This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: [PATCH 12/28] math: Use wordsize-64 version for isnan
Hi, Adhemerval, Wilco,
On Wed, Jun 12 2019, Wilco Dijkstra wrote:
>
> > My take is math wordsize-64 was added to overcome some old compiler
> > deficiencies which most likely does not show any gain nowadays. Also,
> > specifically for isnan/isinf/isfinite I think multiple possible
> > optimization is a bit of over-engineering since the libc symbols will
> > be called with recent compiler only in very specific situations (most
> > likely not performance-wise).
>
> Yes we shouldn't worry about old compilers. There is no reason for so
> many duplicated implementations. Given modern compilers inline isinf etc,
> trying to save 1 instruction in GLIBC versions makes no sense.
Makes sense.
> > What I would like to do is just uses wordsize-64 implementations as
> > default one and aim to provide specialized WORDSIZE paths if it really
> > provides a real gain.
>
> Yes that is the best way forward. We should focus optimizations on
> areas where there is clear evidence of significant gains.
>
> > So for this specific usage, isnan/isinf/isfinite, do you agree we can
> > just use wordsize-64 as default?
>
> Absolutely.
Agreed.
Thank you both for all the clarification about 32-bits targets and
optimization focus.