As far as I know there can't be anything in MPIR 1.2 that is slower than MPIR 1.1, but plenty that is substantially faster.
These timings look suspiciously close, but I don't think they should be. Perhaps we are wrong and the time complexity of this benchmark doesn't depend much on MPIR at all, but other things? Or is it possible we've picked up a system wide GMP by accident when linking in this benchmark? Even untuned the FFT should be significantly faster in MPIR 1.2, so I doubt tuning will be the issue. How repeatable are these timings? And can you check with ldd which library it is actually linking against, just to make sure. Bill. On 11 June, 16:20, Jeff Gilchrist <[email protected]> wrote: > On Sat, Jun 6, 2009 at 8:57 PM, Bill Hart<[email protected]> wrote: > > The MPIR team is proud to release MPIR 1.2, which is available on our > > website: > > >http://www.mpir.org/ > > The new features of MPIR 1.2 include: > > > * Further improvements to assembly support for K8/K10/Core2/Pentium 4 > > including improvements to multiplication and division by 2 limbs > > So I had a chance to do some benchmarks, this is using the BPSW code > which is part of the new stuff added to Brian's benchmark. It seems > that 1.2 is slower than 1.1 was on my K8 system. Neither was tuned so > maybe the default parameters are worse off now for my system? > > Testing all base 2 pseudoprimes from 10^15 to 10^16 with BPSW took: > MPIR 1.1.1. = 1m26.068s > MPIR 1.2 = 1m32.788s > GMP 4.3.1 = 1m34.058s > > I guess I could try tuning and see if that helps. > > Jeff. --~--~---------~--~----~------------~-------~--~----~ 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 -~----------~----~----~----~------~----~------~--~---
