| Issue |
208437
|
| Summary |
Teach computeKnownFPClass and other ValueTracking APIs about llvm.convert.from.arbitrary.fp
|
| Labels |
good first issue,
missed-optimization,
floating-point
|
| Assignees |
|
| Reporter |
arsenm
|
The recently added [arbitrary FP intrinsics](https://llvm.org/docs/LangRef.html#llvm-convert-from-arbitrary-fp-intrinsic) are missing handling in analysis APIs. In particular, computeKnownFPClass can report quite a bit of information based solely on the exponent range of the source format type, and if it supports inf/nan.
We probably could do something with `llvm.convert.to.arbitrary.fp` but it's less obviously useful.
![]()
_______________________________________________
llvm-bugs mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
