diff options
| author | Marek Olšák <marek.olsak@amd.com> | 2016-09-13 17:33:23 +0200 |
|---|---|---|
| committer | Marek Olšák <marek.olsak@amd.com> | 2016-09-14 12:33:00 +0200 |
| commit | ab29788250a705eb0dd517cb3d38f37f944eb8ad (patch) | |
| tree | 5b14ecf4f70cc117e4c8e6b47c8c0209a7f4559d /src/gallium/drivers/radeonsi/si_shader.c | |
| parent | 007b512f9d05875b0dda956230fd3882dfea30af (diff) | |
| download | external_mesa3d-ab29788250a705eb0dd517cb3d38f37f944eb8ad.zip external_mesa3d-ab29788250a705eb0dd517cb3d38f37f944eb8ad.tar.gz external_mesa3d-ab29788250a705eb0dd517cb3d38f37f944eb8ad.tar.bz2 | |
radeonsi: reload PS inputs with direct indexing at each use (v2)
The LLVM compiler can CSE interp intrinsics thanks to
LLVMReadNoneAttribute.
26011 shaders in 14651 tests
Totals:
SGPRS: 1146340 -> 1132676 (-1.19 %)
VGPRS: 727371 -> 711730 (-2.15 %)
Spilled SGPRs: 2218 -> 2078 (-6.31 %)
Spilled VGPRs: 369 -> 369 (0.00 %)
Scratch VGPRs: 1344 -> 1344 (0.00 %) dwords per thread
Code Size: 35841268 -> 36009732 (0.47 %) bytes
LDS: 767 -> 767 (0.00 %) blocks
Max Waves: 222559 -> 224779 (1.00 %)
Wait states: 0 -> 0 (0.00 %)
v2: don't call load_input for fragment shaders in emit_declaration
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Diffstat (limited to 'src/gallium/drivers/radeonsi/si_shader.c')
| -rw-r--r-- | src/gallium/drivers/radeonsi/si_shader.c | 27 |
1 files changed, 11 insertions, 16 deletions
diff --git a/src/gallium/drivers/radeonsi/si_shader.c b/src/gallium/drivers/radeonsi/si_shader.c index b034837..3ccff7a 100644 --- a/src/gallium/drivers/radeonsi/si_shader.c +++ b/src/gallium/drivers/radeonsi/si_shader.c @@ -440,7 +440,8 @@ static LLVMValueRef get_instance_index_for_fetch( static void declare_input_vs( struct radeon_llvm_context *radeon_bld, unsigned input_index, - const struct tgsi_full_declaration *decl) + const struct tgsi_full_declaration *decl, + LLVMValueRef out[4]) { struct lp_build_context *base = &radeon_bld->soa.bld_base.base; struct gallivm_state *gallivm = base->gallivm; @@ -498,11 +499,8 @@ static void declare_input_vs( /* Break up the vec4 into individual components */ for (chan = 0; chan < 4; chan++) { LLVMValueRef llvm_chan = lp_build_const_int32(gallivm, chan); - /* XXX: Use a helper function for this. There is one in - * tgsi_llvm.c. */ - ctx->radeon_bld.inputs[radeon_llvm_reg_index_soa(input_index, chan)] = - LLVMBuildExtractElement(gallivm->builder, - input, llvm_chan, ""); + out[chan] = LLVMBuildExtractElement(gallivm->builder, + input, llvm_chan, ""); } } @@ -1463,7 +1461,8 @@ static LLVMValueRef get_interp_param(struct si_shader_context *ctx, static void declare_input_fs( struct radeon_llvm_context *radeon_bld, unsigned input_index, - const struct tgsi_full_declaration *decl) + const struct tgsi_full_declaration *decl, + LLVMValueRef out[4]) { struct lp_build_context *base = &radeon_bld->soa.bld_base.base; struct si_shader_context *ctx = @@ -1482,14 +1481,10 @@ static void declare_input_fs( unsigned offset = SI_PARAM_POS_FIXED_PT + 1 + (i ? util_bitcount(colors_read & 0xf) : 0); - radeon_bld->inputs[radeon_llvm_reg_index_soa(input_index, 0)] = - mask & 0x1 ? LLVMGetParam(main_fn, offset++) : base->undef; - radeon_bld->inputs[radeon_llvm_reg_index_soa(input_index, 1)] = - mask & 0x2 ? LLVMGetParam(main_fn, offset++) : base->undef; - radeon_bld->inputs[radeon_llvm_reg_index_soa(input_index, 2)] = - mask & 0x4 ? LLVMGetParam(main_fn, offset++) : base->undef; - radeon_bld->inputs[radeon_llvm_reg_index_soa(input_index, 3)] = - mask & 0x8 ? LLVMGetParam(main_fn, offset++) : base->undef; + out[0] = mask & 0x1 ? LLVMGetParam(main_fn, offset++) : base->undef; + out[1] = mask & 0x2 ? LLVMGetParam(main_fn, offset++) : base->undef; + out[2] = mask & 0x4 ? LLVMGetParam(main_fn, offset++) : base->undef; + out[3] = mask & 0x8 ? LLVMGetParam(main_fn, offset++) : base->undef; return; } @@ -1513,7 +1508,7 @@ static void declare_input_fs( shader->selector->info.colors_read, interp_param, LLVMGetParam(main_fn, SI_PARAM_PRIM_MASK), LLVMGetParam(main_fn, SI_PARAM_FRONT_FACE), - &radeon_bld->inputs[radeon_llvm_reg_index_soa(input_index, 0)]); + &out[0]); } static LLVMValueRef get_sample_id(struct radeon_llvm_context *radeon_bld) |