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

Double CLBlast speed by disabling OpenBLAS thread workaround

Browse source 
Co-authored-by: Concedo <39025047+LostRuins@users.noreply.github.com>
Co-authored-by: slaren <2141330+slaren@users.noreply.github.com>
This commit is contained in:
0cc4m 2023-04-24 15:15:23 +02:00
parent 309af7fce9
commit f469d9afa0
3 changed files with 15 additions and 2 deletions
Download patch file Download diff file Expand all files Collapse all files

12
ggml.c
View file

@ -12354,6 +12354,18 @@ int ggml_cpu_has_cublas(void) {
#endif
}
int ggml_cpu_has_clblast(void) {
#if defined(GGML_USE_CLBLAST)
return 1;
#else
return 0;
#endif
}
int ggml_cpu_has_gpublas(void) {
return ggml_cpu_has_cublas() || ggml_cpu_has_clblast();
}
int ggml_cpu_has_sse3(void) {
#if defined(__SSE3__)
return 1;

3
ggml.h
View file

@ -852,10 +852,11 @@ extern "C" {
GGML_API int ggml_cpu_has_wasm_simd (void);
GGML_API int ggml_cpu_has_blas (void);
GGML_API int ggml_cpu_has_cublas (void);
GGML_API int ggml_cpu_has_clblast (void);
GGML_API int ggml_cpu_has_gpublas (void);
GGML_API int ggml_cpu_has_sse3 (void);
GGML_API int ggml_cpu_has_vsx (void);
//
// Internal types and functions exposed for tests and benchmarks
//

2
llama.cpp
View file

@ -1076,7 +1076,7 @@ static bool llama_eval_internal(
// for big prompts, if BLAS is enabled, it is better to use only one thread
// otherwise, the threads are spin-lock waiting for the BLAS calls and are degrading the performance
ggml_cgraph gf = {};
gf.n_threads = N >= 32 && ggml_cpu_has_blas() && !ggml_cpu_has_cublas() ? 1 : n_threads;
gf.n_threads = N >= 32 && ggml_cpu_has_blas() && !ggml_cpu_has_gpublas() ? 1 : n_threads;
struct ggml_tensor * embd = ggml_new_tensor_1d(ctx0, GGML_TYPE_I32, N);
memcpy(embd->data, tokens, N*ggml_element_size(embd));