On 2014-04-29 15:51, Peter Meerwald wrote: > Hello, > >>>> Okay, so second iteration. The patch is no longer a draft, and numbers >>>> look good; Peter Meerwald has confirmed my numbers of 15 - 25% less CPU in >>>> low latency scenarios (right)? > >> The ALSA thread logic should remain unchanged, regardless of protocol >> mechanism. So maybe this is mostly a measure of the general accuracy :-) > > right, I was interested in further optimization opportunities, this does > not really relate to srchannel What I'd like to do is to also set up ringbuffers directly between the client and the I/O thread - this should result in even bigger savings for the PA main thread, which would not have to wake up just to receive a message and send it further - but I haven't looked into that yet, so I don't know how difficult it would be. So that's a possible improvement, but I don't want to promise anything at this point. >> I guess a perf on the client would show bigger differences. > > will do this when benchmarking the current patches... Thanks! Btw, I think I found what caused the pactl segfault for you - I got a segfault too (at the end of "pactl list") while working on the 2nd iteration. It's fixed in this patch series. -- David Henningsson, Canonical Ltd. https://launchpad.net/~diwic
