OpenHarmony/third_party_mesa3d
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

tu: Update HS_WAVE_INPUT_SIZE formula

Browse Source 
A better explanation for SP_HS_WAVE_INPUT_SIZE is that it is the size
of local memory to allocate per wave (which can be more than one
patch), in 256B units.

Then the maximum of 64 makes sense because only 16KB of local memory
is reserved for VS<->HS linkage.

The resulting formula matches the blob behaviour, even when
patch_control_points and tcs_vertices_out have different values,
while the past formula gave wrong answers on gen3+.

Signed-off-by: Danylo Piliaiev <dpiliaiev@igalia.com>
Suggested-by: Jonathan Marek <jonathan@marek.ca>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17957>
This commit is contained in:
Danylo Piliaiev
2022-08-12 17:31:59 +03:00
parent a7db1da37d
commit 1eb7a85b55
1 changed files with 23 additions and 12 deletions
Download Patch File Download Diff File Expand all files Collapse all files

35
src/freedreno/vulkan/tu_pipeline.c
View File

@@ -1258,24 +1258,35 @@ tu6_emit_vpc(struct tu_cs *cs,
tu_cs_emit_pkt4(cs, REG_A6XX_PC_TESS_NUM_VERTEX, 1);
tu_cs_emit(cs, hs->tess.tcs_vertices_out);
uint32_t patch_local_mem_size_16b =
patch_control_points * vs->output_size / 4;
/* Total attribute slots in HS incoming patch. */
tu_cs_emit_pkt4(cs, REG_A6XX_PC_HS_INPUT_SIZE, 1);
tu_cs_emit(cs, patch_control_points * vs->output_size / 4);
tu_cs_emit(cs, patch_local_mem_size_16b);
const uint32_t wavesize = 64;
const uint32_t max_wave_input_size = 64;
const uint32_t vs_hs_local_mem_size = 16384;
/* note: if HS is really just the VS extended, then this
* should be by MAX2(patch_control_points, hs->tess.tcs_vertices_out)
* however that doesn't match the blob, and fails some dEQP tests.
*/
uint32_t prims_per_wave = wavesize / hs->tess.tcs_vertices_out;
uint32_t max_prims_per_wave =
max_wave_input_size * wavesize / (vs->output_size * patch_control_points);
prims_per_wave = MIN2(prims_per_wave, max_prims_per_wave);
uint32_t max_patches_per_wave;
if (cs->device->physical_device->info->a6xx.tess_use_shared) {
/* HS invocations for a patch are always within the same wave,
* making barriers less expensive. VS can't have barriers so we
* don't care about VS invocations being in the same wave.
*/
max_patches_per_wave = wavesize / hs->tess.tcs_vertices_out;
} else {
/* VS is also in the same wave */
max_patches_per_wave =
wavesize / MAX2(patch_control_points, hs->tess.tcs_vertices_out);
}
uint32_t total_size = vs->output_size * patch_control_points * prims_per_wave;
uint32_t wave_input_size = DIV_ROUND_UP(total_size, wavesize);
uint32_t patches_per_wave =
MIN2(vs_hs_local_mem_size / (patch_local_mem_size_16b * 16),
max_patches_per_wave);
uint32_t wave_input_size = DIV_ROUND_UP(
patches_per_wave * patch_local_mem_size_16b * 16, 256);
tu_cs_emit_pkt4(cs, REG_A6XX_SP_HS_WAVE_INPUT_SIZE, 1);
tu_cs_emit(cs, wave_input_size);