fma:Exception "Underflow" not set

Wilco Dijkstra Wilco.Dijkstra@arm.com
Mon Apr 7 14:19:39 GMT 2025


Hi,

> "I know x86 and aarch64 are underflows in this situation, but I don't
> think this complies with IEEE Std 754 ™- 2019. There may be evidence
> to refute my calculation process.Thanks."

Lookup "tininess before rounding" - IEEE allows implementations to use this
option, so underflow is correct for this case.

Cheers,
Wilco


More information about the Libc-alpha mailing list