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

QK_K = 64 tests pass on ARM_NEON and Metal

Browse source 
Sadly, that does not mean it actually works.
This commit is contained in:
Iwan Kawrakow 2024-02-27 20:12:54 +02:00
parent 28e6146c11
commit de64e061da
2 changed files with 31 additions and 29 deletions
Download patch file Download diff file Expand all files Collapse all files

2
ggml-quants.c
View file

@ -10262,7 +10262,7 @@ void ggml_vec_dot_iq1_s_q8_K (int n, float * GGML_RESTRICT s, size_t bs, const
const int nb = n / QK_K;
#if defined __ARM_NEON
#if defined __ARM_NEON && QK_K != 64
const uint8x16_t m8 = vdupq_n_u8(0x08);
const uint8x16_t m7 = vdupq_n_u8(0x07);