Skip to content
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

F32-IGEMM AVX512 generate up to 16x64 #7493

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

F32-IGEMM AVX512 generate up to 16x64 #7493

Show file tree
Hide file tree
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
154 changes: 154 additions & 0 deletions bench/f32-gemm-minmax.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1595,6 +1595,160 @@

BENCHMARK_GEMM(f32_gemm_minmax_ukernel_16x32__avx512f_broadcast)

static void f32_gemm_minmax_ukernel_1x48__avx512f_broadcast(benchmark::State& state, const char* net) {
GEMMBenchmark(state,
xnn_f32_gemm_minmax_ukernel_1x48__avx512f_broadcast,
xnn_init_f32_minmax_scalar_params,
xnn_pack_f32_gemm_goi_w,
/*mr=*/1, /*nr=*/48, /*kr=*/1, /*sr=*/1,
benchmark::utils::CheckAVX512F);
}

BENCHMARK_GEMM(f32_gemm_minmax_ukernel_1x48__avx512f_broadcast)

static void f32_gemm_minmax_ukernel_4x48__avx512f_broadcast(benchmark::State& state, const char* net) {
GEMMBenchmark(state,
xnn_f32_gemm_minmax_ukernel_4x48__avx512f_broadcast,
xnn_init_f32_minmax_scalar_params,
xnn_pack_f32_gemm_goi_w,
/*mr=*/4, /*nr=*/48, /*kr=*/1, /*sr=*/1,
benchmark::utils::CheckAVX512F);
}

BENCHMARK_GEMM(f32_gemm_minmax_ukernel_4x48__avx512f_broadcast)

static void f32_gemm_minmax_ukernel_5x48__avx512f_broadcast(benchmark::State& state, const char* net) {
GEMMBenchmark(state,
xnn_f32_gemm_minmax_ukernel_5x48__avx512f_broadcast,
xnn_init_f32_minmax_scalar_params,
xnn_pack_f32_gemm_goi_w,
/*mr=*/5, /*nr=*/48, /*kr=*/1, /*sr=*/1,
benchmark::utils::CheckAVX512F);
}

BENCHMARK_GEMM(f32_gemm_minmax_ukernel_5x48__avx512f_broadcast)

static void f32_gemm_minmax_ukernel_6x48__avx512f_broadcast(benchmark::State& state, const char* net) {
GEMMBenchmark(state,
xnn_f32_gemm_minmax_ukernel_6x48__avx512f_broadcast,
xnn_init_f32_minmax_scalar_params,
xnn_pack_f32_gemm_goi_w,
/*mr=*/6, /*nr=*/48, /*kr=*/1, /*sr=*/1,
benchmark::utils::CheckAVX512F);
}

BENCHMARK_GEMM(f32_gemm_minmax_ukernel_6x48__avx512f_broadcast)

static void f32_gemm_minmax_ukernel_7x48__avx512f_broadcast(benchmark::State& state, const char* net) {
GEMMBenchmark(state,
xnn_f32_gemm_minmax_ukernel_7x48__avx512f_broadcast,
xnn_init_f32_minmax_scalar_params,
xnn_pack_f32_gemm_goi_w,
/*mr=*/7, /*nr=*/48, /*kr=*/1, /*sr=*/1,
benchmark::utils::CheckAVX512F);
}

BENCHMARK_GEMM(f32_gemm_minmax_ukernel_7x48__avx512f_broadcast)

static void f32_gemm_minmax_ukernel_8x48__avx512f_broadcast(benchmark::State& state, const char* net) {
GEMMBenchmark(state,
xnn_f32_gemm_minmax_ukernel_8x48__avx512f_broadcast,
xnn_init_f32_minmax_scalar_params,
xnn_pack_f32_gemm_goi_w,
/*mr=*/8, /*nr=*/48, /*kr=*/1, /*sr=*/1,
benchmark::utils::CheckAVX512F);
}

BENCHMARK_GEMM(f32_gemm_minmax_ukernel_8x48__avx512f_broadcast)

static void f32_gemm_minmax_ukernel_9x48__avx512f_broadcast(benchmark::State& state, const char* net) {
GEMMBenchmark(state,
xnn_f32_gemm_minmax_ukernel_9x48__avx512f_broadcast,
xnn_init_f32_minmax_scalar_params,
xnn_pack_f32_gemm_goi_w,
/*mr=*/9, /*nr=*/48, /*kr=*/1, /*sr=*/1,
benchmark::utils::CheckAVX512F);
}

