6 by Takanori Nakane at 2019-10-21T12:28:29+01:00
Clean clutter
- - - - -
145b0b13 by Takanori Nakane at 2019-10-21T12:28:29+01:00
loadSafely: assign a correct table name after conversion (GitHub issue #521)
- - - - -
69a4dd4a by Takanori Nakane at 2019-10-21T13:42:18+01:00
relion_display: repair --lowpass (GitHub issue #523)
- - - - -
f0b8b48a by Takanori Nakane at 2019-10-21T17:16:33+01:00
Improve CtfRefine GUI (GitHub issue #517)
- - - - -
3600ede5 by Takanori Nakane at 2019-10-21T17:18:18+01:00
STAR conversion: stronger warning when box size cannot be determined
- - - - -
a01ec785 by Takanori Nakane at 2019-10-21T17:23:59+01:00
Import: Make sure voltage and angpix is positive
- - - - -
6903cc46 by Takanori Nakane at 2019-10-22T10:28:29+01:00
STAR conversion: make the default optics group name to OpticsGroup1, not OpticsGroup0.
- - - - -
151f15bc by Joaquin Oton at 2019-10-22T11:17:13+01:00
Autodetect data_dim in case --3d_rot is not provided
- - - - -
f1f054fc by Joaquin Oton at 2019-10-22T11:17:19+01:00
Merge branch 'master' of bitbucket.org:scheres/relion-devel-lmb
- - - - -
6f8ddb70 by Takanori Nakane at 2019-10-22T11:53:02+01:00
GUI: Repaired the output file name for training
- - - - -
0b4f84cf by Takanori Nakane at 2019-10-22T12:22:42+01:00
Polish: Repaired false warnings about pixel size.
- - - - -
296f7519 by Takanori Nakane at 2019-10-22T12:24:28+01:00
GUI: Repaired the output file name for training
- - - - -
fa2ea304 by Takanori Nakane at 2019-10-22T12:24:28+01:00
Polish: Repaired false warnings about pixel size.
- - - - -
0f91b47b by Takanori Nakane at 2019-10-22T12:31:18+01:00
Merge branch 'ver3.1' of https://github.com/3dem/relion
- - - - -
e22c93f2 by Sjors Scheres at 2019-10-22T12:49:36+01:00
added warning if data dimensionality was missing from particles STAR
- - - - -
9babce43 by Sjors Scheres at 2019-10-22T12:50:01+01:00
write out reprojections and summed classes
- - - - -
a97ac093 by Sjors Scheres at 2019-10-22T12:50:23+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
7f17498b by Takanori Nakane at 2019-10-22T12:51:08+01:00
GUI: Repaired the output file name for training
- - - - -
d43db717 by Takanori Nakane at 2019-10-22T12:51:08+01:00
Polish: Repaired false warnings about pixel size.
- - - - -
dd833470 by Takanori Nakane at 2019-10-22T12:51:43+01:00
Merge branch 'ver3.1' of https://github.com/3dem/relion
- - - - -
a7b0bd52 by Takanori Nakane at 2019-10-22T13:57:02+01:00
GUI: repair crash when deleting operators with missing arguments
- - - - -
f13e1800 by Takanori Nakane at 2019-10-22T17:05:36+01:00
ObsModel: Clearer messages when the box size is not available.
- - - - -
41c4ad14 by Takanori Nakane at 2019-10-22T17:07:51+01:00
ObsModel: more warning
- - - - -
780dd7ba by Takanori Nakane at 2019-10-23T17:13:58+01:00
relion_refine: clearer error message when rlnModelStarFile2 is empty.
- - - - -
f1f21888 by Takanori Nakane at 2019-10-24T15:00:37+01:00
Polish: repaired --recenter
- - - - -
1f22d6a7 by Sjors Scheres at 2019-10-25T11:26:34+01:00
repaired bug reported by Clinton Lau
- - - - -
13869ee4 by Sjors Scheres at 2019-10-25T11:26:51+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
f8a215df by Takanori Nakane at 2019-10-28T15:03:41+00:00
scratch with MPI: repair crash when the disk is full (Github issue #529)
- - - - -
e023ce55 by Sjors Scheres at 2019-10-28T15:08:38+00:00
not ori_size but image_size in exp_imagedata array when not doing parallel disc i/o; die if multiple image sizes in input star file with not parallel disc i/o
- - - - -
d3d5419e by Takanori Nakane at 2019-10-28T15:10:50+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
ef36c564 by Takanori Nakane at 2019-10-28T15:26:39+00:00
GUI: Improved help messages for parallel disc I/O.
- - - - -
95dfc62c by Takanori Nakane at 2019-10-29T11:33:50+00:00
Subtract: restore box sizes when reverting to original particles (GitHub issue #531).
- - - - -
198bddea by Takanori Nakane at 2019-10-30T16:54:21+00:00
GUI: minor tweak for Polish parameter files
- - - - -
9efbe650 by Takanori Nakane at 2019-10-30T17:49:24+00:00
STAR conversion: Do not set EMDL_MICROGRAPH_ORIGINAL_PIXEL_SIZE for particles, because particles might have been down-sampled.
- - - - -
16ba44ad by Takanori Nakane at 2019-10-31T11:17:18+00:00
STAR conversion: Do not set EMDL_MICROGRAPH_ORIGINAL_PIXEL_SIZE for particles, because particles might have been down-sampled.
(the last commit was a wrong file ...)
