fdo-mirrors/mesa
1
0
Fork 0
mirror of https://gitlab.freedesktop.org/mesa/mesa.git synced 2026-05-04 18:28:12 +02:00
Code Issues Projects Releases Packages Wiki Activity Actions 2

radv/nir/lower_cmat: clean up GFX11 ACC->B convert
Some checks are pending
macOS-CI / macOS-CI (dri) (push) Waiting to run
Details
macOS-CI / macOS-CI (xlib) (push) Waiting to run
Details

Browse source 
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/37213>
This commit is contained in:
Georg Lehmann 2025-09-06 17:13:33 +02:00 • committed by Marge Bot
parent 5c0ebcdaef
commit 4143f0725a
1 changed files with 9 additions and 28 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

37
src/amd/vulkan/nir/radv_nir_lower_cooperative_matrix.c
View file

@ -471,11 +471,12 @@ convert_use(nir_builder *b, nir_def *src, enum glsl_cmat_use src_use, enum glsl_
nir_def *tmp[NIR_MAX_VEC_COMPONENTS];
if (src->bit_size == 32) {
if (params->wave_size == 64) {
nir_def *low_lanes = nir_inverse_ballot_imm(b, UINT32_MAX, 64);
for (int cross32 = params->wave_size == 64; cross32 >= 0; cross32--) {
uint64_t low_mask = cross32 ? UINT32_MAX : 0xffff0000ffffull;
nir_def *low_lanes = nir_inverse_ballot_imm(b, low_mask, params->wave_size);
for (int i = 0; i < num_comps; i++) {
nir_def *comp = components[i];
nir_def *half_swap = shuffle_xor_imm(b, comp, 32);
nir_def *half_swap = shuffle_xor_imm(b, comp, cross32 ? 32 : 16);
tmp[i * 2] = nir_bcsel(b, low_lanes, comp, half_swap);
tmp[i * 2 + 1] = nir_bcsel(b, low_lanes, half_swap, comp);
@ -483,16 +484,6 @@ convert_use(nir_builder *b, nir_def *src, enum glsl_cmat_use src_use, enum glsl_
num_comps *= 2;
memcpy(components, tmp, sizeof(components));
}
nir_def *low_lanes = nir_inverse_ballot_imm(b, 0xffff0000ffffull, params->wave_size);
for (int i = 0; i < num_comps; i++) {
nir_def *half_swap = shuffle_xor_imm(b, components[i], 16);
tmp[i * 2] = nir_bcsel(b, low_lanes, components[i], half_swap);
tmp[i * 2 + 1] = nir_bcsel(b, low_lanes, half_swap, components[i]);
}
num_comps *= 2;
memcpy(components, tmp, sizeof(components));
} else {
/* Same as above, but operate on 32 bits at once, using byte_perm as vectorized bcsel. */
assert(src->bit_size < 32);
@ -504,13 +495,15 @@ convert_use(nir_builder *b, nir_def *src, enum glsl_cmat_use src_use, enum glsl_
nir_def *low_sel = nir_imm_int(b, src->bit_size == 8 ? 0x05010400 : 0x05040100);
nir_def *high_sel = nir_imm_int(b, src->bit_size == 8 ? 0x01050004 : 0x01000504);
if (params->wave_size == 64) {
nir_def *low_lanes = nir_inverse_ballot_imm(b, UINT32_MAX, 64);
for (int cross32 = params->wave_size == 64; cross32 >= 0; cross32--) {
uint64_t low_mask = cross32 ? UINT32_MAX : 0xffff0000ffffull;
nir_def *low_lanes = nir_inverse_ballot_imm(b, low_mask, params->wave_size);
nir_def *first_perm = nir_bcsel(b, low_lanes, low_sel, high_sel);
nir_def *second_perm = nir_ior_imm(b, first_perm, 0x02020202);
for (int i = 0; i < num_comps; i++) {
nir_def *comp = components[i];
nir_def *half_swap = shuffle_xor_imm(b, comp, 32);
nir_def *half_swap = shuffle_xor_imm(b, comp, cross32 ? 32 : 16);
tmp[i * 2] = nir_byte_perm_amd(b, half_swap, comp, first_perm);
tmp[i * 2 + 1] = nir_byte_perm_amd(b, half_swap, comp, second_perm);
@ -519,18 +512,6 @@ convert_use(nir_builder *b, nir_def *src, enum glsl_cmat_use src_use, enum glsl_
memcpy(components, tmp, sizeof(components));
}
nir_def *low_lanes = nir_inverse_ballot_imm(b, 0xffff0000ffffull, params->wave_size);
nir_def *first_perm = nir_bcsel(b, low_lanes, low_sel, high_sel);
nir_def *second_perm = nir_ior_imm(b, first_perm, 0x02020202);
for (int i = 0; i < num_comps; i++) {
nir_def *comp = components[i];
nir_def *half_swap = shuffle_xor_imm(b, comp, 16);
tmp[i * 2] = nir_byte_perm_amd(b, half_swap, comp, first_perm);
tmp[i * 2 + 1] = nir_byte_perm_amd(b, half_swap, comp, second_perm);
}
num_comps *= 2;
memcpy(components, tmp, sizeof(components));
nir_def *unpacked =
nir_extract_bits(b, components, num_comps, 0, num_comps * (32 / src->bit_size), src->bit_size);
num_comps *= 32 / src->bit_size;