Valery Pykhtin
b8025d1482
Reapply "[AMDGPU] Add InstCombine rule for ballot.i64 intrinsic in wave32 mode." ( #80303 )
...
Reapply #71556 with added lit test constraint: `REQUIRES: amdgpu-registered-target`.
This reverts commit 9791e5414960f92396582b9e9ee503ac15799312.
2024-02-02 13:09:25 +01:00
Valery Pykhtin
9791e54149
Revert "[AMDGPU] Add InstCombine rule for ballot.i64 intrinsic in wave32 mode." ( #78429 )
...
Reverts llvm/llvm-project#71556
Fixes failures:
https://lab.llvm.org/buildbot/#/builders/188/builds/40541
https://lab.llvm.org/buildbot/#/builders/91/builds/21847
https://lab.llvm.org/buildbot/#/builders/98/builds/31671
https://lab.llvm.org/buildbot/#/builders/139/builds/57289
2024-01-17 14:12:07 +01:00
Valery Pykhtin
57b50ef017
[AMDGPU] Add InstCombine rule for ballot.i64 intrinsic in wave32 mode. ( #71556 )
...
Substitute with zero-extended to i64 ballot.i32 intrinsic.
2024-01-17 17:02:05 +07:00
Rana Pratap Reddy
13ea1146a7
[AMDGPU] Lower __builtin_amdgcn_read_exec_hi to use amdgcn_ballot ( #69567 )
...
Currently __builtin_amdgcn_read_exec_hi lowers to llvm.read_register,
this patch lowers it to use amdgcn_ballot.
2023-10-26 10:26:11 +05:30
Matt Arsenault
f4bcd7f598
AMDGPU/clang: Add builtins for llvm.amdgcn.ballot
...
Use explicit _w32/_w64 suffixes for the wave size to be consistent
with the existing other wave dependent intrinsics. Also start
diagnosing trying to use both wave32 and wave64.
I would have preferred to avoid the +wavefrontsize64 spam on targets
where that's the only option, but avoiding this seems to be more work
than I expected.
2022-12-29 17:58:55 -05:00
