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

update according to review comments

Browse source
This commit is contained in:
Jianyu Zhang 2024-03-01 22:30:32 +08:00
parent 5db8896f25
commit e4cc412114
4 changed files with 3 additions and 8 deletions
Download patch file Download diff file Expand all files Collapse all files

4
common/common.cpp
View file

@ -641,8 +641,8 @@ bool gpt_params_parse_ex(int argc, char ** argv, gpt_params & params) {
params.split_mode = LLAMA_SPLIT_MODE_LAYER;
} else if (arg_next == "row") {
#ifdef GGML_USE_SYCL
fprintf(stderr, "warning: The split mode value:[row] is not supported by llama.cpp with SYCL. It's developing.\nExit!\n");
exit(1);
fprintf(stderr, "warning: The split mode value:[row] is not supported by llama.cpp with SYCL. It's developing.\nExit!\n");
exit(1);
#endif // GGML_USE_SYCL
params.split_mode = LLAMA_SPLIT_MODE_ROW;
} else {

2
examples/llama-bench/llama-bench.cpp
View file

@ -123,8 +123,6 @@ static std::string get_gpu_info() {
}
#endif
#ifdef GGML_USE_SYCL
int device_list[GGML_SYCL_MAX_DEVICES];
ggml_sycl_get_gpu_list(device_list, GGML_SYCL_MAX_DEVICES);
int count = ggml_backend_sycl_get_device_count();
for (int i = 0; i < count; i++) {
char buf[128];

2
examples/sycl/run-llama2.sh
View file

@ -19,7 +19,7 @@ echo "use $GGML_SYCL_DEVICE as main GPU"
#ZES_ENABLE_SYSMAN=1, Support to get free memory of GPU by sycl::aspect::ext_intel_free_memory. Recommended to use when --split-mode = layer.
#use all GPUs with same max compute units
ZES_ENABLE_SYSMAN=1 ./build/bin/main -m models/llama-2-7b.Q4_0.gguf -p "${INPUT2}" -n 400 -e -ngl 33 -s 0 -mg $GGML_SYCL_DEVICE
ZES_ENABLE_SYSMAN=1 ./build/bin/main -m models/llama-2-7b.Q4_0.gguf -p "${INPUT2}" -n 400 -e -ngl 33 -s 0
#use main GPU only
#ZES_ENABLE_SYSMAN=1 ./build/bin/main -m models/llama-2-7b.Q4_0.gguf -p "${INPUT2}" -n 400 -e -ngl 33 -s 0 -mg $GGML_SYCL_DEVICE -sm none

3
llama.cpp
View file

@ -104,9 +104,6 @@
#define LLAMA_MAX_NODES 8192
#define LLAMA_MAX_EXPERTS 8
#if (defined(GGML_USE_CUBLAS) || defined(GGML_USE_SYCL))
#define GGML_USE_CUBLAS_SYCL
#endif
//
// logging