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

also applied to ggml_vec_dot_q4_1_q8_1

Browse source
This commit is contained in:
Ilya Kurdyukov 2023-05-16 21:17:49 +07:00 • committed by GitHub
parent c68faa991b
commit c1b3a8e24d
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
1 changed files with 1 additions and 1 deletions
Download patch file Download diff file Expand all files Collapse all files

2
ggml.c
View file

@ -2449,7 +2449,7 @@ static void ggml_vec_dot_q4_1_q8_1(const int n, float * restrict s, const void *
const __m256i bx = bytes_from_nibbles_32(x[i].qs);
const __m256i by = _mm256_loadu_si256( (const __m256i *)y[i].qs );
const __m256 xy = mul_sum_i8_pairs_float(bx, by);
const __m256 xy = mul_sum_us8_pairs_float(bx, by);
// Accumulate d0*d1*x*y
#if defined(__AVX2__)