| Issue |
107607
|
| Summary |
[RISCV] Miscompilation on rv64
|
| Labels |
new issue
|
| Assignees |
|
| Reporter |
mikhailramalho
|
The following program miscompiles when using clang++ for rv64:
```
$ cat foo.cpp
#include <cmath>
int main() {
long double y = NAN;
_Float16 a = static_cast<_Float16>(y);
return (a != a);
}
```
The cast of `NAN` from `long double` to `_Float16` generates `0`, instead of `NAN`.
It only happens with clang++ in rv64. It works just fine in intel 64 and rv32. Compiling with g++ for rv64 also works.
![]()
_______________________________________________
llvm-bugs mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
