holm/whisper.cpp
1
1
Fork 0
forked from extern/whisper.cpp
Code Issues Pull Requests Packages Projects Releases Wiki Activity

ggml : multi-thread the ggml_add operator

Browse Source
This commit is contained in:
Georgi Gerganov 2022-11-03 20:53:44 +02:00
parent a2eeb941f6
commit 75171c2b79
No known key found for this signature in database
GPG Key ID: 449E073F9DC10735
1 changed files with 5 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

7
ggml.c
View File

@ -3150,7 +3150,10 @@ void ggml_compute_forward_add_f32(
GGML_ASSERT(nb00 == sizeof(float)); GGML_ASSERT(nb00 == sizeof(float));
if (nb10 == sizeof(float)) { if (nb10 == sizeof(float)) {
for (int j = ith; j < n; j += nth) { const int j0 = (n/nth)*ith;
const int j1 = ith == nth - 1 ? n : (n/nth)*(ith + 1);
for (int j = j0; j < j1; j++) {
ggml_vec_add_f32(nc, ggml_vec_add_f32(nc,
(float *) ((char *) dst->data + j*nb1), (float *) ((char *) dst->data + j*nb1),
(float *) ((char *) src0->data + j*nb01), (float *) ((char *) src0->data + j*nb01),
@ -6857,7 +6860,7 @@ void ggml_graph_compute(struct ggml_context * ctx, struct ggml_cgraph * cgraph)
} break; } break;
case GGML_OP_ADD: case GGML_OP_ADD:
{ {
node->n_tasks = 1; node->n_tasks = n_threads;
} break; } break;
case GGML_OP_SUB: case GGML_OP_SUB:
case GGML_OP_MUL: case GGML_OP_MUL: