From: torbenh <torbenh(a)gmx.de>
Subject: Re: [LAD] Kernel Opinions: .31-rt or .33-rt?
To: linux-audio-dev(a)lists.linuxaudio.org
Date: Thursday, April 29, 2010, 3:41 PM
On Thu, Apr 29, 2010 at 09:23:37PM
+0200, Ralf Mardorf wrote:
Gabriel M. Beddingfield wrote:
> On Thu, 29 Apr 2010, Ralf Mardorf wrote:
>
>
>> using, might have an advantage regarding to
>>
heat. I keep my
>> fingers crossed that less activity for
you
>>
fan, won't shutdown
>> your laptop, when doing audio work.
>>
> Well, I've used it for several kernel builds with
>
-j2. Gets hot,
but
hasn't overheated... so I think I'm good. :-)
-gabriel
This is a good test :), when I do build a kernel the
CPU is working
to capacity for around 1 hour, when set to
MAKEFLAGS="CONCURRENCY-LEVEL=2", quasi -j2. I guess
your mileage
won't vary too much, so several kernel builds
should
be harder than
audio work.
1 hour ... with both cores ?
thats why i dont use these weird kernel buildsystems.
i build a kernel in 10minutes on one core.
simply because i turn the stuff i dont need off...
but this time i missed some important option and it took
pretty long to
get the new kernel working...
for the record:
this was the solution.
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/571164
I tried 2.6.33.x-rty with a config from 2.6.32. Did the usual make-kpkg thing with the
env variable CONCURRENCY_LEVEL=2 set prior to make-kpkg (dual core system). It built fine
in a reasonable time (my kernel config is home-made, a lot of stuff is not checked in).
But when booting, I got a kernel panic (first time in years!). So back to vanilla 2.6.32
which works great for me even for RT (see xruns only when unstable apps are sometimes
crashing - I use jack2 by the way).
J.