- - - - -
02460531 by Takanori Nakane at 2019-10-31T15:45:08+00:00
ExpModel: more efficient sorting (hopefully related to GitHub issue #514)
- - - - -
7f8d1d64 by Takanori Nakane at 2019-10-31T19:14:02+00:00
ExpModel: remove debug message
- - - - -
76fa3d28 by Takanori Nakane at 2019-11-01T17:23:11+00:00
ExpModel: particle shuffling and sorting in place to save memory
- - - - -
82aff6f0 by Takanori Nakane at 2019-11-02T21:23:36+00:00
Polish: remove debug output
- - - - -
0841d068 by Takanori Nakane at 2019-11-04T14:49:14+00:00
GUI: Fixed help message for External job (GitHub issue #534)
- - - - -
72574efe by Sjors Scheres at 2019-11-06T16:10:26+00:00
multiple by average MTF in divideByMtf function to prevent oversharpening of 2D and 3D classes
- - - - -
a8e935cd by Sjors Scheres at 2019-11-06T16:51:02+00:00
remove do_mtf from postprocess GUI
- - - - -
6c0ecf2f by Takanori Nakane at 2019-11-11T11:54:40+00:00
MultidimArray: added const to some functions (although not sure if this helps optimizations)
- - - - -
87a80ded by Takanori Nakane at 2019-11-11T12:47:21+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
414b04ef by Takanori Nakane at 2019-11-11T21:20:07+00:00
ml_optimiser: repair a false warning introduced in e023ce55d
- - - - -
ee1f5913 by Takanori Nakane at 2019-11-11T21:30:26+00:00
Subtraction: Write the pixel size to the header
- - - - -
e1416949 by Sjors Scheres at 2019-11-12T12:02:22+00:00
fixed issue #535: apply orientations in particles assigned to a class
- - - - -
885dff3a by Takanori Nakane at 2019-11-12T14:37:37+00:00
Subtract: Fix 3D tomo (GitHub issue #532) and fix particles with different box sizes (GitHub issue #538)
- - - - -
f10a9a88 by Takanori Nakane at 2019-11-12T15:39:11+00:00
Subtraction: repair compiler warning
- - - - -
63ec0029 by Takanori Nakane at 2019-11-12T17:52:50+00:00
Image: supported writing integer MRCs
- - - - -
62f7d2c3 by Takanori Nakane at 2019-11-12T21:55:34+00:00
Subtract: fixed an indexing bug (thanks to Bojian Ding)
- - - - -
9d75256d by Takanori Nakane at 2019-11-12T22:04:58+00:00
Revert accidental addition of relion_convert_to_tiff, which is not yet ready.
- - - - -
aa8daa6a by Takanori Nakane at 2019-11-12T22:06:53+00:00
Merge branch 'master' into tiff_converter
- - - - -
d9696c63 by Takanori Nakane at 2019-11-12T22:12:24+00:00
relion_convert_to_tiff: simple conversion and gain estimation done. gain unnormalization work in progress.
- - - - -
c7599659 by Takanori Nakane at 2019-11-12T23:44:15+00:00
relion_convert_to_tiff: implemented lossless strategy
- - - - -
d29a2d3c by Takanori Nakane at 2019-11-13T14:46:21+00:00
relion_convert_to_tiff: implemented unnormalisation, tidied up I/O.
- - - - -
4c18b2de by Takanori Nakane at 2019-11-13T15:06:49+00:00
relion_convert_to_tiff: repair --estimate_gain
- - - - -
48bc8332 by Takanori Nakane at 2019-11-13T15:58:27+00:00
relion_convert_to_tiff: fix SEGV and preparing for MPI version
- - - - -
9a299847 by Takanori Nakane at 2019-11-13T20:14:33+00:00
relion_convert_to_tiff: MPI version
- - - - -
7db84830 by Takanori Nakane at 2019-11-13T21:43:41+00:00
relion_convert_to_tiff: use temporary files in simple conversion, fix message
- - - - -
1c53280f by Takanori Nakane at 2019-11-14T15:01:17+00:00
relion_convert_to_tiff: differentiate unsigned and signed short (mode 6 and 1)
- - - - -
62c59153 by Sjors Scheres at 2019-11-14T16:04:14+00:00
repaired bug reported by Wim Hagen in getAverageMtf
- - - - -
3493645a by Sjors Scheres at 2019-11-14T16:04:31+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
f29379ee by Takanori Nakane at 2019-11-14T17:09:37+00:00
Merge branch 'tiff_converter'
- - - - -
d928eb13 by Takanori Nakane at 2019-11-15T21:35:06+00:00
relion_convert_to_tiff: make frame number 1-indexed
- - - - -
f2c3d8d8 by Takanori Nakane at 2019-11-15T22:34:43+00:00
relion_convert_to_tiff: supported cases where the number of frames is not uniform
- - - - -
f045222c by Takanori Nakane at 2019-11-18T12:42:16+00:00
GUI: allow an empty MTF file name to skip MTF correction
- - - - -
74beeb58 by Takanori Nakane at 2019-11-18T17:20:01+00:00
Added MPI version of TIFF converter (pointed out by Vish)
- - - - -
da823c1f by Takanori Nakane at 2019-11-19T11:20:19+00:00
MRC: Treat mode 0 as SIGNED char to be compatible with the official speficifation and SerialEM. (It was treated as unsigned before)
TIFF: Support both signed and unsigned char
- - - - -
eb89fc97 by Takanori Nakane at 2019-11-21T11:41:29+00:00
TIFF: read and write a pixel size in the header
relion_convert_to_tiff: more checks
- - - - -
c9fadd81 by Takanori Nakane at 2019-11-21T13:44:53+00:00
relion_convert_to_tiff: accept a simple list (.lst)
- - - - -
f2335a48 by Takanori Nakane at 2019-11-21T17:45:07+00:00
JoinStar: Removed false warning reported by John Heumann
- - - - -
8bdeba85 by Takanori Nakane at 2019-11-22T15:44:27+00:00
ml_optimiser: write at the final iteration regardless of --sgd_write_iter..