BENCHMARK_GEMM(f32_gemm_minmax_ukernel_9x48__avx512f_broadcast)

static void f32_gemm_minmax_ukernel_10x48__avx512f_broadcast(benchmark::State& state, const char* net) {
GEMMBenchmark(state,
xnn_f32_gemm_minmax_ukernel_10x48__avx512f_broadcast,
xnn_init_f32_minmax_scalar_params,
xnn_pack_f32_gemm_goi_w,
/*mr=*/10, /*nr=*/48, /*kr=*/1, /*sr=*/1,
benchmark::utils::CheckAVX512F);
}

BENCHMARK_GEMM(f32_gemm_minmax_ukernel_10x48__avx512f_broadcast)

static void f32_gemm_minmax_ukernel_11x48__avx512f_broadcast(benchmark::State& state, const char* net) {
GEMMBenchmark(state,
xnn_f32_gemm_minmax_ukernel_11x48__avx512f_broadcast,
xnn_init_f32_minmax_scalar_params,
xnn_pack_f32_gemm_goi_w,
/*mr=*/11, /*nr=*/48, /*kr=*/1, /*sr=*/1,
benchmark::utils::CheckAVX512F);
}

BENCHMARK_GEMM(f32_gemm_minmax_ukernel_11x48__avx512f_broadcast)

static void f32_gemm_minmax_ukernel_12x48__avx512f_broadcast(benchmark::State& state, const char* net) {
GEMMBenchmark(state,
xnn_f32_gemm_minmax_ukernel_12x48__avx512f_broadcast,
xnn_init_f32_minmax_scalar_params,
xnn_pack_f32_gemm_goi_w,
/*mr=*/12, /*nr=*/48, /*kr=*/1, /*sr=*/1,
benchmark::utils::CheckAVX512F);
}

BENCHMARK_GEMM(f32_gemm_minmax_ukernel_12x48__avx512f_broadcast)

static void f32_gemm_minmax_ukernel_13x48__avx512f_broadcast(benchmark::State& state, const char* net) {
GEMMBenchmark(state,
xnn_f32_gemm_minmax_ukernel_13x48__avx512f_broadcast,
xnn_init_f32_minmax_scalar_params,
xnn_pack_f32_gemm_goi_w,
/*mr=*/13, /*nr=*/48, /*kr=*/1, /*sr=*/1,
benchmark::utils::CheckAVX512F);
}

BENCHMARK_GEMM(f32_gemm_minmax_ukernel_13x48__avx512f_broadcast)

static void f32_gemm_minmax_ukernel_14x48__avx512f_broadcast(benchmark::State& state, const char* net) {
GEMMBenchmark(state,
xnn_f32_gemm_minmax_ukernel_14x48__avx512f_broadcast,
xnn_init_f32_minmax_scalar_params,
xnn_pack_f32_gemm_goi_w,
/*mr=*/14, /*nr=*/48, /*kr=*/1, /*sr=*/1,
benchmark::utils::CheckAVX512F);
}

BENCHMARK_GEMM(f32_gemm_minmax_ukernel_14x48__avx512f_broadcast)

static void f32_gemm_minmax_ukernel_15x48__avx512f_broadcast(benchmark::State& state, const char* net) {
GEMMBenchmark(state,
xnn_f32_gemm_minmax_ukernel_15x48__avx512f_broadcast,
xnn_init_f32_minmax_scalar_params,
xnn_pack_f32_gemm_goi_w,
/*mr=*/15, /*nr=*/48, /*kr=*/1, /*sr=*/1,
benchmark::utils::CheckAVX512F);
}

BENCHMARK_GEMM(f32_gemm_minmax_ukernel_15x48__avx512f_broadcast)

static void f32_gemm_minmax_ukernel_16x48__avx512f_broadcast(benchmark::State& state, const char* net) {
GEMMBenchmark(state,
xnn_f32_gemm_minmax_ukernel_16x48__avx512f_broadcast,
xnn_init_f32_minmax_scalar_params,
xnn_pack_f32_gemm_goi_w,
/*mr=*/16, /*nr=*/48, /*kr=*/1, /*sr=*/1,
benchmark::utils::CheckAVX512F);
}

