tqcq/llama.cpp
0
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-07-28 03:55:06 -04:00
Code Issues Packages Projects Releases Wiki Activity

ggml-cpu: activate nnpa for ggml_cpu_fp32_to_fp16

Browse Source 
Signed-off-by: Aaron Teo <aaron.teo1@ibm.com>
This commit is contained in:
Aaron Teo
2025-06-21 16:50:05 +08:00
parent e0f8fb930b
commit bb9345ca8a
1 changed files with 10 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

10
ggml/src/ggml-cpu/ggml-cpu.c
View File

@ -3142,6 +3142,16 @@ void ggml_cpu_fp32_to_fp16(const float * x, ggml_fp16_t * y, int64_t n) {
_mm_storel_epi64((__m128i *)(y + i), y_vec); _mm_storel_epi64((__m128i *)(y + i), y_vec);
} }
#endif #endif
#if defined(__NNPA__)
for (; i + 7 < n; i += 8) {
float32x4_t v_xh = vec_xl(0, (const float *)(x + i + 0));
float32x4_t v_xl = vec_xl(0, (const float *)(x + i + 4));
uint16x8_t v_xd = vec_round_from_fp32(v_xh, v_xl, 0);
vec_xst(v_xd, 0, (ggml_fp16_t *)(y + i));
}
#endif
for (; i < n; ++i) { for (; i < n; ++i) {
y[i] = GGML_FP32_TO_FP16(x[i]); y[i] = GGML_FP32_TO_FP16(x[i]);
} }