- - - - -
7b3195ec by Jeffrey Salmond at 2019-11-22T16:46:22+00:00
Fix typo in tiff_converter
currently preventing ver3.1 from building with TIFF support
- - - - -
b8b4faf0 by Takanori Nakane at 2019-11-22T16:47:28+00:00
GUI: don't free return values from fl_input(). Fixes GitHub issue #545.
- - - - -
76b9772e by biochem_fan at 2019-11-22T16:48:40+00:00
Merge pull request #546 from js947/patch-1
Fix typo in tiff_converter
- - - - -
0babfbec by Takanori Nakane at 2019-11-22T16:49:54+00:00
Merge branch 'ver3.1' of https://github.com/3dem/relion
- - - - -
55e57fca by Takanori Nakane at 2019-11-22T17:09:16+00:00
star_handler: fixed the output node type (solves #544)
- - - - -
932818d1 by Takanori Nakane at 2019-11-25T12:13:33+00:00
relion_project: make it clear that --ang accepts only particle STAR files
- - - - -
6dddc666 by Takanori Nakane at 2019-11-25T12:17:24+00:00
ManualPicker: added --coord_scale. This is only for display. CoordinateX/Y should nonetheless refer to the original micrograph pixels, otherwise Extract and Polish will fail.
- - - - -
afe75e89 by TJ Ragan at 2019-11-25T16:49:09+00:00
Add environment variable for default scratch directory.
- - - - -
94d5d64f by biochem_fan at 2019-11-25T16:57:33+00:00
Merge pull request #550 from LISCB/ver3.1
Add environment variable for default scratch directory.
- - - - -
6997e2df by Takanori Nakane at 2019-11-25T17:40:02+00:00
relion_star_handler: make minumum and maximum inclusive
- - - - -
a76cf5e4 by Takanori Nakane at 2019-11-25T17:41:35+00:00
Merge branch 'ver3.1' of https://github.com/3dem/relion
- - - - -
6503c00b by Bjoern Forsberg at 2019-11-26T17:45:12+01:00
display of model.star rot 3D-helix recon by 90 deg
but only when "apply orient" is ticked
- - - - -
f511adda by Takanori Nakane at 2019-11-27T11:58:07+00:00
relion_flex_analyse: write the optics group table (pointed out by Rafael Leiro)
- - - - -
6b6f9fe9 by Takanori Nakane at 2019-11-28T17:24:43+00:00
JoinSTAR: fixed a type error
- - - - -
3bfef2bd by Sjors Scheres at 2019-11-29T12:20:45+00:00
repaired bug reported by Sean Connell about refinement hanging with multiple different MTFs
- - - - -
734866ce by Takanori Nakane at 2019-11-29T14:25:16+00:00
MetaDataTable: added MDT_TYPE_CHECK flag to cmake (only for developers). This checks type compatibility on every call to setValue and getValue.
- - - - -
fd7dc1a2 by Takanori Nakane at 2019-11-29T14:36:15+00:00
ObsModel: make critical sections independent
- - - - -
699f3dab by Takanori Nakane at 2019-12-03T11:56:25+00:00
EPS colorimage: limit the number of elements in string / array literals for compatibility with earlier Ghostscript and Preview in Mac OS.
- - - - -
f6ce7a7c by Takanori Nakane at 2019-12-03T12:24:24+00:00
TIFF: made write_tiff_one_page public.
Resolved Conflicts:
src/tiff_converter.cpp
src/tiff_converter.h
- - - - -
7d0b093a by Sjors Scheres at 2019-12-04T09:07:23+00:00
added option to debug split_random_half and repaired bug in subtracting COMs with different vdims
- - - - -
f981a5ec by Sjors Scheres at 2019-12-04T09:07:42+00:00
repaired bug in subtracting COMs with different vdims
- - - - -
29f1b42d by Sjors Scheres at 2019-12-04T09:08:21+00:00
added cisTEM-like FSC-part calculation to postprocessing
- - - - -
e01882af by Sjors Scheres at 2019-12-04T09:08:43+00:00
built in protection against operations of unequal vdim
- - - - -
a06d9cdb by Sjors Scheres at 2019-12-04T12:27:23+00:00
repaired bug in inverseFFT in multibody refinement on CPU
- - - - -
ffe049a8 by Sjors Scheres at 2019-12-04T13:54:35+00:00
removed strange initialisation of Faux in acc ml_optimiser
- - - - -
c2d7a46c by Sjors Scheres at 2019-12-04T13:54:55+00:00
check whether fFourier is expected size when deciding whether to recompute the plan
- - - - -
7a75ef36 by Sjors Scheres at 2019-12-04T14:10:21+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
2fc87bf0 by Sjors Scheres at 2019-12-04T14:11:45+00:00
for now disable FSC plot with cisTEM comparison
- - - - -
0d47c401 by Takanori Nakane at 2019-12-04T14:17:22+00:00
GUI: fix crash when choosing "job options" without any selected job (GitHub issue #552)
- - - - -
c75aba4c by Takanori Nakane at 2019-12-04T14:18:43+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
01cf0b35 by Sjors Scheres at 2019-12-04T14:41:49+00:00
made CL option to pipeliner to add a job from STAR file, and change some options
- - - - -
9f499d2f by Sjors Scheres at 2019-12-04T14:42:13+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
fe4a2efa by Sjors Scheres at 2019-12-04T17:18:44+00:00
only count pixels in mask>0.5 for cisTEM fraction f
- - - - -
db787df4 by Takanori Nakane at 2019-12-05T11:03:25+00:00
relion_display: working toward colorful manual picker. Refactored the color selection dialog.
- - - - -
ee065021 by Takanori Nakane at 2019-12-05T11:03:25+00:00
relion_display: initialise color
- - - - -
39664d85 by Takanori Nakane at 2019-12-05T11:03:25+00:00
relion_display: display and save color particles.
TODO: BUG: somehow I cannot unpick particles now
- - - - -
aab11976 by Takanori Nakane at 2019-12-05T11:03:25+00:00
ManualPicker: repair particle deletion with --coord_scale
- - - - -
8af52e32 by Takanori Nakane at 2019-12-05T11:03:26+00:00
relion_pipeliner: make --addJobFromStar working
- - - - -
9ff7c7ba by Takanori Nakane at 2019-12-05T11:03:26+00:00
bfactor_plot.py: supported job.star
- - - - -
ad3cd69c by Takanori Nakane at 2019-12-05T11:54:28+00:00
bfactor_plot.py: repair twin X axes for old matplotlib
- - - - -
9c2ab921 by Sjors Scheres at 2019-12-06T15:54:35+00:00
added define flag for cisTEM plot FSC
- - - - -
2085051e by Takanori Nakane at 2019-12-11T21:43:59+00:00
Repaired opening of coordinate suffix STAR files
- - - - -
7a05f1c3 by Takanori Nakane at 2019-12-12T14:00:00+00:00
ManualPicker: removed a debug message
- - - - -
6c49f845 by Takanori Nakane at 2019-12-12T15:57:18+00:00
GUI: fixed JobOptions::read (again!)
- - - - -
4854bbaf by Takanori Nakane at 2019-12-13T13:59:42+00:00
relion_convert_to_tiff: don't use string::back() for compatibility with older compilers
- - - - -
914542ea by Takanori Nakane at 2019-12-16T11:02:11+00:00
JoinStar: repaired a bug introduced in f2335a48, where optics group numbers are not renamed when rlnGroupName is missing (reported by Giuseppe)
- - - - -
f5721ee7 by Takanori Nakane at 2019-12-22T21:58:21+00:00
relion_convert_to_tiff: repaired support for RELION 3.1 movie STAR file
- - - - -
96bef338 by Takanori Nakane at 2019-12-22T22:37:52+00:00
STAR parser: trim trailing spaces
- - - - -
a451e833 by Takanori Nakane at 2019-12-28T11:15:02+00:00
relion_convert_to_tiff: removed compiler warnings
- - - - -
7d4fc2e6 by Kiyo Tsunezumi at 2020-01-10T18:16:49+09:00
Fixed a bug that "Browse" button was not displayed in "Raw input files".
- - - - -
5e744e41 by Kiyo Tsunezumi at 2020-01-11T10:02:19+09:00
Added support for tiff files and others.
- - - - -
2bd3ad1b by Joaquin Oton at 2020-01-13T14:47:18+00:00
Bug fixed when --sigma parameter is provided. Number of images per group is now calculated
- - - - -
e6949ab5 by Takanori Nakane at 2020-01-14T12:56:57+00:00
ml_optimiser: fixed timing code
- - - - -
0b35ccca by Sjors Scheres at 2020-01-14T12:57:22+00:00
inserted statement about SAGD patent from Structura to section 8
- - - - -
8144d6fc by Takanori Nakane at 2020-01-14T12:57:44+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
408f51a5 by biochem_fan at 2020-01-14T12:58:43+00:00
Merge pull request #564 from kiyotune/ver3.1
Fixed a bug that "Browse" button was not displayed in "Raw input files".
- - - - -
09b7e6a5 by Takanori Nakane at 2020-01-14T12:59:23+00:00
Merge branch 'ver3.1' of https://github.com/3dem/relion
- - - - -
b4203926 by Sjors Scheres at 2020-01-14T14:03:41+00:00
repaired typos and changed SGD into SAGD
- - - - -
06afb9fe by Sjors Scheres at 2020-01-14T14:04:01+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
7da6711d by Takanori Nakane at 2020-01-27T12:37:13+00:00
Converted micrograph STAR files should use rlnMicrographPixelSize not rlnImagePixelSize.
Up to now, rlnImagePixelSize was used, leading to (1) failure to open ManualPicker and (2) Autopick using a wrong pixel size.
- - - - -
7e0a5d2d by Takanori Nakane at 2020-01-27T12:43:02+00:00
ObsModel: repair a loop variable
- - - - -
c11be6a6 by Sjors Scheres at 2020-01-27T15:11:14+00:00
added option to read tau2 or FSC back in from external reconstruct programme
- - - - -
b4b08d03 by Takanori Nakane at 2020-01-28T10:38:53+00:00
Extract: "Manually set pixel size" (--set_angpix) has been non-functional since 3b7de54. Remove it from the GUI and the command-line option.
- - - - -
4e60d1ff by Takanori Nakane at 2020-01-28T16:29:40+00:00
Don't regroup particles from different optics groups into a same (noise-estimation) group
- - - - -
a5d691e8 by Takanori Nakane at 2020-01-31T12:38:30+00:00
AutoPicker: Repair a bug where non rank-0 MPI processes failed to read the reference pixel size from the header (solves #572)
- - - - -
0d19adf0 by Vahid Abrishami at 2020-02-03T15:29:02+02:00
Conflict has been removed
- - - - -
9231b187 by Takanori Nakane at 2020-02-03T15:53:21+00:00
relion_reconstruct: clean clutter (preparing for per-frame reconstruction)
- - - - -
9575df8e by Takanori Nakane at 2020-02-04T12:17:26+00:00
Don't allow padding factor less than 1.