BENCHMARK_GEMM(f32_gemm_minmax_ukernel_16x48__avx512f_broadcast)

static void f32_gemm_minmax_ukernel_1x64__avx512f_broadcast(benchmark::State& state, const char* net) {
GEMMBenchmark(state,
xnn_f32_gemm_minmax_ukernel_1x64__avx512f_broadcast,
Expand Down
70 changes: 64 additions & 6 deletions cmake/gen/avx512f_microkernels.cmake
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -16,9 +16,11 @@ SET(PROD_AVX512F_MICROKERNEL_SRCS
src/f32-dwconv/gen/f32-dwconv-9p16c-minmax-avx512f.c
src/f32-dwconv/gen/f32-dwconv-25p16c-minmax-avx512f.c
src/f32-gemm/gen/f32-gemm-1x32-minmax-avx512f-broadcast.c
src/f32-gemm/gen/f32-gemm-7x32-minmax-avx512f-broadcast.c
src/f32-gemm/gen/f32-gemm-8x32-minmax-avx512f-broadcast.c
src/f32-gemm/gen/f32-gemm-13x32-minmax-avx512f-broadcast.c
src/f32-igemm/gen/f32-igemm-1x32-minmax-avx512f-broadcast.c
src/f32-igemm/gen/f32-igemm-7x32-minmax-avx512f-broadcast.c
src/f32-igemm/gen/f32-igemm-8x32-minmax-avx512f-broadcast.c
src/f32-igemm/gen/f32-igemm-13x32-minmax-avx512f-broadcast.c
src/f32-raddstoreexpminusmax/gen/f32-raddstoreexpminusmax-avx512f-rr2-p5-u64-acc2.c
src/f32-rdsum/gen/f32-rdsum-7p7x-minmax-avx512f-c64.c
src/f32-rminmax/gen/f32-rmax-avx512f-u64-acc4.c
Expand Down Expand Up @@ -65,7 +67,6 @@ SET(PROD_AVX512F_MICROKERNEL_SRCS
src/f32-vunary/gen/f32-vneg-avx512f.c
src/f32-vunary/gen/f32-vsqr-avx512f.c
src/s32-f32-vcvt/gen/s32-f32-vcvt-avx512f.c
src/x32-packw/gen/x32-packw-x32-gemm-gio-avx512f-u8.c
src/x32-packw/gen/x32-packw-x32-gemm-goi-avx512f-u4-prfm.c)

