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

block_q5_k const hoist

Browse source
This commit is contained in:
Steven Roussey 2023-06-14 12:02:40 -07:00
parent 222c679842
commit 1556bbb6a3
1 changed files with 6 additions and 5 deletions
Download patch file Download diff file Expand all files Collapse all files

11
ggml-metal.metal
View file

@ -1303,16 +1303,17 @@ kernel void kernel_mul_mat_q5_k_f32(
float sumf = 0;
for (int i = tpitg.x; i < nb; i += tptg.x) {
device const uint8_t * q1 = (x + i)->qs + q_offset;
device const block_q5_k * xi = x + i;
device const uint8_t * q1 = xi->qs + q_offset;
device const uint8_t * q2 = q1 + 64;
device const uint8_t * qh = (x + i)->qh + l0;
device const uint8_t * qh = xi->qh + l0;
device const float * y1 = yy + i*QK_K + y_offset;
device const float * y2 = y1 + 128;
const float dall = (float)((x + i)->d);
const float dmin = (float)((x + i)->dmin);
const float dall = (float)(xi->d);
const float dmin = (float)(xi->dmin);
device const uint16_t * a = (device const uint16_t *)(x + i)->scales;
device const uint16_t * a = (device const uint16_t *)xi->scales;
sc1 = as_type<uchar2>((uint16_t)(a[im+0] & kmask1));
sc2 = as_type<uchar2>((uint16_t)(a[im+2] & kmask1));
sc3 = as_type<uchar2>((uint16_t)(((a[im+4] >> 0) & kmask2) | ((a[im+0] & kmask3) >> 2)));