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

vulkan: optimize Q3_K by removing branches

Browse source
This commit is contained in:
Rémy O 2025-01-19 11:52:19 +01:00
parent 055e4287f7
commit 5fa89da950
1 changed files with 2 additions and 4 deletions
Download patch file Download diff file Expand all files Collapse all files

6
ggml/src/ggml-vulkan/vulkan-shaders/mul_mm.comp
View file

@ -343,10 +343,8 @@ void main() {
const uint qsshift = halfsplit * 2; // 0,2,4,6 const uint qsshift = halfsplit * 2; // 0,2,4,6
const uint m = 1 << (4 * n + halfsplit); // 1,2,4,8,16,32,64,128 const uint m = 1 << (4 * n + halfsplit); // 1,2,4,8,16,32,64,128
const int8_t us = int8_t(is < 4 ? (data_a[ib].scales[is-0] & 0xF) | (((data_a[ib].scales[is+8] >> 0) & 3) << 4) : const int8_t us = int8_t(((data_a[ib].scales[is % 8] >> (4 * int(is / 8))) & 0xF)
is < 8 ? (data_a[ib].scales[is-0] & 0xF) | (((data_a[ib].scales[is+4] >> 2) & 3) << 4) : | (((data_a[ib].scales[8 + (is % 4)] >> (2 * int(is / 4))) & 3) << 4));
is < 12 ? (data_a[ib].scales[is-8] >> 4) | (((data_a[ib].scales[is+0] >> 4) & 3) << 4) :
(data_a[ib].scales[is-8] >> 4) | (((data_a[ib].scales[is-4] >> 6) & 3) << 4));
const float dl = float(data_a[ib].d) * float(us - 32); const float dl = float(data_a[ib].d) * float(us - 32);
buf_a[buf_idx ] = FLOAT_TYPE(dl * float(int8_t((data_a[ib].qs[qsi ] >> qsshift) & 3) - (((data_a[ib].hmask[hmi ] & m) != 0) ? 0 : 4))); buf_a[buf_idx ] = FLOAT_TYPE(dl * float(int8_t((data_a[ib].qs[qsi ] >> qsshift) & 3) - (((data_a[ib].hmask[hmi ] & m) != 0) ? 0 : 4)));