ikawrakow/ik_llama.cpp
1
0
Fork 0
mirror of https://github.com/ikawrakow/ik_llama.cpp.git synced 2026-04-25 08:59:30 +00:00
Code Issues Packages Projects Releases Wiki Activity

Fix Zen4 implementation of iq3_k, iq4_k, iq5_k

Browse Source 
See comments in f3a823ce72
This commit is contained in:
Iwan Kawrakow
2024-08-09 10:32:07 +03:00
committed by Kawrakow
parent c77dba5273
commit f0d7a0d53b
3 changed files with 63 additions and 56 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
ggml/src/iqk/iqk_quantize.cpp
View File

@@ -1868,7 +1868,7 @@ size_t quantize_iq6_k(const float * src, void * dst, int64_t nrows, int64_t n_pe
char * qrow = (char *)dst;
float values[128];
for (int i = 0; i < 64; ++i) {
values[i] = A_IQ6K + B_IQ6K*i - C_IQ6K*i*i + D_IQ6K*i*i*i;
values[i] = iq6nl_values[i];
values[i+64] = values[i] + S_IQ6K;
}
for (int64_t row = 0; row < nrows; ++row) {