ttn: Use 1-bit NIR comparison opcodes
authorConnor Abbott <cwabbott0@gmail.com>
Fri, 2 Aug 2019 13:00:30 +0000 (15:00 +0200)
committerMarek Olšák <marek.olsak@amd.com>
Tue, 6 Aug 2019 22:03:19 +0000 (18:03 -0400)
commit4b10949482d18f5df1e35f9711d32b4b9428c954
tree94edfcd257e7a6a119b3735e88686a1a7fd5a6f5
parente7fd90e8ef6dc9cf12108f53409fb71f0133cb01
ttn: Use 1-bit NIR comparison opcodes

We shouldn't be using the versions that output a 32-bit boolean, since
nir_opt_algebraic won't optimize them as well. Drivers will lower these
to the 32-bit versions after optimizing, if appropriate. Also, this will
make implementing 64-bit comparisons easier.

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
src/gallium/auxiliary/nir/tgsi_to_nir.c