Bug#547207: ogg123: widely different CPU usage between .. ( different systems? )

Florian Weimer fw at deneb.enyo.de
Sun Jan 29 19:43:08 UTC 2012


* Maximiliano Curia:

> Hola Florian Weimer!
>
> El 29/01/2012 a las 19:53 escribiste:
>> >> Cpu0  : 19.6%us, 27.2%sy,  0.0%ni, 53.2%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
>> > Which tool are you using to gather this data for a particular process?
>> This is printed by "top".
>
> But that's a summary for the whole system usage. We need to test it
> with tools that only measure the usage of ogg123.

Not really, the system is otherwise idle.

> That's what it seems. Anyway, I would like to know if the kernel
> change produces a factor of 2 improvement in your environment as
> well. There's also a factor of 4 between the i5 with sid in amd64
> and the core 2 duo in squeeze in i686, which would be interesting to
> know if it's due to a particular package or to an architectural
> difference.

I've looked at this with GDB and it appears to be an ALSA issue.
libao's ALSA plugin calls into the ALSA library, and that call blocks
properly (i.e., it doesn't return before playing a few dozen kilobytes
of audio data).

Do you know how to tell libao not to use the mmap-based ALSA
interface?

You're right that it's probably hardware-dependent.  I see this with
an USB audio device.





More information about the pkg-xiph-maint mailing list