projects / platform / upstream / llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 20c873c) | patch
AMDGPU: More mad_64_32 test cases for multiple uses
authorNicolai Hähnle <nicolai.haehnle@amd.com>
Tue, 19 Apr 2022 22:34:23 +0000 (17:34 -0500)
committerNicolai Hähnle <nicolai.haehnle@amd.com>
Tue, 19 Apr 2022 23:00:05 +0000 (18:00 -0500)
commitb39d34de5e6b05012989e96d83fd3267d74db4cb
treeba92c3e6569f0b95f644b01a909d00bbe30fab73tree | snapshot
parent20c873c12fa5563bbfb07c880db51fa68db571cccommit | diff
AMDGPU: More mad_64_32 test cases for multiple uses

Also use gfx90a for the gfx9 test, whose code gen should be affected by
faster multiply-add instructions.
llvm/test/CodeGen/AMDGPU/mad_64_32.ll diff | blob | history
Domain: System / Toolchain;