Having min and max on floats is useful for lowering to special
hardware intrinsics.
Details
Details
- Reviewers
ftynse mehdi_amini
Diff Detail
Diff Detail
- Repository
- rG LLVM Github Monorepo
Event Timeline
Comment Actions
Please document and test the semantics of these operations in presence of NaNs, and update docs/Rationale.md to mention why we have minf/maxf (presumably because they offer NaN-propagation semantics that cannot be expressed with a single cmpf+select).
Comment Actions
I think the intended semantics is to propagate NaNs, which can be implemented as various combinations of 4 operations that becomes less straightforward to match.