- - - - -
07586c3b by Vahid Abrishami at 2020-02-06T13:15:43+02:00
Merge branch 'ver3.1' of https://github.com/3dem/relion into va_relion_relax
- - - - -
f9d266bd by Sjors Scheres at 2020-02-10T09:31:42+00:00
repaired critical bug in subtraction with auto-refine results
- - - - -
40a21f24 by Sjors Scheres at 2020-02-10T13:47:04+00:00
also replaced part_id to ori_img_id for ctf
- - - - -
f92c60f2 by Sjors Scheres at 2020-02-10T17:45:35+00:00
further debugging subtractor
- - - - -
1645d183 by Sjors Scheres at 2020-02-11T15:03:04+00:00
no longer pass orients around, instead modify output MDt and combine those in the end
- - - - -
57bac07a by Alexey Romaneko at 2020-02-11T21:50:08-05:00
OPTIM: merge data transfers. part 1
- - - - -
9eff2a3b by Sjors Scheres at 2020-02-12T14:22:26+00:00
repaired bu in plotting phase differences for beam tilt estimation in logfile.pdf
- - - - -
3112c488 by Takanori Nakane at 2020-02-12T14:32:29+00:00
Micrograph: more friendly error messages
- - - - -
257ec094 by Takanori Nakane at 2020-02-12T14:33:06+00:00
README: add explanation about "make install" and MPI version
- - - - -
237a506f by Alexey Romanenko at 2020-02-12T22:41:34-08:00
OPTIM: merge data transfers. part 2
- - - - -
2b1f5aba by Alexey Romanenko at 2020-02-12T22:42:14-08:00
OPTIM: merge data transfers. part 3
- - - - -
eb55b063 by Alexey Romanenko at 2020-02-12T22:42:59-08:00
OPTIM: merge data transfers. part 4
- - - - -
245c441d by biochem_fan at 2020-02-14T10:45:28+00:00
Merge pull request #575 from arom4github/optim
OPTIM: merge data transfers
- - - - -
c17c8967 by Sjors Scheres at 2020-02-14T11:59:33+00:00
repaired bug in reconstruct when dealing with multiple optics groups
- - - - -
de27462f by Takanori Nakane at 2020-02-14T16:09:52+00:00
Merge branch 'master' into nv-optim-datatransfer
- - - - -
0de3f2d5 by Alexey Romanenko at 2020-02-17T00:40:44-08:00
move init_pdf_orientation on GPU
by this commit, we eliminate one data transfer and move calculation of init_pdf_orientation on GPU
- - - - -
7f1ec1c3 by Vahid Abrishami at 2020-02-17T11:36:45+02:00
Merge branch 'ver3.1' of https://github.com/3dem/relion into va_relion_relax
- - - - -
baea39fd by Takanori Nakane at 2020-02-20T11:20:16+00:00
Removed dependency on libjpeg (JPEG is never used in RELION).
- - - - -
f8693be7 by Takanori Nakane at 2020-02-20T11:21:08+00:00
Merge branch 'nv-optim-datatransfer'
- - - - -
2784a5f5 by biochem_fan at 2020-02-20T11:24:16+00:00
Merge pull request #578 from arom4github/optim1
move init_pdf_orientation on GPU
- - - - -
1ca266fc by Vahid Abrishami at 2020-02-25T15:56:06+02:00
Merge branch 'ver3.1' of https://github.com/3dem/relion into va_relion_relax
- - - - -
98676205 by Vahid Abrishami at 2020-02-25T16:13:30+02:00
New flag added to check for relax option
- - - - -
7fb53e5f by Sjors Scheres at 2020-02-26T09:35:42+00:00
set default skip_gridding to true. In future perhaps remove from GUI and make a --gridding instead of a --skip_gridding refine CL option
- - - - -
a8382367 by Sjors Scheres at 2020-02-26T09:36:23+00:00
removed some debug statements in skip_gridding
- - - - -
ff549659 by Sjors Scheres at 2020-02-26T09:37:50+00:00
also changed wording of help text
- - - - -
fa896427 by Sjors Scheres at 2020-02-26T13:01:41+00:00
repaired bug #580 reported by JoeBeton
- - - - -
8ba188c0 by Sjors Scheres at 2020-02-26T14:08:02+00:00
implemented faster convergence in auto-refine and put it on GUI
- - - - -
16183114 by Sjors Scheres at 2020-02-26T14:13:55+00:00
Use finer angular sampling faster instead of converge faster on auto-refine GUI
- - - - -
a0af57df by Takanori Nakane at 2020-02-26T14:30:39+00:00
Merge branch 'nv-optim-datatransfer'
- - - - -
12d9d527 by Alexey Romanenko at 2020-02-26T08:12:48-08:00
Fix timers, optimize two kernels, fixed issue with pixel coordinate calcuation
- - - - -
29de8077 by Sjors Scheres at 2020-02-27T14:32:55+00:00
allow saving coordinates with CTRL-S in the manual picker
- - - - -
4463657e by Sjors Scheres at 2020-02-27T17:37:55+00:00
add also CTRL-Q to close picker window
- - - - -
6b948768 by Sjors Scheres at 2020-02-27T17:38:57+00:00
add option to CTRL-click a Pick button to open multiple micrographs simultaneously
- - - - -
04b25258 by Sjors Scheres at 2020-02-27T17:39:33+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
2399ef4d by biochem_fan at 2020-02-28T12:35:26+00:00
Merge pull request #581 from arom4github/optim2
Fix timers, optimize two kernels, fixed issue with pixel coordinate calculation
- - - - -
4a17fbea by Sjors Scheres at 2020-02-28T14:11:34+00:00
faster 2D/3D classifications by allowing coarser angular and translational samplings in initial iterations
- - - - -
0053ba19 by Sjors Scheres at 2020-02-28T17:44:52+00:00
first phase of replacing centerfft with centerfftbysign
- - - - -
1a4163b4 by Takanori Nakane at 2020-02-28T23:10:39+00:00
Nelder-Mead: Repaired memory leak
- - - - -
2d1a6e49 by Sjors Scheres at 2020-03-02T10:09:20+00:00
centerfftbysign in autopicker
- - - - -
25c642d4 by Sjors Scheres at 2020-03-02T10:51:46+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
18abf519 by Takanori Nakane at 2020-03-02T14:35:54+00:00
Merge branch 'nvidia-optim2'
- - - - -
7893c918 by Joaquin Oton at 2020-03-03T12:33:43+00:00
Merge branch 'master' of bitbucket.org:scheres/relion-devel-lmb
- - - - -
6ee9bdef by Joaquin Oton at 2020-03-03T14:03:57+00:00
Bugfixed: removeDuplicatedParticles in metadata_table.cpp won't consider Z coordinate #584
- - - - -
06971980 by Sjors Scheres at 2020-03-04T15:09:15+00:00
centerfftbysign for reconstructor and subtractor; ewald sphere tested now
- - - - -
284137ba by Sjors Scheres at 2020-03-04T15:09:45+00:00
Takanori suggested further optimisation of centerfftbysign
- - - - -
c3cb32b1 by Sjors Scheres at 2020-03-04T17:34:14+00:00
go back to without sorted_idx
- - - - -
f57bd847 by Takanori Nakane at 2020-03-04T17:50:53+00:00
ManualPicker: added Ctrl+1 to Ctrl+6 for color selection. TODO: synchronize this with the pop up window. prevent multiple pop up windows.
