tqcq/llama.cpp
0
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-06-27 20:05:20 +00:00
Code Issues Packages Projects Releases Wiki Activity

ggml-cpu: clarify naming of dlf16

Browse Source 
Signed-off-by: Aaron Teo <aaron.teo1@ibm.com>
This commit is contained in:
Aaron Teo
2025-06-21 23:34:42 +08:00
parent 07de57c69a
commit 489cdf44bf
1 changed files with 2 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
ggml/src/ggml-cpu/simd-mappings.h
View File

@ -965,8 +965,8 @@ static inline void __lsx_f16x4_store(ggml_fp16_t * x, __m128 y) {
static inline float32x4_t __lzs_f16cx4_load(const ggml_fp16_t * x) { static inline float32x4_t __lzs_f16cx4_load(const ggml_fp16_t * x) {
#if defined(__NNPA__) #if defined(__NNPA__)
uint16x8_t v_x = vec_xl(0, (const ggml_fp16_t *)x); uint16x8_t v_x = vec_xl(0, (const ggml_fp16_t *)x);
uint16x8_t nnpa_dlf16 = vec_convert_from_fp16(v_x, 0); uint16x8_t v_xd = vec_convert_from_fp16(v_x, 0);
return vec_extend_to_fp32_hi(nnpa_dlf16, 0); return vec_extend_to_fp32_hi(v_xd, 0);
#else #else
float tmp[4]; float tmp[4];