Dynamic locking in Pthread Backend to allow multiple BLAS calls to be executed parallelly

This commit is contained in:
shivammonaka 2024-06-07 08:40:17 +05:30
parent f955616f98
commit 9e22d70957
1 changed files with 12 additions and 0 deletions

View File

@ -570,6 +570,8 @@ static int gemm_driver(blas_arg_t *args, BLASLONG *range_m, BLASLONG
InitializeCriticalSection((PCRITICAL_SECTION)&level3_lock);
#else
static pthread_mutex_t level3_lock = PTHREAD_MUTEX_INITIALIZER;
static pthread_cond_t level3_wakeup = PTHREAD_COND_INITIALIZER;
volatile static BLASLONG CPU_AVAILABLE = MAX_CPU_NUMBER;
#endif
blas_arg_t newarg;
@ -639,6 +641,12 @@ static int gemm_driver(blas_arg_t *args, BLASLONG *range_m, BLASLONG
EnterCriticalSection((PCRITICAL_SECTION)&level3_lock);
#else
pthread_mutex_lock(&level3_lock);
while(CPU_AVAILABLE < nthreads) {
pthread_cond_wait(&level3_wakeup, &level3_lock);
}
CPU_AVAILABLE -= nthreads;
WMB;
pthread_mutex_unlock(&level3_lock);
#endif
#ifdef USE_ALLOC_HEAP
@ -783,6 +791,10 @@ static int gemm_driver(blas_arg_t *args, BLASLONG *range_m, BLASLONG
#elif defined(OS_WINDOWS)
LeaveCriticalSection((PCRITICAL_SECTION)&level3_lock);
#else
pthread_mutex_lock(&level3_lock);
CPU_AVAILABLE += nthreads;
WMB;
pthread_cond_signal(&level3_wakeup);
pthread_mutex_unlock(&level3_lock);
#endif