- - - - -
68bbeea9 by Joaquin Oton at 2020-03-05T12:30:36+00:00
Bug fixed when --sigma parameter is provided. Number of images per group is now calculated in MPI version
- - - - -
29632197 by Sjors Scheres at 2020-03-06T11:38:47+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
7f66ca02 by Sjors Scheres at 2020-03-06T12:17:00+00:00
sort particles indices instead of whole particle vector; calls to getFT, getDelta2 etc inside expectationOneParticle still work on original part_id, but above that, use the index from the sorted vector
- - - - -
01abbf86 by Takanori Nakane at 2020-03-06T12:45:41+00:00
relion_reconstruct: we cannot perform Ewald sphere correction on CTF premultiplied particles
- - - - -
57dd8011 by Takanori Nakane at 2020-03-06T12:45:57+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
b6c9f439 by Sjors Scheres at 2020-03-06T16:09:39+00:00
repaired bug in half-set division
- - - - -
3eef41f6 by Takanori Nakane at 2020-03-07T22:08:45+00:00
obs_model: repaired a false warning about rlnImagePixelSize reported by John M. Heumann in CCPEM.
- - - - -
6d9a0da7 by Takanori Nakane at 2020-03-09T14:11:41+00:00
Merge branch 'newsort'
- - - - -
016bc040 by Vahid Abrishami at 2020-03-16T12:53:12+02:00
Confflict solved
- - - - -
d86d3c94 by Sjors Scheres at 2020-03-20T15:32:54+00:00
Subtract job wasn't taking additional arguments
- - - - -
50fd10de by Takanori Nakane at 2020-03-27T09:58:51+00:00
relion_star_handler: added --random_seed
- - - - -
164b0f02 by Takanori Nakane at 2020-03-27T10:00:23+00:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
5d1eaca4 by Takanori Nakane at 2020-03-31T15:58:57+01:00
Write a pixel size to the header of shiny particles (solves GitHub issue #597)
- - - - -
d92f3f28 by Takanori Nakane at 2020-04-01T14:23:21+01:00
TBB: Updated URL and MD5SUM for TBB 2018 U3 (GitHub issue #594)
- - - - -
ca101fab by Takanori Nakane at 2020-04-02T21:26:39+01:00
relion_display: repaired training set contribution (applies to LMB only)
- - - - -
f6b9e2ed by Takanori Nakane at 2020-04-13T17:53:27+01:00
ml_optimiser: repaired a bug reported by James Conway on CCPEM
- - - - -
3533f637 by Sjors Scheres at 2020-04-16T14:58:56+01:00
repiared bug reported by Max wilkinson in subtraction with different image sizes in optics groups
- - - - -
94e2802f by Sjors Scheres at 2020-04-16T14:59:24+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
2bed2f63 by Takanori Nakane at 2020-04-16T17:17:17+01:00
relion_reconstruct_mpi: fixed the class name in error message
- - - - -
e3fa664c by Takanori Nakane at 2020-04-16T17:17:52+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
84d160fc by Takanori Nakane at 2020-04-20T11:03:38+01:00
flex_analyser: don't allow MPI when performing PCA
- - - - -
cd096c0c by Sjors Scheres at 2020-04-21T13:20:50+01:00
repaired bugs by KThurber2
- - - - -
649aa389 by Sjors Scheres at 2020-04-21T13:21:35+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
9d4609b4 by Sjors Scheres at 2020-04-21T14:12:27+01:00
removed orientational_prior_mode from the sampling object
- - - - -
2258321d by Sjors Scheres at 2020-04-21T14:18:26+01:00
changed orientational_prior_mode in sampling object in GPU code too
- - - - -
75dfe143 by Sjors Scheres at 2020-04-21T14:55:41+01:00
renamed do_local_search to include _helical for improved clarity
- - - - -
4e274013 by Sjors Scheres at 2020-04-21T16:22:56+01:00
further simplified the handling of helical priors, still to be tested
- - - - -
b741ab5b by Dari Kimanius at 2020-04-21T21:47:51+01:00
Relion now compiles if libTIFF was not found.
Also, cmake messages regarding system occurrence of libPNG, libopenMP and libTIFF are outputted only once in proper cmake format.
