On Mon, Sep 30, 2013 at 10:15:24PM +1000, Patrick Shirkey wrote:
> I'm asking for additional suggestions on how to authoritatively test this
In a system such as the one you are testing latency is determined by
the amount of buffering, and nothing else. C-states shouldn't enter
in the picture as long as the CPU wakeup time is a small fraction of
a period (1.333ms in this case). If the wakeup time is too high, the
whole setup just fails. In other words, if you measure e.g. 100ms
of latency, that can only mean that somewhere along the pipeline
100ms of audio is being stored. It can't be the result of a CPU
being 100ms late - that would interrupt the signal instead.
A world of exhaustive, reliable metadata would be an utopia.
It's also a pipe-dream, founded on self-delusion, nerd hubris
and hysterically inflated market opportunities. (Cory Doctorow)
Linux-audio-dev mailing list