[pymvpa] Time for single permutation varies between a few seconds and several minutes

[Michael Hanke]

Hey,

On Thu, Apr 9, 2015 at 11:42 AM, Jan Derrfuss <derrfuss at gmail.com> wrote:

> Some additional observations:
>
> - I ran 100 searchlights using the same dataset to see if the issue also
> occurs; it does not, all searchlights took 11-12 seconds
>
> - when the permutation takes minutes, the times are not random; they were
> (in s): 225, 236, 225, 611, 241, 247, 267, 232; for the times in the 200's,
> this is roughly the time for one fast permutation multiplied by 20
>
> - the effect also occurs with assure=False



Random thought: Maybe your learning algorithm takes a long time to converge
in some cases -- maybe especially so when there is nothing to learn...

Michael

-- 
Michael Hanke
http://mih.voxindeserto.de
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.alioth.debian.org/pipermail/pkg-exppsy-pymvpa/attachments/20150409/450f4172/attachment.html>