Re: [linux-audio-user] Latency issues with laptops

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

Subject: Re: [linux-audio-user] Latency issues with laptops
From: torbenh_AT_gmx.de
Date: Fri Oct 15 2004 - 13:34:53 EEST


On Mon, Oct 11, 2004 at 09:06:23PM -0400, Lee Revell wrote:
> On Mon, 2004-10-11 at 20:51, Mark Knecht wrote:
> > On Mon, 11 Oct 2004 21:02:03 -0400, Peter Lutek <plutek_AT_infinity.net> wrote:
> > > On Mon, 2004-10-11 at 19:01, Lee Revell wrote:
> > > --snip--
> > > > But we can at least identify the problematic laptops,
> > > > complain to the manufacturers, and warn people not to buy them.
> > >
> > > has anyone yet compiled such a list?
> > >
>
> > I suppose my laptop (also a Compaq - an R3070us) could be considered a
> > 'problematic' laptop also, but my thought is that I don't really care
> > about ACPI when I'm doing audio anyway. I'm not running on batteries
> > or doing any of that sort of stuff when I'm recording and mixing. I
> > want ACPI and every other distraction to the machine turned off. For
> > instance, I'm seeing far fewer xruns still with fluxbox. I see one or
> > two under Gnome. I see bunches of them under KDE. So, at least with
> > ACPI off, I think it's not so certain that ACPI is the *only* problem
> > here.
> >
>
> The LKML poster's main point was that if your laptop implements ACPI
> with SMM you WILL get xruns because SMM disables interrupts. Even
> disabling ACPI in the kernel will not eliminate the problem 100% because
> the hardware sensors will cause the machine to go to SMM mode to enable
> the fan for example.
>
> So this _is_ definitely a problem IF your laptop uses SMM to implement
> ACPI. It remains to be seen how widespread the issue is, and whether
> there is a way to tell whether a given machine has the problem other
> than trial and error.

APM is implemented with SMM. turning off ACPI will give you problems.
(we are talking about the sub milisecond latency range here)

i still see some hope to fix this stuff by disabling APM (which is done
by the ACPI code) but not using ACPI.

will experiment with this stuff during the weekend....

-- 
torben Hohn
http://galan.sourceforge.net -- The graphical Audio language


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

This archive was generated by hypermail 2b28 : Fri Oct 15 2004 - 13:31:53 EEST