aco: CSE readlane/readfirstlane/permute/reduce with the same exec mask
authorRhys Perry <pendingchaos02@gmail.com>
Mon, 23 Sep 2019 13:31:24 +0000 (14:31 +0100)
committerRhys Perry <pendingchaos02@gmail.com>
Thu, 26 Sep 2019 12:19:51 +0000 (13:19 +0100)
commit06ea3325c32025a1adc36adb84276ba737e63acd
tree3db63cbcb3cfb1c2890fa1d3165cbfa998dc57ab
parent86ecf92c23ccd4a5493a27104e9c85b8fa41f3a3
aco: CSE readlane/readfirstlane/permute/reduce with the same exec mask

v2: rename pass_temp to pass_flags
v2: also CSE reductions
v3: add ds_swizzle_b32 support
v3: check gds/offset0/offset1 fields

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
src/amd/compiler/aco_ir.h
src/amd/compiler/aco_opt_value_numbering.cpp