|
[Sponsors] |
July 12, 2013, 07:23 |
openmp with single processor
|
#1 |
Member
Ravindra Shende
Join Date: Feb 2011
Location: Pune, India
Posts: 45
Rep Power: 15 |
Hi everyone,
I found that my openmp parallelized fortran code runs slower when run on one processor as compared to the same but unparallelized code. (I have used ifort compiler) Can someone tell me why is this so? Is it because the compiler itself tries to parallelise a serial code and when the code is parallel it doesn't? |
|
July 15, 2013, 11:50 |
|
#2 |
Super Moderator
Alex
Join Date: Jun 2012
Location: Germany
Posts: 3,427
Rep Power: 49 |
How high is the slowdown you observe for the parallel code?
A few percent could be explained by the overhead introduced by the parallelization. |
|
July 15, 2013, 11:51 |
|
#3 |
Senior Member
Filippo Maria Denaro
Join Date: Jul 2010
Posts: 6,896
Rep Power: 73 |
how do you set the OS parameters NCPUS and OMP_NUM_THREADS?
|
|
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
AMI speed performance | danny123 | OpenFOAM | 21 | October 24, 2020 05:13 |
decomposePar pointfield | flying | OpenFOAM Running, Solving & CFD | 28 | December 30, 2013 16:05 |
AMI interDyMFoam for mixer | danny123 | OpenFOAM Running, Solving & CFD | 4 | June 19, 2013 05:49 |
DecomposePar unequal number of shared faces | maka | OpenFOAM Pre-Processing | 6 | August 12, 2010 10:01 |
Unaligned accesses on IA64 | andre | OpenFOAM | 5 | June 23, 2008 11:37 |