Hi all,
When you write reports like this, cc to linux-kernel too.
It is important to build acceptance that it CAN be done.
Next step is to get it into Linus kernel.
/RogerL
Benno Senoner wrote:
>
> Hi,
>
> Chris Baugher sent me a patch (thanks Chris) for latencytest in order to
> support PNG.
> So if you got compilation errors because your libgd does not support
> GIF anymore, use this version please.
>
> http://www.gardena.net/benno/linux/latencytest-0.42-png.tar.gz
>
> I just finished to compile 2.4.0-test5-pre4 + Andrew-lowlatency ,
> so the latencytests on this kernel should follow in a couple of hours.
>
> Andrew: read below why I think that 2.4 lowlatency patches
> (not the ones accepted by linus but the kludgeized ones)
> should deliner Ingo-paches-like performance.
> Swapping my occur on a highly loaded multimedia system,
> and no one can guarantee that other apps (not our killer
> multimedia app) does not make heavy use of mmap().
>
> The 2.2.16-lowlatency RPM I posted today works great,
> and, I stressed the disksampler for this purpose:
>
> PII400 + 256MB RAM
>
> - started a small app which mlocks 128MB and then sleeps
> for infinite time, in order to increase the memory pressure.
> - my usual 10 xterms , 10-15 netscape windows , kmail
> - compilation of the 2.4-test kernel
> - find /
> - about 60MB of ulilized swapaea
>
> disksampler was running with 35MB of mlocked mem
>
> and I tried my usual triggering of 20 voices (my MIDI keyboard
> does not seem to be able to manage more than 20 active notes)
>
> I hoped at least that the disk stream dropped out
> (256KB buffers = 3sec latencies per track), but no I was unable
> to generate the "associated stream not ready" message.
>
> nice nice .....
>
> cheers,
> Benno.
--
Home page:
http://www.norran.net/nra02596/
