accuracy of double precision functions (binary64)

[zimmermann6]

I am about to release a new update of the accuracy comparison of mathematical functions. It will be with Openlibm 0.7.5, the last release. The previous update (with Openlibm 0.7.4) is available here: https://members.loria.fr/PZimmermann/papers/accuracy.pdf. For double precision, in 0.7.5, there is still a large maximal error for pow (636 ulps) and tgamma (1.03e3 ulps) and huge errors for j0,j1,y0,y1 and lgamma. All other functions have a maximal known error less than 4.04 ulps.

[zimmermann6]

a new version is available, with OpenLibm 0.8.1: https://members.loria.fr/PZimmermann/papers/accuracy.pdf. No major change with respect to the above, except the maximal known error for "other functions" is now 4.08 ulps.

[zimmermann6]

a new version is available, still with OpenLibm 0.8.1: https://members.loria.fr/PZimmermann/papers/accuracy.pdf.
The main change is that the LLVM library now implements a few correctly-rounded routines in single precision (log, log10, log2, hypot), and hypot in double precision.

[zimmermann6]

a new version is available, still with OpenLibm 0.8.1: https://members.loria.fr/PZimmermann/papers/accuracy.pdf.
The main change is that the LLVM library now implements more correctly-rounded routines in single precision.

[zimmermann6]

a new version is available, still with OpenLibm 0.8.1: https://members.loria.fr/PZimmermann/papers/accuracy.pdf.
This is a new version with updated versions of the different libraries and:

• the Microsoft math library is now included (it was clearly missing)
• we also added the FreeBSD math library
• and we added some new C23 functions (acospi, cospi, ...) which are available
  in the Intel and FreeBSD math libraries

[zimmermann6]

a new version is available, still with OpenLibm 0.8.1: https://members.loria.fr/PZimmermann/papers/accuracy.pdf.
This is a new version with updated versions of the different libraries and the ARM Performance Library.

[zimmermann6]

a new version is available, with OpenLibm 0.8.3: https://members.loria.fr/PZimmermann/papers/accuracy.pdf.
Remaining issues in OpenLibm:

• large error for pow in double precision
• powl is not thread-safe
• huge error for tgammal for tiny x
• expm1l yields NaN instead of a normal value near the overflow threshold