Mirrors/llama.cpp
1
0
Fork 0
mirror of https://github.com/ggerganov/llama.cpp.git synced 2024-12-27 22:59:24 +01:00
Code Issues Packages Projects Releases Wiki Activity

sycl : offload of get_rows set to 0 (#10432)

Browse Source
This commit is contained in:
Alberto Cabrera Pérez 2024-11-29 12:38:45 +00:00 committed by GitHub
parent 266b8519ee
commit 0f77aae560
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 1 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
ggml/src/ggml-sycl/ggml-sycl.cpp
View File

@ -4493,7 +4493,7 @@ static bool ggml_backend_sycl_device_supports_buft(ggml_backend_dev_t dev, ggml_
static int64_t get_op_batch_size(const ggml_tensor * op) { static int64_t get_op_batch_size(const ggml_tensor * op) {
switch (op->op) { switch (op->op) {
case GGML_OP_GET_ROWS: case GGML_OP_GET_ROWS:
return op->ne[1]; // this will increse the speed of prefill in test return 0;
case GGML_OP_MUL_MAT: case GGML_OP_MUL_MAT:
return op->ne[1]; return op->ne[1];
case GGML_OP_MUL_MAT_ID: case GGML_OP_MUL_MAT_ID: