On Sun, Oct 21, 2012 at 10:15 AM, Bill Hart <[email protected]> wrote: > Hi all, > > I have just uploaded alpha4 to our website http://mpir.org/ > > This will hopefully be the last alpha. It merely adds a patch to speed > up the FFT slightly in certain instances. > > Testing can now begin in earnest. > > Bill. Tests pass on Fedora 17 x64 on an AMD K102 platform.
I've done some performance testing using gmpy/gmpy2. The minor changes to the FFT made a significant improvement to one of my benchmarks.The benchmark measures multiplication over a wide range of sizes - from 1000x1000 bit to 1273945x21954313 bit. Compared to alpha1, the performance, especially for different sized operands, improved by ~20%. That is about ~15% faster than 2.5.1. The benchmark isn't very portable since it is written in Python and uses versions of gmpy and gmpy2 linked to different versions of MPIR and/or GMP. The actual performance increase is higher than the numbers indicate since I don't subtract the overhead. I also ran the mpmath test suite. There is ~10% performance improvement in the test suit running time. Case > > -- > You received this message because you are subscribed to the Google Groups > "mpir-devel" group. > To post to this group, send email to [email protected]. > To unsubscribe from this group, send email to > [email protected]. > For more options, visit this group at > http://groups.google.com/group/mpir-devel?hl=en. > -- You received this message because you are subscribed to the Google Groups "mpir-devel" group. To post to this group, send email to [email protected]. To unsubscribe from this group, send email to [email protected]. For more options, visit this group at http://groups.google.com/group/mpir-devel?hl=en.
