Intel® oneAPI Math Kernel Library Developer Reference - C
Specifies the number of OpenMP* threads to use.
void mkl_set_num_threads( int nt );
Name |
Type |
Description |
|---|---|---|
nt |
int |
nt > 0 - The number of threads suggested by the user. nt≤ 0 - Invalid value, which is ignored. |
This function enables you to specify how many OpenMP threads Intel® oneAPI Math Kernel Library should use for internal parallel regions. If this number is not set (default), Intel® oneAPI Math Kernel Library functions use the default number of threads for the OpenMP run-time library. The specified number of threads applies:
To all Intel® oneAPI Math Kernel Library functions except the functions from the domains where the number of threads is set withmkl_domain_set_num_threads
To all execution threads except the threads where the number of threads is set with mkl_set_num_threads_local
The number specified is a hint, and Intel® oneAPI Math Kernel Library may actually use a smaller number.
This function takes precedence over the MKL_NUM_THREADS environment variable.
#include "mkl.h" … mkl_set_num_threads(4); my_compute_using_mkl(); // Intel MKL uses up to 4 OpenMP threads