[linux-audio-dev] Re: more 2.5 experiences

New Message Reply About this list Date view Thread view Subject view Author view Other groups

Subject: [linux-audio-dev] Re: more 2.5 experiences
From: Joern Nettingsmeier (nettings_AT_folkwang-hochschule.de)
Date: Wed Oct 16 2002 - 09:15:57 EEST


Bob Ham wrote:
>
> Hi all,
>
> I've been running 2.5 for about a week now without any troubles. I did try
> 2.5.39, but it didn't want to stay up for more than 10 minutes. 41 and 42
> have been solid (and that's with all the usual whizbang usb, bttv, etc, stuff
> as well.)

good news.

> Overall, imho, 2.5 seemed less responsive than the 2.4.19 kernel that I'm
> coming from. The mouse cursor sticks on the screen under load, and I had to
> increase the buffer size in sweep because it was dropping out like nobody's
> business.

yeah, same here. but it *does* seem to boot faster :)

> But then I gave 2.5.42-mm3 a whirl, and it seems to be
> considerably more responsive, and sweep drops out much less.

i have not tried it. did you hand-apply the patch, or is there an
updated version against 2.5.42 somwhere ?
 
> I've done some latency tests as well:
>
> http://pkl.net/~node/lad/latency-tests/
>
> The machine is an athlon 1.4GHz, 512MB ram, xfs fs, sblive sound card,
> matrox g550 graphics card.
>
> There's quite(!) a descrepency between Joern's and my X11 performance. This
> makes me wonder 2 things: what card are you using Joern?, and what (if
> anything) can be done to make the X11 performance less ludicrously bad?

it's a trusty ole 3dfx voodoo 3 3000 agp, running in DRI mode with full
3d acceleration (although quake3 performance drops from 41.5 fps to 28),
driving a 17" screen at 1280x1024, with a ps2 mouse and keyboard
attached. that's it.

x11 is the least of my latency problems :)

> The 2.4.19 benchmarks in there are probably bad due to the fact that it's
> xfs+lowlatency, but apparently there's ways to do that properly which I
> didn't do. Other than that, fairly dry benchmarks; just showing slightly
> better performance with each patch.

btw, i see you are using only 2x256 buffers with latencytest. since the
measured results are more-or-less independent of the setting, i'm going
to to that, too and remove the other results for clarity. or is there
some wisdom to be gathered for the wise if i leave them in ? (it's takes
painfully long to create them all...)

regards,

jörn

-- 
Jörn Nettingsmeier     
Kurfürstenstr 49, 45138 Essen, Germany      
http://spunk.dnsalias.org (my server)
http://www.linuxdj.com/audio/lad/ (Linux Audio Developers)


New Message Reply About this list Date view Thread view Subject view Author view Other groups

This archive was generated by hypermail 2b28 : Wed Oct 16 2002 - 09:25:21 EEST