[linux-audio-user] Advice needed re: latency tuning/optimization, esp. video

[Chris Metzler]

Hi.  I'm sorry for my tardiness in replying to this; it's been a really
difficult last couple of weeks.  I'm CC'ing you directly as well because
of it being so long; hope that's OK.

On Tue, 17 Feb 2004 16:01:36 +0100
Maarten de Boer <mdeboer at iua.upf.es> wrote:
>On Tue, 17 Feb 2004 15:43:20 +0100
>Maarten de Boer <mdeboer at iua.upf.es> wrote:
>> 
>>> Matrox Millenium G550 video card, AGP4x, 32MB RAM
>> 
>> Hmm. I remember there was an issue with Matrox and the low latency
>> patch... It should be in the archives...
> 
> See:
> 
> http://www.eca.cx/lad/2002/06/0076.html
> 
> though it could very well be that by now this has been included by
> Andrew Morton in the LL patch...

Yeah, when I first read your message, I went hunting in the archives
and found that thread.  Before my original post, I'd searched the
archives on "latency" and got a zillion responses, too many to read;
but it never occurred to me to search on "Matrox".  Duh.

Anyway, that was an interesting thread, thanks.  Looking at the
2.4.23 sources, it doesn't look like that patch (from Jussi Laako)
has been merged in to the kernel source; looking at the diffs in
the lowlatency patch, it doesn't look like they're there, either.
(although it's odd; there's a diff for mga_dma.c present in the
LL patch -- that's the same file patched in Jussi Laako's patch
-- but the diff is empty.  It's as if Andrew Morton set up to do
it, and then didn't.

So I guess I'm wondering how to proceed at this point.  I could
ask Andrew Morton what he thought of the patch, and if there was
some reason it didn't make it into his LL patch.  I could troll
here for people who've been using it for a while, to see what they
think.  I could just apply it and see what happens . . .if you
have any thoughts/suggestions, I'd be grateful.

Anyway, thanks muchly.

-c

P.S.
<off-topic>
I find everything with DRI and hardware acceleration in Linux almost
impossible to understand.  Maybe I'd get it if I were a kernel or X
hacker; but as it stands, from the outside, it's impenetrable.  Case
in point:  the lockups I mentioned in another post when doing very
intensive 3D/OpenGL stuff.  At one point, I did a lot of research,
and came up with the following possibilities:

1.  A problem with the kernel's support for my motherboard's chipset
-- various people had complained about such lockups (during 3D
endering) for Via chipsets on the LKML.

2.  A problem with the kernel modules for my video card.

3.  A problem with the XFree86 modules for my video card.

4.  A problem with the Mesa (OpenGL) libraries.

Without understanding these components in a fair amount of detail,
it's impossible to figure out where the problem is coming from, and
thus impossible to file a useful bug report (against what do I file
it?).  So here I am a year and a half later, still not able to do
OpenGL, except for the simplest things like glxgears and occasionally
tuxracer (but sometimes even that causes a lockup).  Anyway.
</off-topic>

-- 
Chris Metzler			cmetzler at speakeasy.snip-me.net
		(remove "snip-me." to email)

"As a child I understood how to give; I have forgotten this grace since I
have become civilized." - Chief Luther Standing Bear
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
Url : http://lists.linuxaudio.org/pipermail/linux-audio-user/attachments/20040225/b82635c6/attachment.pgp