Mirrors/llama.cpp
1
0
Fork 0
mirror of https://github.com/ggerganov/llama.cpp.git synced 2024-12-26 14:20:31 +01:00
Code Issues Packages Projects Releases Wiki Activity

Update

Browse Source
This commit is contained in:
Aidan 2024-04-22 10:42:32 +01:00
parent 907df4459c
commit 124e4dced2
2 changed files with 9 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1
examples/benchmark/benchmark-matmult.cpp
View File

@ -147,7 +147,6 @@ int main(int argc, char ** argv) {
};
int main_gpu = 0;
ggml_backend_t backend = ggml_backend_sycl_init(main_gpu);
if (!ctx) {
fprintf(stderr, "%s: ggml_init() failed\n", __func__);
return 1;

11
examples/test-bench/test-bench.cpp
View File

@ -415,9 +415,16 @@ struct test_mul_mat : public test_case {
};
void bench_mul_mat(ggml_backend_t backend) {
auto test = test_mul_mat(GGML_TYPE_Q4_0, GGML_TYPE_F16, 512, 256, 1024, { 1, 1}, {1, 1});
auto test1 = test_mul_mat(GGML_TYPE_Q4_0, GGML_TYPE_F16, 11008, 1, 4096, { 1, 1}, {1, 1});
auto test2 = test_mul_mat(GGML_TYPE_Q4_0, GGML_TYPE_Q4_0, 11008, 1, 4096, { 1, 1}, {1, 1});
test.eval_perf(backend, "MUL_MAT", 8000 /*n_runs*/);
auto test3 = test_mul_mat(GGML_TYPE_Q4_0, GGML_TYPE_F16, 512, 256, 1024, { 1, 1}, {1, 1});
auto test4 = test_mul_mat(GGML_TYPE_Q4_0, GGML_TYPE_Q4_0, 512, 256, 1024, { 1, 1}, {1, 1});
test1.eval_perf(backend, "MUL_MAT", 8000 /*n_runs*/);
test2.eval_perf(backend, "MUL_MAT", 8000 /*n_runs*/);
test3.eval_perf(backend, "MUL_MAT", 8000 /*n_runs*/);
test4.eval_perf(backend, "MUL_MAT", 8000 /*n_runs*/);
}
int main() {