Paul.Zimmermann at loria.fr
Sat Feb 27 10:21:50 CET 2010
> Your http://gmplib.org/pi-with-gmp.html doc
> has a paragraph that states to "Compare the numbers":
> "Many π programs proclaim themselves as "the fastest", but then they are
> actually several times slower than gmp-chudnovsky.c
> with the current GMP release. Compare the numbers!"
> There are however no numbers in the table to make any comparisons, other
> than across GMP versions, to.
a comparison was made by Fabrice Bellard with his own implementation TPI
on http://misn.loria.fr/fichiers/bellard-20100209.pdf, slide 34 (it is in
french, but you should understand the table).
His TPI code is available at http://bellard.org/pi/pi2700e9/tpi.html.
More information about the gmp-discuss