OpenMP-based parallel processing support
Brought to you by:
yeehector
Menu
▾
▴
#9 OpenMP-based parallel processing support
Here is a patch that enables parallel processing via OpenMP. I parallelized two small sections of code. Most of the speedup comes from parallelizing the convolution algorithm. On a 4 core OS X machine, this patch provides a speedup of about 1.5.
The patch by default disables OpenMP. It can be enabled via CMake option.
I've uploaded a new version of openmp.patch that also parallelizes the testing section. This results in a better overall speedup of about 2 (instead of the 1.5 mentioned previously).