tqcq/llama.cpp
0
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-07-27 11:43:35 -04:00
Code Issues Packages Projects Releases Wiki Activity

ggml-cpu: reattempt fp32->fp16

Browse Source 
Signed-off-by: Aaron Teo <aaron.teo1@ibm.com>
This commit is contained in:
Aaron Teo
2025-06-21 17:12:22 +08:00
parent 946c78ebde
commit 433d587426
1 changed files with 6 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
ggml/src/ggml-cpu/ggml-cpu.c
View File

@ -3141,6 +3141,12 @@ void ggml_cpu_fp32_to_fp16(const float * x, ggml_fp16_t * y, int64_t n) {
__m128i y_vec = _mm_cvtps_ph(x_vec, _MM_FROUND_TO_NEAREST_INT);
_mm_storel_epi64((__m128i *)(y + i), y_vec);
}
#elif defined(__NNPA__)
for (; i + 3 < n; i += 4) {
float32x4_t v_x = vec_xl(0, (const float *)(x + i));
uint16x8_t v_xd = vec_convert_to_fp16(v_xh, 0);
vec_xst(v_xd, 0, (ggml_fp16_t *)(y + i));
}
#endif
for (; i < n; ++i) {
y[i] = GGML_FP32_TO_FP16(x[i]);