tqcq/llama.cpp
0
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-07-27 03:33:46 -04:00
Code Issues Packages Projects Releases Wiki Activity

ggml-cpu: fp16 load ensured to hit

Browse Source 
Signed-off-by: Aaron Teo <aaron.teo1@ibm.com>
This commit is contained in:
Aaron Teo
2025-06-21 15:00:46 +08:00
parent 8f3a5af6c0
commit 575ea9f6c6
1 changed files with 0 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
ggml/src/ggml-cpu/simd-mappings.h
View File

@ -967,7 +967,6 @@ static inline float32x4_t __lzs_f16cx4_load(const ggml_fp16_t * x) {
#ifdef __NNPA__ #ifdef __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 nnpa_dlf16 = vec_convert_from_fp16(v_x, 0);
raise(SIGINT); // TODO: Ensure it is called
return vec_extend_to_fp32_hi(nnpa_dlf16, 0); return vec_extend_to_fp32_hi(nnpa_dlf16, 0);
#else #else
float tmp[4]; float tmp[4];