floraachy
/
OpenBLAS
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Add openblas_set_num_threads_local() and use of blas_omp_threads_local in OMP parallel regions

This commit is contained in:
Martin Kroeker 2024-01-13 20:02:43 +01:00 committed by GitHub
parent f9b2d7f225
commit 2e2e538b7c
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
2 changed files with 5 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1
cblas.h
View File

@ -12,6 +12,7 @@ extern "C" {
/*Set the number of threads on runtime.*/
void openblas_set_num_threads(int num_threads);
void goto_set_num_threads(int num_threads);
int openblas_set_num_threads_local(int num_threads);
/*Get the number of threads on runtime.*/
int openblas_get_num_threads(void);

7
common_thread.h
View File

@ -137,19 +137,20 @@ typedef struct blas_queue {
extern int blas_server_avail;
extern int blas_omp_number_max;
extern int blas_omp_threads_local;
static __inline int num_cpu_avail(int level) {
#ifdef USE_OPENMP
int openmp_nthreads;
openmp_nthreads=omp_get_max_threads();
if (omp_in_parallel()) openmp_nthreads = blas_omp_threads_local;
#endif
#ifndef USE_OPENMP
if (blas_cpu_number == 1
#endif
#ifdef USE_OPENMP
if (openmp_nthreads == 1 || omp_in_parallel()
#else
if (openmp_nthreads == 1
#endif
) return 1;