Unrelated, but why is up here reproducing logic that's already in APFloat?
wzssyqa219 days ago
I have no idea. I will try to fixed in future patches. #111991
wzssyqa218 days ago (edited 218 days ago)
I guess it was due to that the APFloat::minnum claims that it fellow IEEE-754 2019 minimumNumber semantics. Now, we have minimumnum, we should change that.
and fmin needs to follow the libc's behavior.
We need to fix them.
This doesn't have tests showing the evaluation, similar to those added for fmin/fmax in ec32386
wzssyqa219 days ago
Thanks. @tbaederr It seems that __builtin_fmaximum_num cannot work with -fexperimental-new-constant-interpreter.
I guess I must miss something.
Any idea?
tbaederr219 days ago
Well you didn't implement that. If you add a new file anyway, you don't have to add a run line for it.
We just introduce llvm.minimumnum and llvm.maximumnum intrinsics support to llvm. Let's support them in Clang.
See: #93033