From: Marek Olšák <marek.ol...@amd.com> Diff from shader-db: Scratch: 3221504 -> 17408 (-99.46 %) bytes per wave --- src/gallium/drivers/radeonsi/si_shader.c | 17 +++++++++++++++-- 1 file changed, 15 insertions(+), 2 deletions(-)
diff --git a/src/gallium/drivers/radeonsi/si_shader.c b/src/gallium/drivers/radeonsi/si_shader.c index 6e8eefb..d392807 100644 --- a/src/gallium/drivers/radeonsi/si_shader.c +++ b/src/gallium/drivers/radeonsi/si_shader.c @@ -5825,6 +5825,18 @@ void si_shader_binary_read_config(struct radeon_shader_binary *binary, unsigned i; const unsigned char *config = radeon_shader_binary_config_start(binary, symbol_offset); + bool really_needs_scratch = false; + + /* LLVM adds SGPR spills to the scratch size. + * Find out if we really need the scratch buffer. + */ + for (i = 0; i < binary->reloc_count; i++) { + const struct radeon_shader_reloc *reloc = &binary->relocs[i]; + + if (!strcmp(scratch_rsrc_dword0_symbol, reloc->name) || + !strcmp(scratch_rsrc_dword1_symbol, reloc->name)) + really_needs_scratch = true; + } /* XXX: We may be able to emit some of these values directly rather than * extracting fields to be emitted later. @@ -5859,8 +5871,9 @@ void si_shader_binary_read_config(struct radeon_shader_binary *binary, case R_0286E8_SPI_TMPRING_SIZE: case R_00B860_COMPUTE_TMPRING_SIZE: /* WAVESIZE is in units of 256 dwords. */ - conf->scratch_bytes_per_wave = - G_00B860_WAVESIZE(value) * 256 * 4 * 1; + if (really_needs_scratch) + conf->scratch_bytes_per_wave = + G_00B860_WAVESIZE(value) * 256 * 4; break; default: { -- 2.7.4 _______________________________________________ mesa-dev mailing list mesa-dev@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/mesa-dev