ikawrakow/ik_llama.cpp
1
0
Fork 0
mirror of https://github.com/ikawrakow/ik_llama.cpp.git synced 2026-02-24 15:14:10 +00:00
Code Issues Packages Projects Releases Wiki Activity

iq5_ks_r4: AVX2 works

Browse Source
This commit is contained in:
Iwan Kawrakow
2025-05-15 20:03:13 +03:00
parent 25ce0f0372
commit 2b040aef21
1 changed files with 1 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
ggml/src/iqk/iqk_mul_mat.cpp
View File

@@ -7509,6 +7509,7 @@ static void mul_mat_iq5_ks_r4_q8_k(int n, const void * vx, size_t bx, const Data
using helper_t = union { __m256i vec; uint32_t val[8]; };
#ifndef HAVE_FANCY_SIMD
helper_t h, h_shift;
auto s_shuffle = _mm256_set_epi64x(0x0f0e0f0e0d0c0d0c, 0x0b0a0b0a09080908, 0x0706070605040504, 0x0302030201000100);
#else
using helper512_t = union { __m512i vec; uint64_t val[8]; };
helper_t h;