[sane-devel] Bad Performance uhci (vs. usb-ohci ,ohci)

[Theodore Kilgore]

Hello,

I am running usb-uhci on a box with 1G Athlon and SIS chipset (FIC AZ11
board) and kernel 2.4.17. I compiled the usb-uhci support in the kernel
(CONFIG_USB_UHCI_ALT=y),
not a module (not, I suspect, that this makes any difference in time
tests; I just felt like doing it that way).

I have not had any experience with ohci (no boards using it), and I have
never tried out the "old" uhci support on this machine or any other.

My time result for a scan with merlin670 and the Canoscan N670U is

40 seconds, beginning with ./merlin670 <Enter> and ending with exit of
program and reappearance of command prompt. This time includes the
initialization sequence at the beginning and also includes parking the
sensor head at the end. Actual time for scanning is approximately 30
seconds (or less). Head parking takes a while, by its nature.

Theodore Kilgore

On Sun, 10 Feb 2002, Gene Heskett wrote:

> On Sunday 10 February 2002 06:38 am, Bertrik Sikken wrote:
> >Hi,
> >
> >Gunther Mayer wrote:
> >> the "uhci" module suffers from very bad performance
> >> (linux-2.4.17).
> >>
> >> Scanning 150dpi A4 (merlin670 for Canon N670U over
> >> libusb-0.1.4) gives: uhci:    135 sec.
> >>   usb-uhci: 43 sec.
> >>   ohci:     35 sec
> >>
> >> This corresponds to about 50K/sec for 'uhci' (and the scanner
> >> must stop'n'start) !
> >>
> >> Is this a known bug in "uhci" ?
> >
> >I found something similar, with transfer rates below 100 kB/s
> >while it was something like 700 kB/s before, resulting in
> >heavy backtracking of the scanner head of my HP3300 scanner.
> >I too used (a custom compiled) kernel 2.4.17, libusb 0.1.4 and
> >uhci (compiled as module). Using the usb-uhci module instead of
> >the uhci module solved it for me. I did not further research the
> >issue, so I can't really tell for sure what caused the problem.
> >
> >Regards,
> >Bertrik Sikken
>
> How are you fellows effecting this change in usb drivers?  Here,
> on an M7VIB mobo, 1400mhz AthlonDX, I have uhci compiled in and
> have noted no such problems.  Now on my other machine, a TYAN
> TRINITY 100 mobo, any uhci eventulluy crashes if more than a
> mouse is used on the bus.
>
> --
> Cheers, Gene
> AMD K6-III at 500mhz 320M
> Athlon1600XP at 1400mhz  512M
> 98.5+% setiathome rank, not too shabby for a hillbilly
> _______________________________________________
> Sane-devel mailing list
> Sane-devel at www.mostang.com
> http://www.mostang.com/mailman/listinfo/sane-devel
>