fdo-mirrors/mesa
1
0
Fork 0
mirror of https://gitlab.freedesktop.org/mesa/mesa.git synced 2026-01-03 05:00:09 +01:00
Code Issues Projects Releases Packages Wiki Activity Actions 1

ac/nir: return 0x76543210 for NULL FMASK fetch

Browse source 
This can replace several v_cndmask_b32 with a single v_cndmask_b32.

fossil-db (Sienna Cichlid):
Totals from 1169 (0.78% of 150170) affected shaders:
SGPRs: 82032 -> 81984 (-0.06%); split: -0.19%, +0.13%
VGPRs: 68368 -> 68280 (-0.13%); split: -0.30%, +0.18%
SpillSGPRs: 8754 -> 8609 (-1.66%); split: -4.06%, +2.40%
CodeSize: 5460132 -> 5446308 (-0.25%); split: -0.42%, +0.17%
MaxWaves: 24456 -> 24462 (+0.02%); split: +0.03%, -0.01%

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12214>
This commit is contained in:
Rhys Perry 2021-08-04 17:02:14 +01:00 • committed by Marge Bot
parent bf0cc05227
commit 91a5c9fdf2
2 changed files with 20 additions and 13 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

25
src/amd/llvm/ac_llvm_build.c
View file

@ -3348,6 +3348,16 @@ void ac_apply_fmask_to_sample(struct ac_llvm_context *ac, LLVMValueRef fmask, LL
LLVMValueRef fmask_value = ac_build_image_opcode(ac, &fmask_load);
fmask_value = LLVMBuildExtractElement(ac->builder, fmask_value, ac->i32_0, "");
/* Don't rewrite the sample index if WORD1.DATA_FORMAT of the FMASK
* resource descriptor is 0 (invalid).
*/
LLVMValueRef tmp;
tmp = LLVMBuildBitCast(ac->builder, fmask, ac->v8i32, "");
tmp = LLVMBuildExtractElement(ac->builder, tmp, ac->i32_1, "");
tmp = LLVMBuildICmp(ac->builder, LLVMIntNE, tmp, ac->i32_0, "");
fmask_value =
LLVMBuildSelect(ac->builder, tmp, fmask_value, LLVMConstInt(ac->i32, 0x76543210, false), "");
/* Apply the formula. */
unsigned sample_chan = is_array_tex ? 3 : 2;
LLVMValueRef final_sample;
@ -3357,20 +3367,9 @@ void ac_apply_fmask_to_sample(struct ac_llvm_context *ac, LLVMValueRef fmask, LL
LLVMBuildZExt(ac->builder, final_sample, ac->i32, ""), "");
/* Mask the sample index by 0x7, because 0x8 means an unknown value
* with EQAA, so those will map to 0. */
final_sample = LLVMBuildAnd(ac->builder, final_sample, LLVMConstInt(ac->i32, 0x7, 0), "");
addr[sample_chan] = LLVMBuildAnd(ac->builder, final_sample, LLVMConstInt(ac->i32, 0x7, 0), "");
if (fmask_load.a16)
final_sample = LLVMBuildTrunc(ac->builder, final_sample, ac->i16, "");
/* Don't rewrite the sample index if WORD1.DATA_FORMAT of the FMASK
* resource descriptor is 0 (invalid).
*/
LLVMValueRef tmp;
tmp = LLVMBuildBitCast(ac->builder, fmask, ac->v8i32, "");
tmp = LLVMBuildExtractElement(ac->builder, tmp, ac->i32_1, "");
tmp = LLVMBuildICmp(ac->builder, LLVMIntNE, tmp, ac->i32_0, "");
/* Replace the MSAA sample index. */
addr[sample_chan] = LLVMBuildSelect(ac->builder, tmp, final_sample, addr[sample_chan], "");
addr[sample_chan] = LLVMBuildTrunc(ac->builder, final_sample, ac->i16, "");
}
static LLVMValueRef _ac_build_readlane(struct ac_llvm_context *ctx, LLVMValueRef src,

8
src/amd/llvm/ac_nir_to_llvm.c
View file

@ -4850,6 +4850,14 @@ static void visit_tex(struct ac_nir_context *ctx, nir_tex_instr *instr)
LLVMValueRef two = LLVMConstInt(ctx->ac.i32, 2, false);
LLVMValueRef layers = LLVMBuildExtractElement(ctx->ac.builder, result, two, "");
result = LLVMBuildInsertElement(ctx->ac.builder, result, layers, ctx->ac.i32_1, "");
} else if (instr->op == nir_texop_fragment_mask_fetch_amd) {
/* Use 0x76543210 if the image doesn't have FMASK. */
LLVMValueRef tmp = LLVMBuildBitCast(ctx->ac.builder, args.resource, ctx->ac.v8i32, "");
tmp = LLVMBuildExtractElement(ctx->ac.builder, tmp, ctx->ac.i32_1, "");
tmp = LLVMBuildICmp(ctx->ac.builder, LLVMIntNE, tmp, ctx->ac.i32_0, "");
result = LLVMBuildSelect(ctx->ac.builder, tmp,
LLVMBuildExtractElement(ctx->ac.builder, result, ctx->ac.i32_0, ""),
LLVMConstInt(ctx->ac.i32, 0x76543210, false), "");
} else if (nir_tex_instr_result_size(instr) != 4)
result = ac_trim_vector(&ctx->ac, result, instr->dest.ssa.num_components);