diff options
| author | Kenneth Graunke <[email protected]> | 2015-06-10 00:52:07 -0700 |
|---|---|---|
| committer | Kenneth Graunke <[email protected]> | 2015-06-22 14:08:36 -0700 |
| commit | 1762568fd39b9be42d963d335e36daea25df7044 (patch) | |
| tree | 763dd0431662741306c078b53b11278bf81f8cf8 /src/glsl/nir | |
| parent | 94e3864707e48d4b1d5fb5f88a01370a73ddb0cb (diff) | |
nir: Allow vec2/vec3/vec4 instructions in the select peephole pass.
These are basically just moves, so they should be safe as well.
When disabling i965's GLSL IR level scalarizer (channel expressions)
pass, I started seeing NIR code like this:
if ssa_21 {
block block_1:
/* preds: block_0 */
vec4 ssa_120 = vec4 ssa_82, ssa_83, ssa_84, ssa_30
/* succs: block_3 */
} else {
block block_2:
/* preds: block_0 */
/* succs: block_3 */
}
block block_3:
/* preds: block_1 block_2 */
vec4 ssa_33 = phi block_1: ssa_120, block_2: ssa_2
Previously, the GLSL IR scalarizer pass would break the vec4 into a
series of fmovs, which were allowed by the peephole pass. But with
the vec4 operation, they were not. We want to keep getting selects.
Normal i965 on Broadwell:
instructions in affected programs: 200 -> 176 (-12.00%)
helped: 4
With brw_fs_channel_expressions() disabled:
instructions in affected programs: 1832 -> 1646 (-10.15%)
helped: 30
Signed-off-by: Kenneth Graunke <[email protected]>
Reviewed-by: Jason Ekstrand <[email protected]>
Reviewed-by: Connor Abbott <[email protected]>
Diffstat (limited to 'src/glsl/nir')
| -rw-r--r-- | src/glsl/nir/nir_opt_peephole_select.c | 4 |
1 files changed, 3 insertions, 1 deletions
diff --git a/src/glsl/nir/nir_opt_peephole_select.c b/src/glsl/nir/nir_opt_peephole_select.c index 82c65bb442f..ef7c9775aa3 100644 --- a/src/glsl/nir/nir_opt_peephole_select.c +++ b/src/glsl/nir/nir_opt_peephole_select.c @@ -86,7 +86,9 @@ block_check_for_allowed_instrs(nir_block *block) nir_alu_instr *mov = nir_instr_as_alu(instr); if (mov->op != nir_op_fmov && mov->op != nir_op_imov && mov->op != nir_op_fneg && mov->op != nir_op_ineg && - mov->op != nir_op_fabs && mov->op != nir_op_iabs) + mov->op != nir_op_fabs && mov->op != nir_op_iabs && + mov->op != nir_op_vec2 && mov->op != nir_op_vec3 && + mov->op != nir_op_vec4) return false; /* Can't handle saturate */ |