Re: [LMMS-devel] threads...

2014-04-27 Thread oeai
This is something about #628 qthreadpools http://qt-project.org/doc/qt-4.8/qthreadpool.html briefly pools will suspend unused threads and reuse them in needed time. and i think that semaphores should be implemented in patterns, then this will give exact time +/-500ms for wake or suspend thread.

Re: [LMMS-devel] threads...

2014-04-26 Thread Vesa
On 04/26/2014 03:50 PM, Jonathan Aquilina wrote: > Wouldnt something like a mixer with each channel having its own thread? AFAIK, no. Thread count should be optimized to the system, too many threads just loses performance. --

Re: [LMMS-devel] threads...

2014-04-26 Thread Tres Finocchiaro
> > I'm not very familiar with threads so I might be way off here, but seems > to me that in FxMixer, we go through all of the channels in a recursive > function just to add them on the job queue - and we do this every > period. Additionally it seems (I could be wrong as I'm not sure how the > queu

Re: [LMMS-devel] threads...

2014-04-26 Thread Jonathan Aquilina
Wouldnt something like a mixer with each channel having its own thread? On Sat, Apr 26, 2014 at 2:49 PM, Vesa wrote: > I'm a bit all over the place with these questions lately, but bear with > me... > > I'm not very familiar with threads so I might be way off here, but seems > to me that in FxM

[LMMS-devel] threads...

2014-04-26 Thread Vesa
I'm a bit all over the place with these questions lately, but bear with me... I'm not very familiar with threads so I might be way off here, but seems to me that in FxMixer, we go through all of the channels in a recursive function just to add them on the job queue - and we do this every period. A