[flang][OpenMP] Change the lowering of max and min reduction operator #62
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This patch changes the lowering of max and min to be lowered to arith::MaxNumFop and arith::MinNumFOp instead of using arith::MaximumFOp and arith::MinimumFOp to match the code generated in clang. The arith::MaximumFOp and arith::MinimumFOp map to the corresponding intrinsics llvm.maximum.* and llvm.minimum.* intrinsics which conform to the semantics specified in the draft of IEEE 754-2019. which is not supported by all hardware. Instead using arith::MaximumFOp and arith::MinimumFOp will allow code generation for more targets and match the code generated by clang OpenMP.
jsjodin
requested review from
skatrak,
agozillon,
DominikAdamski,
dpalermo and
TIFitis
|
This was upstreamed. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This patch changes the lowering of max and min to be lowered to arith::MaxNumFop and arith::MinNumFOp instead of using arith::MaximumFOp and arith::MinimumFOp. The arith::MaximumFOp and arith::MinimumFOp map to the corresponding intrinsics llvm.maximum.* and llvm.minimum.* intrinsics which conform to the semantics specified in the draft of IEEE 754-2019. which is not supported by all hardware. Instead using arith::MaximumFOp and arith::MinimumFOp will allow code generation for more targets and match the code generated by clang OpenMP.