Subject: Re: [linux-audio-dev] user lowlatency kernel experience
From: Paul Davis (pbd_AT_Op.Net)
Date: Wed Jun 13 2001 - 19:56:58 EEST
>so, since people here seem to think that the low-latency kernel is
>really great, i suspect i have missed something. is there a way to get
>processes to get sceduled promptly, but at the same time not hog the cpu
>if they try?
welcome to the complex world of RT-scheduled cooperating apps.
the short answer is no.
the long answer is that the apps have to be written to cooperate, and
that since they are not written to cooperate, they need to be run by a
scheduler that understands what to do. this has a lot to do with the
LAAGA discussion now under way on this list.
if the linux kernel scheduler had way of telling csound "generate
1.3msecs worth of audio data" rather than just telling it "you can run
until i take the processor away", this would all work just fine. as it
is, csound thinks it can run for as long as it can without blocking,
which is often until the kernel scheduler forcefully takes the
processor back. since this is 100ms (maximum), this does lead to the
kind of behaviour you want to see.
so, you either need to change csound and/or timidity, or run them both
from a scheduler that understands the issues. LAAGA would provide such
a scheduler, though the apps would still need to be modified (and in
csound's case, as us csound "fans" know, this is not trivial to do; it
is, however, necessary)
--p
This archive was generated by hypermail 2b28 : Wed Jun 13 2001 - 21:36:46 EEST