- - - - -
d7a58811 by Sjors Scheres at 2020-04-24T10:42:22+01:00
revsered sign rot prior in code by KThurber
- - - - -
d9309ddc by Sjors Scheres at 2020-04-24T13:29:01+01:00
only write out number of opposites if psi or rot prior is used
- - - - -
6427ef63 by Sjors Scheres at 2020-04-27T08:49:02+01:00
added helical_nstart to helical prior calculation to go back the N-start twist and rise to avoid very large fluctuations in predicted rot priors for small errors in twist and rise
- - - - -
d28fb9f5 by Vahid Abrishami at 2020-04-27T11:44:48+03:00
Merge branch 'ver3.1' of https://github.com/3dem/relion into va_relion_relax
- - - - -
911bdc55 by Sjors Scheres at 2020-04-27T18:04:37+01:00
introduced search for both directions of distance along helix for rot prior, reverse direction if needed by making distances negative
- - - - -
ae50cb77 by Takanori Nakane at 2020-04-27T20:59:31+01:00
GUI: Removed harmful spaces from the help message of MultiBody body STAR file.
- - - - -
c6c5c7c2 by Takanori Nakane at 2020-04-27T21:00:31+01:00
Merge branch 'master' of https://bitbucket.org/scheres/relion-devel-lmb
- - - - -
23bbe42e by Sjors Scheres at 2020-05-06T12:31:23+00:00
Merged in kthurber2 (pull request #33)
Kthurber2
- - - - -
40c8d214 by Björn Forsberg at 2020-05-08T09:09:18+02:00
Merge pull request #1 from 3dem/ver3.1
Ver3.1
- - - - -
6782e81b by Takanori Nakane at 2020-05-12T12:43:03+01:00
Allow Intel MKLFFT for GNU compilers
- - - - -
a67564a7 by Takanori Nakane at 2020-05-12T14:28:54+01:00
MultidimArray: always use aligned malloc for better FFTW performance. This should fix an unexpected crash when an FFTW plan is re-used illegaly with different memory alignment.
- - - - -
5cebbd53 by Takanori Nakane at 2020-05-13T23:05:44+01:00
Regrouping: Fixed issue #621 reported by Huw Jenkins.
- - - - -
307c8766 by Takanori Nakane at 2020-05-13T23:10:46+01:00
Merge branch 'master' into devel-fft-improvement
- - - - -
335f2d5c by Vahid Abrishami at 2020-05-20T11:36:02+03:00
Merge with ver3.1
- - - - -
8a7eb111 by Takanori Nakane at 2020-05-24T22:11:40+01:00
relion_ctf_toolbox: repair defV
- - - - -
3a49598c by Takanori Nakane at 2020-05-24T22:11:57+01:00
GUI: Check either Polish or Tune is selected.
- - - - -
b4a3aa5a by Takanori Nakane at 2020-05-27T15:41:39+01:00
relion_image_handler: supported writing a PNG file (only single image per call; e.g. "--i 1 at stack.mrcs --o 1.png")
Colouring functions were moved from displayer.h/cpp to image.h/cpp because displayer.cpp is not compiled without GUI.
- - - - -
4268a182 by biochem_fan at 2020-05-27T15:47:25+01:00
Merge pull request #620 from bforsbe/ver3.1
rotate helical display from model.star
- - - - -
e176a91d by Takanori Nakane at 2020-05-27T16:43:12+01:00
Extract: make the program robust to absent columns
- - - - -
8275ac61 by Takanori Nakane at 2020-05-29T15:11:18+01:00
relion_image_handler: write the REAL pixel size when the box size is rounded by --rescale_angpix. --force_header_angpix can overwrite the header pixel size. --stats shows the header pixel size.
- - - - -
50bfe719 by Takanori Nakane at 2020-06-01T12:05:55+01:00
CMake: clean clutter
- - - - -
5997001f by Takanori Nakane at 2020-06-01T12:06:20+01:00
relion_refine: --print_symmetry_ops also shows Euler angles
- - - - -
b781fe80 by Vahid Abrishami at 2020-06-01T14:11:19+03:00
Merge branch 'ver3.1' of https://github.com/3dem/relion into ver3.1-relax
- - - - -
9090fdeb by Takanori Nakane at 2020-06-09T11:47:13+01:00
regrouping: repaired a bug reported by @huwjenkins (GitHub issue #629)
- - - - -
35b22622 by Takanori Nakane at 2020-06-12T12:26:55+01:00
MetadataTable: repaired the label index when skipping rlnSortedIndex (Reported by Kailu Yang on CCPEM)
- - - - -
3dbc2635 by Takanori Nakane at 2020-06-12T16:41:39+01:00
relion_pipeliner: supported External job type as suggested by Colin Palmer
- - - - -
c619a1a7 by Takanori Nakane at 2020-06-13T21:44:25+01:00
CTF: Respect gammaOffset in CTFP
- - - - -
aa088f11 by Jasenko Zivanov at 2020-06-13T22:56:18+02:00
aberrations are now interpreted as given in undistorted coordinates when they are applied
- - - - -
5857e17f by Takanori Nakane at 2020-06-13T22:22:10+01:00
Merge branch 'aniso_aberrations_fix'
- - - - -
a72dcc25 by Vahid Abrishami at 2020-06-15T11:12:50+03:00
Merge branch 'ver3.1' of https://github.com/3dem/relion into ver3.1-relax
- - - - -
bea738fe by biochem_fan at 2020-06-15T21:11:13+01:00
Merge pull request #632 from LSB-Helsinki/ver3.1-relax
Symmetry relaxation contributed by Vahid Abrishami.
- - - - -
a01b0800 by Takanori Nakane at 2020-06-16T16:11:43+01:00
Added contributors
- - - - -
1bc958fa by Sjors Scheres at 2020-06-17T10:55:28+01:00
automatically rescale the reference to the smallest pixel size of the input optics groups
- - - - -
cfbffbbc by Sjors Scheres at 2020-06-17T10:57:49+01:00
included fabs in pixel size check: dont use smaller pixels than smallest input optics group either!
- - - - -
226d3e41 by Sjors Scheres at 2020-06-17T13:37:52+01:00
added functionality to opitmise scale and bfactor prior to subtraction
- - - - -
61cf47ba by Takanori Nakane at 2020-06-17T14:49:34+01:00
GUI and pipeliner: Fixed illegal memory access in HealPix order choice. Made the code more robust against illegal inputs.
