[PATCH] math: Add fast-path to fma

[Wilco Dijkstra]

Hi Adhemerval,

On 09/03/26 18:21, Florian Weimer wrote:
> * Adhemerval Zanella:
>
>> For normal numbers there is no need to issue scalbn, the fma can set
>> the exponend directly.   Performance-wise on x86_64-linux-gnu without
>> multi-arch it shows a latency improvement of ~5% and throughput of %7
>> (and sligth more for ABIs witht tail-call optimization).
>
> typos: slight[ly], with[]
>
> Isn't x86-64 an ABI with tail-call optimization?  So I don't
> understand the parenthetical remark.

> Ooops, it should be "and slightly mode for ABIs *without* tail-call optimization".

LGTM with description fixed.

Reviewed-by: Wilco Dijkstra  <Wilco.Dijkstra@arm.com>

Cheers,
Wilco