vbatts/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

sycl : offload of get_rows set to 0

Browse source
This commit is contained in:
Alberto Cabrera 2024-11-20 13:31:27 +00:00
parent 8fd4b7fa29
commit f4c4ce3767
1 changed files with 1 additions and 1 deletions
Download patch file Download diff file Expand all files Collapse all files

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

@ -4486,7 +4486,7 @@ 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) {
switch (op->op) {
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:
return op->ne[1];
case GGML_OP_MUL_MAT_ID: