On Fri, Feb 10, 2017 at 02:06:55PM -0800, Yuri wrote:
> 0.12% doesn't sound like much, but this consistently brings it up in 'top'.
So what ?
> If jackd doesn't have any connections, why does it wake up and do something?
> It should be purely event driven, and the idle CPU use should be
> disappearingly low. I think it does something redundant.
Jackd is event driven. And even when there are no connections, it
still has to write/read samples to/from the soundcard. Not doing
that would result in a error condition that would take much more
CPU to handle.
Unless you propose that jackd should only attach to the soundcard
when the first connection is made and release it when the last
client disconnects. That would be a very bad idea for a variety
> As a comparison, mpv player playing an audio mpeg stream consumes
> only 1.2% CPU.
If you are concerned about efficiency then you are focussing on
the wrong targets. Take an audio app that consumes 5% CPU. It
can probably be optimised a bit to use only 4.5% CPU. That means
it's wasting four times as much CPU as Jack's idle load.
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)