Re: 2.5.74-mm1

On Fri, 11 Jul 2003, Daniel Phillips wrote:

I suspect you are right.  I'd also like to note that this is ground so
thoroughly trodden that the grass is flat.  Realtime schedulers are a well
researched topic, it's just too bad that committees don't design them as well
as engineers would.

Thinking strictly about the needs of sound processing, what's needed is a
guarantee of so much cpu time each time the timer fires, and a user limit to
prevent cpu hogging.  It's worth pondering the difference between that and
rate-of-forward-progress.  I suspect some simple improvements to the current
scheduler can be made to do the job, and at the same time, avoid the
priorty-based starvation issue that seems to have been practically mandated
by POSIX.

I've been finally able to make my sound card to sing with 2.5 and I was
able to sh*t load my machine running RealPlay with the SOFTRR path :

http://www.xmailserver.org/linux-patches/softrr.html

I was not able to get a single skip. For many kind of applications it is
not strong real time that is needed. For example, in case on those
multimedia pps, I saw that they can live pretty happy with 10-20ms
latencies. The problem is that w/out living in the realtime priority even,
they can be sucked in by interactive tasks running they loong timeslice
multiple times. Plus-latencies of 100-150ms are very easy to get. Even if
the average latency, like graphs show, is very close to the expected one.



- Davide

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"aart@kvack.org"> aart@kvack.org </a>