haraldwolff/llama.cpp
1
0
Fork 0
Code Issues Pull requests Packages Projects Releases Wiki Activity

Fix scalar version of Q5_K when QK_K = 64 (#2362)

Browse source 
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
This commit is contained in:
Kawrakow 2023-07-24 12:55:02 +03:00 committed by GitHub
parent 84e09a7d8b
commit 42f70cb2f6
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
1 changed files with 2 additions and 3 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

5
k_quants.c
View file

@ -3297,8 +3297,7 @@ void ggml_vec_dot_q5_K_q8_K(const int n, float * restrict s, const void * restri
#else
uint8_t aux8[QK_K];
int8_t aux8[QK_K];
int16_t aux16[16];
float sums [8];
memset(sums, 0, 8*sizeof(float));
@ -3308,7 +3307,7 @@ void ggml_vec_dot_q5_K_q8_K(const int n, float * restrict s, const void * restri
const uint8_t * restrict q4 = x[i].qs;
const uint8_t * restrict hm = x[i].qh;
const int8_t * restrict q8 = y[i].qs;
uint8_t * restrict a = aux8;
int8_t * restrict a = aux8;
for (int l = 0; l < 32; ++l) {
a[l+ 0] = q4[l] & 0xF;
a[l+32] = q4[l] >> 4;