block_q5_k const hoist
This commit is contained in:
parent
222c679842
commit
1556bbb6a3
1 changed files with 6 additions and 5 deletions
|
@ -1303,16 +1303,17 @@ kernel void kernel_mul_mat_q5_k_f32(
|
||||||
float sumf = 0;
|
float sumf = 0;
|
||||||
for (int i = tpitg.x; i < nb; i += tptg.x) {
|
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 * 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 * y1 = yy + i*QK_K + y_offset;
|
||||||
device const float * y2 = y1 + 128;
|
device const float * y2 = y1 + 128;
|
||||||
|
|
||||||
const float dall = (float)((x + i)->d);
|
const float dall = (float)(xi->d);
|
||||||
const float dmin = (float)((x + i)->dmin);
|
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));
|
sc1 = as_type<uchar2>((uint16_t)(a[im+0] & kmask1));
|
||||||
sc2 = as_type<uchar2>((uint16_t)(a[im+2] & 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)));
|
sc3 = as_type<uchar2>((uint16_t)(((a[im+4] >> 0) & kmask2) | ((a[im+0] & kmask3) >> 2)));
|
||||||
|
|
Loading…
Add table
Add a link
Reference in a new issue