extern/whisper.cpp
1
0
Fork 1
mirror of https://github.com/ggerganov/whisper.cpp.git synced 2025-01-06 05:58:55 +01:00
Code Issues Packages Projects Releases Wiki Activity

ggml : when using BLAS start only 1 CPU thread

Browse Source
This commit is contained in:
Georgi Gerganov 2023-01-07 19:48:56 +02:00
parent 6394c906af
commit d347a59a5f
No known key found for this signature in database
GPG Key ID: 449E073F9DC10735
1 changed files with 1 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1
ggml.c
View File

@ -7257,6 +7257,7 @@ void ggml_graph_compute(struct ggml_context * ctx, struct ggml_cgraph * cgraph)
node->src1->type == GGML_TYPE_F32) { node->src1->type == GGML_TYPE_F32) {
#if defined(GGML_USE_ACCELERATE) || defined(GGML_USE_OPENBLAS) #if defined(GGML_USE_ACCELERATE) || defined(GGML_USE_OPENBLAS)
if (ggml_compute_forward_mul_mat_use_blas(node->src0, node->src1, node)) { if (ggml_compute_forward_mul_mat_use_blas(node->src0, node->src1, node)) {
node->n_tasks = 1;
cur = sizeof(float)*(node->src0->ne[0]*node->src0->ne[1]); cur = sizeof(float)*(node->src0->ne[0]*node->src0->ne[1]);
} else { } else {
cur = sizeof(ggml_fp16_t)*ggml_nelements(node->src1); cur = sizeof(ggml_fp16_t)*ggml_nelements(node->src1);