[pulseaudio-tickets] [Bug 84585] Issue of CPU usage with HiFiBerry driver snd_soc_hifiberry_amp with Debian Wheezy kernel 3.12.28

[bugzilla-daemon at freedesktop.org]

https://bugs.freedesktop.org/show_bug.cgi?id=84585

--- Comment #8 from Tanu Kaskinen <tanuk at iki.fi> ---
One important difference is that the configured sink latency is much lower with
3.12.25 than with 3.12.28. I don't know how the kernel could affect that,
though, except by varying the maximum sink buffer size, but in this case it
seems that with both kernels the sink buffer size is the same.

The configured sink latency is determined by what applications request. In this
case there are shairport and LightSone connected to the sink (LightSone is
actually connected to the monitor source of the sink, but if LightSone requests
low latency from the source, that request will propagate to the monitored sink
also). If shairport or LightSone have different latency configuration between
the kernels, that would explain the difference in the configured sink latency.
(The configured latency of the monitor source is the same in both cases,
though, which suggests that LightSone is not to blame here). To get information
about what latencies the applications request, you could attach (as an
attachment, not in the comment!) the verbose log of pulseaudio in both cases.
See https://wiki.ubuntu.com/PulseAudio/Log for instructions.

It's unexpected that lower latency leads to lower cpu use. Is it possible that
the cpu runs at a lower speed with 3.12.28, causing higher figures in top? Here
is some information about cpu speed scaling on Raspberry Pi:
http://with-raspberrypi.blogspot.com/2014/03/cpu-frequency.html

-- 
You are receiving this mail because:
You are the QA Contact for the bug.
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.freedesktop.org/archives/pulseaudio-bugs/attachments/20141006/d75c0175/attachment.html>