Accuracy of Mathematical Functions

[H. Peter Anvin]

On 2026-02-13 03:09, Paul Zimmermann wrote:
>        Hi,
> 
> a new update of our comparison is available:
> 
> https://members.loria.fr/PZimmermann/papers/accuracy.pdf
> 
> Main changes in this update (with GNU libc 2.43):
> 
> * GNU libc now provides seven correctly-rounded functions in double
>   precision (acosh, asinh, atanh, erf, erfc, lgamma, tgamma)
> 
> * thanks to Katsuhisa Ozaki, we have updated to MSVC 2026, where several
>   issues reported against MSVC 2022 have been fixed
> 
> * in double precision, for all C99 univariate functions, at least one library
>   now provides a correctly-rounded version, except atan, cosh, sinh, tanh
> 

Hi!

This is a great document for a whole bunch of reasons.

One thing I would personally love to see added is strto{f,d,ld}() --
converting decimal strings to binary floating-point is notoriously complex to
do with proper rounding, but there are algorithms out there now that will do
so correctly without requiring ridiculous amount of computation, and there
really is no good reason not to take advantage of that.

	-hpa