Re: realtime-preempt-2.6.12-final-V0.7.51-11 glitches [no more]

["Rui Nuno Capela" <rncbc@rncbc.org>]

>>   ------------------------------ ------------- -------------
>>                                  RT-V0.7.51-13 RT-V0.7.49-01
>>   ------------------------------ ------------- -------------
>
>>   Delay Maximum . . . . . . . . :      333           295     usecs
>>   Cycle Maximum . . . . . . . . :      970           943     usecs
>>   Average DSP Load. . . . . . . :       45.7          44.4   %
>>   Average CPU System Load . . . :       15.6          16.3   %
>>   Average CPU User Load . . . . :       32.0          30.1   %
>
> i'm wondering - is this slight increase in CPU utilization (and
> latencies) due to natural fluctuations, or is it a genuine overhead
> increase?
>

I'm pretty confortable in saying that both results are almost equivalent,
the slight differences probably due to environmental effects. All tests
were conducted from a regular user konsole screen, under KDE 3.3 and
perfectly functional X.org desktop session. Runlevel 5 with network down
however. Oh, gkrellm is also on screen :)

OTOH, I'll take this chance to show you something that is annoying me for
quite some time. Just look to the attached chart where I've marked the
spot with an arrow and a question mark. Its just one example of a strange
behavior/phenomenon while running the jack_test4.2 test on PREEMPT_RT
kernels: the CPU usage, which stays normally around 50%, suddenly jumps to
60% steady, starting at random points in time but always some time after
the test has been started. Note that this randomness surely adds to the
the slight differences found on the above results.

First suspicions gone to cpu frequency/throttling, but I have it disabled
on every kernel I build. Denormals? Don't think so, as jack_test_client is
specially coded to treat floats as zero if below 1E-6.

Any hints?
-- 
rncbc aka Rui Nuno Capela
rncbc@rncbc.org

Attachment: jack_test4-2.6.12-RT-V0.7.51-13.0-200507080917.png
Description: PNG image