[RFC] x86_64: Simplify powl computation for small integral y [BZ #33411]

[Siddhesh Poyarekar]

On 2025-10-11 01:28, Paul Zimmermann wrote:
>         Hi Siddhesh,
> 
> thank you for looking at this.
> 
>> This does not introduce any regressions other than the following:
>>
>> testing long double (without inline functions)
>> Failure: Test: pown_upward (-0x1p+8192, 3LL)
>> Result:
>>   is:         -inf  -inf
>>   should be:  -1.18973149535723176502e+4932  -0xf.fffffffffffffff0p+16380
>> Failure: Test: pown_upward (-0x1p-8192, -3LL)
> 
> since the result overflows towards -inf and rounding is upward, the result
> should be the smallest regular number, not -inf. Thus it is indeed a regression.

Thanks for confirming.  Let me see how I can deal with this.

Sid