Skip to content

Manual unrolling of loop for vec_dot_q8_q8 #7

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
wants to merge 1 commit into
base: master
Choose a base branch
from
Open

Manual unrolling of loop for vec_dot_q8_q8 #7

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
90 changes: 86 additions & 4 deletions ggml/src/ggml-cpu/ggml-cpu-quants.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3922,10 +3922,92 @@ void ggml_vec_dot_q8_0_q8_0(int n, float * GGML_RESTRICT s, size_t bs, const voi
}
#elif defined(__POWER9_VECTOR__)
const vector signed int v0 = vec_splats((int32_t)0);
vector float vsumf0 = vec_splats(0.0f);
//vector float vsumf0 = vec_splats(0.0f);

#pragma GCC unroll 8
for (; ib < nb; ++ib) {
//#pragma GCC unroll 8
const int unroll_factor = 16;
vector float vsumf[unroll_factor];
for (int i = 0; i < unroll_factor; ++i) {
vsumf[i] = vec_splats(0.0f);
}

for (; ib <= nb - unroll_factor; ib += unroll_factor) {
#pragma GCC unroll 1 // disable inner unroll, already unrolled manually

for (int u = 0; u < unroll_factor; ++u) {
__builtin_prefetch(x[ib+u].qs, 0, 1);
__builtin_prefetch(y[ib+u].qs, 0, 1);

vector float vxd = vec_splats(GGML_FP16_TO_FP32(x[ib+u].d));
vector float vyd = vec_splats(GGML_FP16_TO_FP32(y[ib+u].d));
vector float vd = vec_mul(vxd, vyd);

vector signed char q8x0 = vec_xl( 0, x[ib+u].qs);
vector signed char q8x1 = vec_xl(16, x[ib+u].qs);
vector signed char q8y0 = vec_xl( 0, y[ib+u].qs);
vector signed char q8y1 = vec_xl(16, y[ib+u].qs);

vector signed short qv0 = vec_mule(q8x0, q8y0);
vector signed short qv1 = vec_mulo(q8x0, q8y0);
vector signed short qv2 = vec_mule(q8x1, q8y1);
vector signed short qv3 = vec_mulo(q8x1, q8y1);

vector signed int vsumi0 = v0;
vector signed int vsumi1 = v0;

vsumi0 = vec_sum4s(qv0, vsumi0);
vsumi1 = vec_sum4s(qv1, vsumi1);
vsumi0 = vec_sum4s(qv2, vsumi0);
vsumi1 = vec_sum4s(qv3, vsumi1);

vsumi0 = vec_add(vsumi0, vsumi1);

vsumf[u] = vec_madd(vec_ctf(vsumi0, 0), vd, vsumf[u]);
}
}

// Handle leftover elements
for (; ib < nb; ++ib) {
__builtin_prefetch(x[ib].qs, 0, 1);
__builtin_prefetch(y[ib].qs, 0, 1);

vector float vxd = vec_splats(GGML_FP16_TO_FP32(x[ib].d));
vector float vyd = vec_splats(GGML_FP16_TO_FP32(y[ib].d));
vector float vd = vec_mul(vxd, vyd);

vector signed char q8x0 = vec_xl( 0, x[ib].qs);
vector signed char q8x1 = vec_xl(16, x[ib].qs);
vector signed char q8y0 = vec_xl( 0, y[ib].qs);
vector signed char q8y1 = vec_xl(16, y[ib].qs);

vector signed short qv0 = vec_mule(q8x0, q8y0);
vector signed short qv1 = vec_mulo(q8x0, q8y0);
vector signed short qv2 = vec_mule(q8x1, q8y1);
vector signed short qv3 = vec_mulo(q8x1, q8y1);

vector signed int vsumi0 = v0;
vector signed int vsumi1 = v0;

vsumi0 = vec_sum4s(qv0, vsumi0);
vsumi1 = vec_sum4s(qv1, vsumi1);
vsumi0 = vec_sum4s(qv2, vsumi0);
vsumi1 = vec_sum4s(qv3, vsumi1);

vsumi0 = vec_add(vsumi0, vsumi1);

vsumf[0] = vec_madd(vec_ctf(vsumi0, 0), vd, vsumf[0]);
}

// Final reduction of partial sums
vector float vsumf0 = vsumf[0];
for (int i = 1; i < unroll_factor; ++i) {
vsumf0 = vec_add(vsumf0, vsumf[i]);
}
vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 4));
vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 8));
sumf = vec_extract(vsumf0, 0);

/*for (; ib < nb; ++ib) {
__builtin_prefetch(x[ib].qs, 0, 1);
__builtin_prefetch(y[ib].qs, 0, 1);

Expand Down Expand Up @@ -3959,7 +4041,7 @@ void ggml_vec_dot_q8_0_q8_0(int n, float * GGML_RESTRICT s, size_t bs, const voi
vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 4));
vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 8));

sumf = vec_extract(vsumf0, 0);
sumf = vec_extract(vsumf0, 0);*/

#elif defined(__loongarch_asx)
// Initialize accumulator with zeros
Expand Down
Loading