summaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
authorSamuel Pitoiset <[email protected]>2018-03-01 11:54:22 +0100
committerSamuel Pitoiset <[email protected]>2018-03-06 10:38:35 +0100
commitf3275ca01cddd5d1e999e3805eff42e06ce6e974 (patch)
tree12764dd976976206055198a313b420192315169d
parent675dde13b2decc92a1babd545289a9912323861c (diff)
ac/nir: only enable used channels when exporting parameters
This allows us to generate, for example, "exp param0 v0, off, off, off" if only the first channel is needed. Not sure if this improves performance but it's worth trying. Signed-off-by: Samuel Pitoiset <[email protected]> Reviewed-by: Dave Airlie <[email protected]>
-rw-r--r--src/amd/common/ac_nir_to_llvm.c24
1 files changed, 20 insertions, 4 deletions
diff --git a/src/amd/common/ac_nir_to_llvm.c b/src/amd/common/ac_nir_to_llvm.c
index 83d5d19fb82..39d92ca0e0a 100644
--- a/src/amd/common/ac_nir_to_llvm.c
+++ b/src/amd/common/ac_nir_to_llvm.c
@@ -5928,11 +5928,11 @@ si_llvm_init_export_args(struct radv_shader_context *ctx,
static void
radv_export_param(struct radv_shader_context *ctx, unsigned index,
- LLVMValueRef *values)
+ LLVMValueRef *values, unsigned enabled_channels)
{
struct ac_export_args args;
- si_llvm_init_export_args(ctx, values, 0xf,
+ si_llvm_init_export_args(ctx, values, enabled_channels,
V_008DFC_SQ_EXP_PARAM + index, &args);
ac_build_export(&ctx->ac, &args);
}
@@ -6092,7 +6092,23 @@ handle_vs_outputs_post(struct radv_shader_context *ctx,
for (unsigned j = 0; j < 4; j++)
values[j] = ac_to_float(&ctx->ac, radv_load_output(ctx, i, j));
- radv_export_param(ctx, param_count, values);
+ unsigned output_usage_mask;
+
+ if (ctx->stage == MESA_SHADER_VERTEX &&
+ !ctx->is_gs_copy_shader) {
+ output_usage_mask =
+ ctx->shader_info->info.vs.output_usage_mask[i];
+ } else if (ctx->stage == MESA_SHADER_TESS_EVAL) {
+ output_usage_mask =
+ ctx->shader_info->info.tes.output_usage_mask[i];
+ } else {
+ /* Enable all channels for the GS copy shader because
+ * we don't know the output usage mask currently.
+ */
+ output_usage_mask = 0xf;
+ }
+
+ radv_export_param(ctx, param_count, values, output_usage_mask);
outinfo->vs_output_param_offset[i] = param_count++;
}
@@ -6106,7 +6122,7 @@ handle_vs_outputs_post(struct radv_shader_context *ctx,
for (unsigned j = 1; j < 4; j++)
values[j] = ctx->ac.f32_0;
- radv_export_param(ctx, param_count, values);
+ radv_export_param(ctx, param_count, values, 0xf);
outinfo->vs_output_param_offset[VARYING_SLOT_PRIMITIVE_ID] = param_count++;
outinfo->export_prim_id = true;