ver4a/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

backend cpu: add online flow for aarch64 Q4_0 GEMV/GEMM kernels (#9921)

Browse source 
* backend-cpu: add online flow for aarch64 Q4_0 GEMV/GEMM kernels

---------

Co-authored-by: Diego Devesa <slarengh@gmail.com>
This commit is contained in:
Charles Xu 2024-11-15 01:28:50 +01:00 • committed by GitHub
parent ae8de6d50a
commit 1607a5e5b0
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
9 changed files with 273 additions and 22 deletions
Download patch file Download diff file Expand all files Collapse all files

3
ggml/src/ggml-cpu/ggml-cpu-aarch64.h
View file

@ -21,6 +21,9 @@ void ggml_gemm_q4_0_4x4_q8_0(int n, float * GGML_RESTRICT s, size_t bs, const vo
void ggml_gemm_q4_0_4x8_q8_0(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc);
void ggml_gemm_q4_0_8x8_q8_0(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc);
void ggml_aarch64_repack_tensor(struct ggml_tensor * cur, enum ggml_type repack_type, const void * data, size_t data_size);
enum ggml_type ggml_aarch64_get_optimal_repack_type(const struct ggml_tensor * cur);
#ifdef __cplusplus
}
#endif