SET(NON_PROD_AVX512F_MICROKERNEL_SRCS
Expand All @@ -85,44 +86,57 @@ SET(NON_PROD_AVX512F_MICROKERNEL_SRCS
src/f32-dwconv/gen/f32-dwconv-25p32c-minmax-avx512f-acc2.c
src/f32-dwconv/gen/f32-dwconv-25p32c-minmax-avx512f.c
src/f32-gemm/gen/f32-gemm-1x16-minmax-avx512f-broadcast.c
src/f32-gemm/gen/f32-gemm-1x48-minmax-avx512f-broadcast.c
src/f32-gemm/gen/f32-gemm-1x64-minmax-avx512f-broadcast.c
src/f32-gemm/gen/f32-gemm-4x16-minmax-avx512f-broadcast.c
src/f32-gemm/gen/f32-gemm-4x32-minmax-avx512f-broadcast.c
src/f32-gemm/gen/f32-gemm-4x48-minmax-avx512f-broadcast.c
src/f32-gemm/gen/f32-gemm-4x64-minmax-avx512f-broadcast.c
src/f32-gemm/gen/f32-gemm-5x16-minmax-avx512f-broadcast.c
src/f32-gemm/gen/f32-gemm-5x32-minmax-avx512f-broadcast.c
src/f32-gemm/gen/f32-gemm-5x48-minmax-avx512f-broadcast.c
src/f32-gemm/gen/f32-gemm-5x64-minmax-avx512f-broadcast.c
src/f32-gemm/gen/f32-gemm-6x16-minmax-avx512f-broadcast.c
src/f32-gemm/gen/f32-gemm-6x32-minmax-avx512f-broadcast.c
src/f32-gemm/gen/f32-gemm-6x48-minmax-avx512f-broadcast.c
src/f32-gemm/gen/f32-gemm-6x64-minmax-avx512f-broadcast.c
src/f32-gemm/gen/f32-gemm-7x16-minmax-avx512f-broadcast.c
src/f32-gemm/gen/f32-gemm-7x32-minmax-avx512f-broadcast.c
src/f32-gemm/gen/f32-gemm-7x48-minmax-avx512f-broadcast.c
src/f32-gemm/gen/f32-gemm-7x64-minmax-avx512f-broadcast.c
src/f32-gemm/gen/f32-gemm-8x16-minmax-avx512f-broadcast.c
src/f32-gemm/gen/f32-gemm-8x32-minmax-avx512f-broadcast.c
src/f32-gemm/gen/f32-gemm-8x48-minmax-avx512f-broadcast.c
src/f32-gemm/gen/f32-gemm-8x64-minmax-avx512f-broadcast.c
src/f32-gemm/gen/f32-gemm-9x16-minmax-avx512f-broadcast.c
src/f32-gemm/gen/f32-gemm-9x32-minmax-avx512f-broadcast.c
src/f32-gemm/gen/f32-gemm-9x48-minmax-avx512f-broadcast.c
src/f32-gemm/gen/f32-gemm-9x64-minmax-avx512f-broadcast.c
src/f32-gemm/gen/f32-gemm-10x16-minmax-avx512f-broadcast.c
src/f32-gemm/gen/f32-gemm-10x32-minmax-avx512f-broadcast.c
src/f32-gemm/gen/f32-gemm-10x48-minmax-avx512f-broadcast.c
src/f32-gemm/gen/f32-gemm-10x64-minmax-avx512f-broadcast.c
src/f32-gemm/gen/f32-gemm-11x16-minmax-avx512f-broadcast.c
src/f32-gemm/gen/f32-gemm-11x32-minmax-avx512f-broadcast.c
src/f32-gemm/gen/f32-gemm-11x48-minmax-avx512f-broadcast.c
src/f32-gemm/gen/f32-gemm-11x64-minmax-avx512f-broadcast.c
src/f32-gemm/gen/f32-gemm-12x16-minmax-avx512f-broadcast.c
src/f32-gemm/gen/f32-gemm-12x32-minmax-avx512f-broadcast.c
src/f32-gemm/gen/f32-gemm-12x48-minmax-avx512f-broadcast.c
src/f32-gemm/gen/f32-gemm-12x64-minmax-avx512f-broadcast.c
src/f32-gemm/gen/f32-gemm-13x16-minmax-avx512f-broadcast.c
src/f32-gemm/gen/f32-gemm-13x32-minmax-avx512f-broadcast.c
src/f32-gemm/gen/f32-gemm-13x48-minmax-avx512f-broadcast.c
src/f32-gemm/gen/f32-gemm-13x64-minmax-avx512f-broadcast.c
src/f32-gemm/gen/f32-gemm-14x16-minmax-avx512f-broadcast.c
src/f32-gemm/gen/f32-gemm-14x32-minmax-avx512f-broadcast.c
src/f32-gemm/gen/f32-gemm-14x48-minmax-avx512f-broadcast.c
src/f32-gemm/gen/f32-gemm-14x64-minmax-avx512f-broadcast.c
src/f32-gemm/gen/f32-gemm-15x16-minmax-avx512f-broadcast.c
src/f32-gemm/gen/f32-gemm-15x32-minmax-avx512f-broadcast.c
src/f32-gemm/gen/f32-gemm-15x48-minmax-avx512f-broadcast.c
src/f32-gemm/gen/f32-gemm-15x64-minmax-avx512f-broadcast.c
src/f32-gemm/gen/f32-gemm-16x16-minmax-avx512f-broadcast.c
src/f32-gemm/gen/f32-gemm-16x32-minmax-avx512f-broadcast.c
src/f32-gemm/gen/f32-gemm-16x48-minmax-avx512f-broadcast.c
src/f32-gemm/gen/f32-gemm-16x64-minmax-avx512f-broadcast.c
src/f32-gemminc/gen/f32-gemminc-1x16-minmax-avx512f-broadcast.c
src/f32-gemminc/gen/f32-gemminc-4x16-minmax-avx512f-broadcast.c
Expand All @@ -131,15 +145,58 @@ SET(NON_PROD_AVX512F_MICROKERNEL_SRCS
src/f32-gemminc/gen/f32-gemminc-7x16-minmax-avx512f-broadcast.c
src/f32-gemminc/gen/f32-gemminc-8x16-minmax-avx512f-broadcast.c
src/f32-igemm/gen/f32-igemm-1x16-minmax-avx512f-broadcast.c
