tqcq/llama.cpp
0
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-06-27 12:05:03 +00:00
Code Issues Packages Projects Releases Wiki Activity

metal : reduce register pressure

Browse Source 
ggml-ci
This commit is contained in:
Georgi Gerganov
2025-03-26 21:24:28 +02:00
parent c6a1be6c0b
commit 70b063a550
1 changed files with 9 additions and 9 deletions
Download Patch File Download Diff File Expand all files Collapse all files

18
ggml/src/ggml-metal/ggml-metal.metal
View File

@ -4311,7 +4311,7 @@ kernel void kernel_cpy_f32_iq4_nl(
float amax = 0.0f; // absolute max float amax = 0.0f; // absolute max
float max = 0.0f; float max = 0.0f;
for (int j = 0; j < QK4_0; j++) { for (int j = 0; j < QK4_NL; j++) {
const float v = src[j]; const float v = src[j];
if (amax < fabs(v)) { if (amax < fabs(v)) {
amax = fabs(v); amax = fabs(v);
@ -4555,12 +4555,12 @@ void kernel_mul_mv_q3_K_f32_impl(
//const uint16_t kmask1 = 0x3030; //const uint16_t kmask1 = 0x3030;
//const uint16_t kmask2 = 0x0f0f; //const uint16_t kmask2 = 0x0f0f;
const int tid = tiisg/4; const short tid = tiisg/4;
const int ix = tiisg%4; const short ix = tiisg%4;
const int ip = tid/4; // 0 or 1 const short ip = tid/4; // 0 or 1
const int il = 2*((tid%4)/2); // 0 or 2 const short il = 2*((tid%4)/2); // 0 or 2
const int ir = tid%2; const short ir = tid%2;
const int l0 = 8*ir; const short l0 = 8*ir;
// One would think that the Metal compiler would figure out that ip and il can only have // One would think that the Metal compiler would figure out that ip and il can only have
// 4 possible states, and optimize accordingly. Well, no. It needs help, and we do it // 4 possible states, and optimize accordingly. Well, no. It needs help, and we do it
@ -4585,8 +4585,8 @@ void kernel_mul_mv_q3_K_f32_impl(
const uint16_t s_shift1 = 4*ip; const uint16_t s_shift1 = 4*ip;
const uint16_t s_shift2 = s_shift1 + il; const uint16_t s_shift2 = s_shift1 + il;
const int q_offset = 32*ip + l0; const short q_offset = 32*ip + l0;
const int y_offset = 128*ip + 32*il + l0; const short y_offset = 128*ip + 32*il + l0;
device const float * y1 = yy + ix*QK_K + y_offset; device const float * y1 = yy + ix*QK_K + y_offset;