[Mesa-dev] [PATCH 3/3] amd/common: only load used channels when sampling buffer views

[Samuel Pitoiset]

This allows to reduce the number of dwords that are loaded
with buffer_load_format_xyzw. For example, when the only used
channel is 1, the driver will emit buffer_load_format_x instead.

Shader stats for DOW3 (with some local hacky scripts for SPIRV):

143 shaders in 143 tests
Totals:
SGPRS: 5344 -> 5352 (0.15 %)
VGPRS: 3476 -> 3452 (-0.69 %)
Spilled SGPRs: 30 -> 29 (-3.33 %)
Spilled VGPRs: 0 -> 0 (0.00 %)
Private memory VGPRs: 0 -> 0 (0.00 %)
Scratch size: 0 -> 0 (0.00 %) dwords per thread
Code Size: 269860 -> 269808 (-0.02 %) bytes
LDS: 0 -> 0 (0.00 %) blocks
Max Waves: 1267 -> 1272 (0.39 %)
Wait states: 0 -> 0 (0.00 %)

The 'const' qualifier has to be removed to avoid a compilation
warning with nir_ssa_def_components_read().

Signed-off-by: Samuel Pitoiset <samuel.pitoiset at gmail.com>
---
 src/amd/common/ac_nir_to_llvm.c | 7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)

diff --git a/src/amd/common/ac_nir_to_llvm.c b/src/amd/common/ac_nir_to_llvm.c
index 8f5df12e3d..bafbdbb250 100644
--- a/src/amd/common/ac_nir_to_llvm.c
+++ b/src/amd/common/ac_nir_to_llvm.c
@@ -2242,16 +2242,19 @@ static LLVMValueRef radv_lower_gather4_integer(struct ac_llvm_context *ctx,
 }
 
 static LLVMValueRef build_tex_intrinsic(struct ac_nir_context *ctx,
-					const nir_tex_instr *instr,
+					nir_tex_instr *instr,
 					bool lod_is_zero,
 					struct ac_image_args *args)
 {
 	if (instr->sampler_dim == GLSL_SAMPLER_DIM_BUF) {
+		unsigned mask = nir_ssa_def_components_read(&instr->dest.ssa);
+
 		return ac_build_buffer_load_format(&ctx->ac,
 						   args->resource,
 						   args->addr,
 						   ctx->ac.i32_0,
-						   4, true);
+						   util_last_bit(mask),
+						   true);
 	}
 
 	args->opcode = ac_image_sample;
-- 
2.15.1