| Issue |
242935
|
| Summary |
[KnownFPClass] Refine `KnownFPClass::fadd`/`KnownFPClass::fsub`
|
| Labels |
floating-point
|
| Assignees |
ZERICO2005
|
| Reporter |
ZERICO2005
|
Tracking issue for the types of refinements we can make for `fadd`/`fsub`/`fadd_self` without the magnitude refinement for `KnownFPClass` https://github.com/llvm/llvm-project/issues/240038
- [ ] Rule out infinity for common cases https://github.com/llvm/llvm-project/pull/239064
- [ ] Rule out infinity for subnormal special cases for `Dynamic` https://github.com/llvm/llvm-project/pull/239064
- [ ] Rule out infinity for subnormal special cases for `NearestTiesToEven` (I have a local branch for this)
- [ ] Rule out infinity for `TowardZero`, `TowardPositive`, `TowardNegative` for `fadd`/`fsub` (I have a local branch for this)
- [ ] Rule out infinity for `TowardZero`, `TowardPositive`, `TowardNegative` for `fadd_self`
- [ ] Rule out infinity for non-finite special cases (`+inf + NaN` is never `+inf`) (I don't think this is worth implementing)
- [ ] Rule out cancellation to zero optimally https://github.com/llvm/llvm-project/pull/242911
- [ ] Rule out subnormal cases optimally (I have a local branch for this)
- [ ] Rule out normal for subnormal special cases (I have a local branch for this)
_______________________________________________
llvm-bugs mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs