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

Speedup the AVX-512 implementation of ggml_vec_dot_q4_0() (#933)

Browse source
This commit is contained in:
Ivan Komarov 2023-04-17 15:10:57 +02:00 • committed by GitHub
parent 47f61aaa5f
commit f266259ad9
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
4 changed files with 235 additions and 44 deletions
Download patch file Download diff file Expand all files Collapse all files

2
ggml.h
View file

@ -808,6 +808,8 @@ size_t ggml_quantize_q4_1(const float * src, void * dst, int n, int k, int64_t *
int ggml_cpu_has_avx(void);
int ggml_cpu_has_avx2(void);
int ggml_cpu_has_avx512(void);
int ggml_cpu_has_avx512_vbmi(void);
int ggml_cpu_has_avx512_vnni(void);
int ggml_cpu_has_fma(void);
int ggml_cpu_has_neon(void);
int ggml_cpu_has_arm_fma(void);