eeva/jamulus
1
0
Fork 0
Code Issues Pull requests Releases Wiki Activity

added some more info regarding the OMP implementation

Browse source
This commit is contained in:
Volker Fischer 2020-06-21 07:52:59 +02:00
parent e705dfaed9
commit e9aef3474c
1 changed files with 2 additions and 0 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

2
src/server.cpp
View file

@ -1001,6 +1001,8 @@ static CTimingMeas JitterMeas ( 1000, "test2.dat" ); JitterMeas.Measure(); // TE
// have an issue using OMP in the OnTimer() function. Even if #pragma omp parallel for is used on a trivial // have an issue using OMP in the OnTimer() function. Even if #pragma omp parallel for is used on a trivial
// for loop for testing, still the CPU usage goes to very high values -> What is the cause of this issue? // for loop for testing, still the CPU usage goes to very high values -> What is the cause of this issue?
// NOTE Most probably it is the overhead of threads creation/destruction which causes this effect. // NOTE Most probably it is the overhead of threads creation/destruction which causes this effect.
// See https://software.intel.com/content/www/us/en/develop/articles/performance-obstacles-for-threading-how-do-they-affect-openmp-code.html
// "[...] overhead numbers are high enough that it doesn’t make sense to thread that code. In those cases, we’re better off leaving the code in its original serial form."
# pragma omp parallel for # pragma omp parallel for
#endif #endif
for ( int i = 0; i < iNumClients; i++ ) for ( int i = 0; i < iNumClients; i++ )