projects / platform / upstream / llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: b90ebbc) | patch
AMDGPU: Convert a test to generated checks
authorMatt Arsenault <Matthew.Arsenault@amd.com>
Fri, 25 Nov 2022 21:52:01 +0000 (16:52 -0500)
committerMatt Arsenault <Matthew.Arsenault@amd.com>
Mon, 5 Dec 2022 18:00:12 +0000 (13:00 -0500)
commit32975934f1106ef4b24b71f9e03ff44dd1eb9fa6
tree1a24153a0755773d5e8b47912f647272d03c9bcftree | snapshot
parentb90ebbc5bd6d90a7dab405004545865d39840277commit | diff
AMDGPU: Convert a test to generated checks

Also switch to not using kernels to test, since it shouldn't matter
for these addressing modes. This should reduce some of the diffs
between subtargets, but it's still not enough to share the same
checks.

The test could use some additional modernization. Probably should also
round this out with the new maximium offsets.
llvm/test/CodeGen/AMDGPU/cgp-addressing-modes-flat.ll diff | blob | history
Domain: System / Toolchain;