- - - - -
e7716483 by Takanori Nakane at 2020-06-17T15:18:15+01:00
Merge branch 'master' of https://github.com/3dem/relion-devel
- - - - -
49c5a583 by scheres at 2020-06-17T16:26:50+01:00
die if reference size incompatible with data, unless --trust_ref_size is given
- - - - -
885321f3 by scheres at 2020-06-17T16:27:12+01:00
Merge branch 'master' of https://github.com/3dem/relion-devel
- - - - -
7be1eb75 by Takanori Nakane at 2020-06-17T17:11:56+01:00
MlModel: tidy up messages
- - - - -
8eecda3b by Takanori Nakane at 2020-06-17T17:13:24+01:00
Merge branch 'master' into devel-sym-relax
- - - - -
d366c360 by Takanori Nakane at 2020-06-19T20:50:52+01:00
MetadataTable: Column numbers should count from 1 (a bug introduced in 35b2262)
- - - - -
3c824b42 by Takanori Nakane at 2020-06-19T20:54:39+01:00
Merge branch 'master' into devel-sym-relax
- - - - -
8476f0eb by Takanori Nakane at 2020-06-24T20:57:20+01:00
Made FileName.decompose more robust to fix a problem with new SerialEM movie names (reported by Dr. Shigematsu). Nonetheless we are against the use of @ in the filename; it might break other programs/scripts using RELION STAR files.
- - - - -
4f1acaf7 by Takanori Nakane at 2020-06-24T20:57:31+01:00
Subtract and MultiBody now respect CTF premultiplication (reported by Sean Connell).
- - - - -
e1bcf0ae by Takanori Nakane at 2020-06-26T10:57:18+01:00
GUI: removed '=' from arguments help (Reported by Juha Huiskonen in GitHub issue #636)
- - - - -
17889467 by Takanori Nakane at 2020-06-27T07:00:48+01:00
Subtract: Repair recentering when the input has anisomag, multiple pixel sizes or boxsizes (reported and analyzed by Max Wilkinson)
- - - - -
9dec2988 by Takanori Nakane at 2020-06-29T12:22:30+01:00
Code clarifications suggested by @do-jason in #637 and #638.
- - - - -
87247459 by Takanori Nakane at 2020-06-29T15:13:09+01:00
Extract: added --ref_angpix
- - - - -
d8f7e6c7 by scheres at 2020-07-02T10:41:43+01:00
repaired bugs in autopicker and local symmetry
- - - - -
53514d7c by Takanori Nakane at 2020-07-03T12:44:46+01:00
relion_reconstruct: renamed a confusing variable name (pointed out by Jasenko)
- - - - -
7607fcb6 by Takanori Nakane at 2020-07-03T14:18:30+01:00
GUI: Update citations
- - - - -
1e738ecb by Takanori Nakane at 2020-07-03T14:18:44+01:00
This is version 3.1.0.
- - - - -
cd4f852b by Roland Fehrenbacher at 2020-09-28T17:07:06+00:00
Add d/dir from 1.4+dfsg-4 version
- - - - -
53b20227 by Roland Fehrenbacher at 2020-09-28T17:07:06+00:00
d/dir changes for new 2.1
- - - - -
414459b9 by Roland Fehrenbacher at 2020-09-29T14:31:20+02:00
d/dir changes for new upstream 3.1.0
- - - - -
59b3cb94 by Roland Fehrenbacher at 2020-09-29T16:23:52+02:00
Update d/copyright.
- - - - -
f365c1e1 by Roland Fehrenbacher at 2020-09-29T17:42:14+02:00
d/rules: Fix uscan and drop describe-current-version
- - - - -
5fab971b by Roland Fehrenbacher at 2020-09-29T19:00:26+00:00
Add simple autopkgtest
- - - - -
36094ad6 by Steffen Moeller at 2020-09-29T19:00:40+00:00
d/u/metadata: yamllint
- - - - -
5ec09817 by Roland Fehrenbacher at 2020-09-29T19:01:43+00:00
Update d/changelog for new release
- - - - -
24 changed files:
- + .github/ISSUE_TEMPLATE/bug_report.md
- .gitignore
- AUTHORS
- CMakeLists.txt
- + LICENSE
- README.md
- + betaGuide.pdf
- cmake/BuildFFTW.cmake
- cmake/BuildFLTK.cmake
- + cmake/BuildTBB.cmake
- cmake/BuildTypes.cmake
- cmake/FindFFTW.cmake
- + cmake/FindTBB.cmake
- + cmake/GetGitRevisionDescription.cmake
- + cmake/GetGitRevisionDescription.cmake.in
- − cmake/fltk-1.3.3.patch
- + data/mtf_de20_300kV.star
- + data/mtf_falcon2_300kV.star
- + data/mtf_falcon3EC_200kV.star
- + data/mtf_falcon3EC_300kV.star
- + data/mtf_k2_300kV.star
- + debian/README.source
- + debian/changelog
- + debian/control
The diff was not included because it is too large.
View it on GitLab: https://salsa.debian.org/med-team/relion/-/compare/216db63dde13facfb21095f01e7ec43d0d5d3134...5ec098176b28bd0df328c1d56f260cb45539d18f
--
View it on GitLab: https://salsa.debian.org/med-team/relion/-/compare/216db63dde13facfb21095f01e7ec43d0d5d3134...5ec098176b28bd0df328c1d56f260cb45539d18f
You're receiving this email because of your account on salsa.debian.org.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://alioth-lists.debian.net/pipermail/debian-med-commit/attachments/20200929/80627b02/attachment-0001.html>
More information about the debian-med-commit
mailing list