hashtags/lmms
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases 1 Wiki Activity

* initial (not yet enabled) support for binding mixer threads to CPUs to reduce scheduling overhead and cache misses (gives avg. 10% performance boost)

Browse Source 
* merged mixer-thread with first worker-thread - no more duplicate code for single- and multicore processing and better scheduling on multicore systems



git-svn-id: https://lmms.svn.sf.net/svnroot/lmms/trunk/lmms@1803 0778d3d1-df1d-0410-868b-ea421aaaa00d
This commit is contained in:
Tobias Doerffel
2008-10-25 17:53:58 +00:00
parent 6fc2eba70b
commit 9530460fce
4 changed files with 207 additions and 200 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
include/mixer.h
View File

@@ -429,7 +429,6 @@ private:
QVector<surroundSampleFrame *> m_bufferPool;
int m_readBuffer;
int m_writeBuffer;
int m_analBuffer;
int m_poolDepth;
surroundSampleFrame m_maxClip;
@@ -439,7 +438,6 @@ private:
bool m_newBuffer[SURROUND_CHANNELS];
int m_cpuLoad;
bool m_multiThreaded;
QVector<mixerWorkerThread *> m_workers;
int m_numWorkers;
QSemaphore m_queueReadySem;