src/f32-igemm/gen/f32-igemm-1x48-minmax-avx512f-broadcast.c
src/f32-igemm/gen/f32-igemm-1x64-minmax-avx512f-broadcast.c
src/f32-igemm/gen/f32-igemm-4x16-minmax-avx512f-broadcast.c
src/f32-igemm/gen/f32-igemm-4x32-minmax-avx512f-broadcast.c
src/f32-igemm/gen/f32-igemm-4x48-minmax-avx512f-broadcast.c
src/f32-igemm/gen/f32-igemm-4x64-minmax-avx512f-broadcast.c
src/f32-igemm/gen/f32-igemm-5x16-minmax-avx512f-broadcast.c
src/f32-igemm/gen/f32-igemm-5x32-minmax-avx512f-broadcast.c
src/f32-igemm/gen/f32-igemm-5x48-minmax-avx512f-broadcast.c
src/f32-igemm/gen/f32-igemm-5x64-minmax-avx512f-broadcast.c
src/f32-igemm/gen/f32-igemm-6x16-minmax-avx512f-broadcast.c
src/f32-igemm/gen/f32-igemm-6x32-minmax-avx512f-broadcast.c
src/f32-igemm/gen/f32-igemm-6x48-minmax-avx512f-broadcast.c
src/f32-igemm/gen/f32-igemm-6x64-minmax-avx512f-broadcast.c
src/f32-igemm/gen/f32-igemm-7x16-minmax-avx512f-broadcast.c
src/f32-igemm/gen/f32-igemm-7x32-minmax-avx512f-broadcast.c
src/f32-igemm/gen/f32-igemm-7x48-minmax-avx512f-broadcast.c
src/f32-igemm/gen/f32-igemm-7x64-minmax-avx512f-broadcast.c
src/f32-igemm/gen/f32-igemm-8x16-minmax-avx512f-broadcast.c
src/f32-igemm/gen/f32-igemm-8x32-minmax-avx512f-broadcast.c
src/f32-igemm/gen/f32-igemm-8x48-minmax-avx512f-broadcast.c
src/f32-igemm/gen/f32-igemm-8x64-minmax-avx512f-broadcast.c
src/f32-igemm/gen/f32-igemm-9x16-minmax-avx512f-broadcast.c
src/f32-igemm/gen/f32-igemm-9x32-minmax-avx512f-broadcast.c
src/f32-igemm/gen/f32-igemm-9x48-minmax-avx512f-broadcast.c
src/f32-igemm/gen/f32-igemm-9x64-minmax-avx512f-broadcast.c
src/f32-igemm/gen/f32-igemm-10x16-minmax-avx512f-broadcast.c
src/f32-igemm/gen/f32-igemm-10x32-minmax-avx512f-broadcast.c
src/f32-igemm/gen/f32-igemm-10x48-minmax-avx512f-broadcast.c
src/f32-igemm/gen/f32-igemm-10x64-minmax-avx512f-broadcast.c
src/f32-igemm/gen/f32-igemm-11x16-minmax-avx512f-broadcast.c
src/f32-igemm/gen/f32-igemm-11x32-minmax-avx512f-broadcast.c
src/f32-igemm/gen/f32-igemm-11x48-minmax-avx512f-broadcast.c
src/f32-igemm/gen/f32-igemm-11x64-minmax-avx512f-broadcast.c
src/f32-igemm/gen/f32-igemm-12x16-minmax-avx512f-broadcast.c
src/f32-igemm/gen/f32-igemm-12x32-minmax-avx512f-broadcast.c
src/f32-igemm/gen/f32-igemm-12x48-minmax-avx512f-broadcast.c
src/f32-igemm/gen/f32-igemm-12x64-minmax-avx512f-broadcast.c
src/f32-igemm/gen/f32-igemm-13x16-minmax-avx512f-broadcast.c
src/f32-igemm/gen/f32-igemm-13x48-minmax-avx512f-broadcast.c
src/f32-igemm/gen/f32-igemm-13x64-minmax-avx512f-broadcast.c
src/f32-igemm/gen/f32-igemm-14x16-minmax-avx512f-broadcast.c
src/f32-igemm/gen/f32-igemm-14x32-minmax-avx512f-broadcast.c
src/f32-igemm/gen/f32-igemm-14x48-minmax-avx512f-broadcast.c
src/f32-igemm/gen/f32-igemm-14x64-minmax-avx512f-broadcast.c
src/f32-igemm/gen/f32-igemm-15x16-minmax-avx512f-broadcast.c
src/f32-igemm/gen/f32-igemm-15x32-minmax-avx512f-broadcast.c
src/f32-igemm/gen/f32-igemm-15x48-minmax-avx512f-broadcast.c
src/f32-igemm/gen/f32-igemm-15x64-minmax-avx512f-broadcast.c
src/f32-igemm/gen/f32-igemm-16x16-minmax-avx512f-broadcast.c
src/f32-igemm/gen/f32-igemm-16x32-minmax-avx512f-broadcast.c
src/f32-igemm/gen/f32-igemm-16x48-minmax-avx512f-broadcast.c
src/f32-igemm/gen/f32-igemm-16x64-minmax-avx512f-broadcast.c
src/f32-raddexpminusmax/gen/f32-raddexpminusmax-avx512f-p5-scalef-u64-acc2.c
src/f32-raddexpminusmax/gen/f32-raddexpminusmax-avx512f-p5-scalef-u64-acc4.c
src/f32-raddexpminusmax/gen/f32-raddexpminusmax-avx512f-p5-scalef-u64.c
Expand Down Expand Up @@ -330,6 +387,7 @@ SET(NON_PROD_AVX512F_MICROKERNEL_SRCS
src/x32-packw/gen/x32-packw-x32-gemm-gio-avx512f-u1-prfm.c
src/x32-packw/gen/x32-packw-x32-gemm-gio-avx512f-u1.c
src/x32-packw/gen/x32-packw-x32-gemm-gio-avx512f-u8-prfm.c
src/x32-packw/gen/x32-packw-x32-gemm-gio-avx512f-u8.c
src/x32-packw/gen/x32-packw-x32-gemm-goi-avx512f-u4.c)

