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

Use more accurate function names

Browse source
This commit is contained in:
Slaren 2023-03-28 14:29:09 +02:00
parent ab6ac3d335
commit 911782cfdd
1 changed files with 4 additions and 5 deletions
Download patch file Download diff file Expand all files Collapse all files

9
ggml.c
View file

@ -1132,7 +1132,7 @@ void dequantize_row_q4_1(const void * restrict x, float * restrict y, int k) {
#define GGML_F32Cx8_LOAD(x) _mm256_cvtph_ps(_mm_loadu_si128((__m128i *)(x)))
#define GGML_F32Cx8_STORE(x, y) _mm_storeu_si128((__m128i *)(x), _mm256_cvtps_ph(y, 0))
#else
static inline __m256 __sse_f16x8_load(ggml_fp16_t *x) {
static inline __m256 __avx_f32cx8_load(ggml_fp16_t *x) {
float tmp[8];
for (int i = 0; i < 8; i++)
@ -1140,8 +1140,7 @@ static inline __m256 __sse_f16x8_load(ggml_fp16_t *x) {
return _mm256_loadu_ps(tmp);
}
static inline void __sse_f16x8_store(ggml_fp16_t *x, __m256 y) {
static inline void __avx_f32cx8_store(ggml_fp16_t *x, __m256 y) {
float arr[8];
_mm256_storeu_ps(arr, y);
@ -1149,8 +1148,8 @@ static inline void __sse_f16x8_store(ggml_fp16_t *x, __m256 y) {
for (int i = 0; i < 8; i++)
x[i] = GGML_FP16_TO_FP32(arr[i]);
}
#define GGML_F32Cx8_LOAD(x) __sse_f16x8_load(x)
#define GGML_F32Cx8_STORE(x, y) __sse_f16x8_store(x, y)
#define GGML_F32Cx8_LOAD(x) __avx_f32cx8_load(x)
#define GGML_F32Cx8_STORE(x, y) __avx_f32cx8_store(x, y)
#endif
#define GGML_F32Cx8_FMA GGML_F32x8_FMA