mirror of
https://github.com/ggerganov/whisper.cpp.git
synced 2025-01-30 17:58:40 +01:00
ggml : add ggml-aarch64 (ggml/0)
This commit is contained in:
parent
dd916a2852
commit
525f190917
2193
ggml/src/ggml-aarch64.c
Normal file
2193
ggml/src/ggml-aarch64.c
Normal file
File diff suppressed because it is too large
Load Diff
39
ggml/src/ggml-aarch64.h
Normal file
39
ggml/src/ggml-aarch64.h
Normal file
@ -0,0 +1,39 @@
|
||||
// SPDX-FileCopyrightText: Copyright 2024 Arm Ltd.
|
||||
#pragma once
|
||||
|
||||
#define GGML_COMMON_DECL_C
|
||||
#include "ggml-common.h"
|
||||
|
||||
#include "ggml.h"
|
||||
|
||||
// GGML internal header
|
||||
|
||||
#ifdef __cplusplus
|
||||
extern "C" {
|
||||
#endif
|
||||
|
||||
// Quantization
|
||||
void quantize_q8_0_4x4(const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t k);
|
||||
void quantize_q8_0_4x8(const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t k);
|
||||
|
||||
void quantize_mat_q8_0(const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t nrows, int64_t n_per_row, int64_t blck_size_interleave);
|
||||
|
||||
// Quantization utilizing an importance matrix (a.k.a. "Activation aWare Quantization")
|
||||
size_t quantize_q4_0_4x4(const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix);
|
||||
size_t quantize_q4_0_4x8(const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix);
|
||||
size_t quantize_q4_0_8x8(const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix);
|
||||
|
||||
// GEMV
|
||||
void ggml_gemv_q4_0_4x4_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_gemv_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_gemv_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);
|
||||
|
||||
// GEMM
|
||||
void ggml_gemm_q4_0_4x4_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_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);
|
||||
|
||||
#ifdef __cplusplus
|
||||
}
|
||||
#endif
|
||||
|
Loading…
Reference in New Issue
Block a user