SET(ALL_AVX512F_MICROKERNEL_SRCS ${PROD_AVX512F_MICROKERNEL_SRCS} + ${NON_PROD_AVX512F_MICROKERNEL_SRCS})
2 changes: 1 addition & 1 deletion cmake/gen/scalar_microkernels.cmake
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -241,6 +241,7 @@ SET(PROD_SCALAR_MICROKERNEL_SRCS
src/x24-transposec/gen/x24-transposec-1x2-scalar.c
src/x32-packw/gen/x32-packw-x2-gemm-goi-scalar-float-u4.c
src/x32-packw/gen/x32-packw-x4-gemm-goi-scalar-float-u4.c
src/x32-packw/gen/x32-packw-x32-gemm-gio-scalar.c
src/x32-transposec/gen/x32-transposec-2x4-scalar-int.c
src/x32-unpool/x32-unpool-scalar.c
src/x32-zip/x32-zip-x2-scalar.c
Expand Down Expand Up @@ -873,7 +874,6 @@ SET(NON_PROD_SCALAR_MICROKERNEL_SRCS
src/x32-packw/gen/x32-packw-x16-gemm-gio-scalar.c
src/x32-packw/gen/x32-packw-x16-gemm-goi-scalar-float-u4.c
src/x32-packw/gen/x32-packw-x16-gemm-goi-scalar-int-u4.c
src/x32-packw/gen/x32-packw-x32-gemm-gio-scalar.c
src/x32-packx/x32-packx-2x-scalar.c
src/x32-packx/x32-packx-3x-scalar.c
src/x32-packx/x32-packx-4x-scalar.c
Expand Down
Loading