happyz
/
llama.cpp
mirror of https://github.com/ggerganov/llama.cpp.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

change quant.c

This commit is contained in:
vithulep 2025-07-04 11:16:32 +05:30
parent 1d47d63d40
commit a318c587cd
1 changed files with 2 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

3
ggml/src/ggml-cpu/arch/arm/quants.c
View File

@ -45,7 +45,7 @@ void quantize_row_q8_0(const float * GGML_RESTRICT x, void * GGML_RESTRICT vy, i
block_q8_0 * GGML_RESTRICT y = vy;
#if defined(__ARM_FEATURE_SVE)
const int sve_register_length = ggml_cpu_get_sve_cnt() * 8;
const int sve_register_length = svcntb() * 8;
const int ggml_f32_epr = sve_register_length / 32;
const svfloat32_t inactive1 = svdup_n_f32(0.0f);
const svbool_t pg = svptrue_b32();
@ -79,6 +79,7 @@ void quantize_row_q8_0(const float * GGML_RESTRICT x, void * GGML_RESTRICT vy, i
svst1b_s32(pg, &y[i].qs[j], result);
}
}
#elif defined(__ARM_NEON)
for (int i = 0; i < nb; i++) {
float32x4_t srcv [8];