On Sat, Feb 28, 2009 at 12:54 PM, <ja...@njkfrudils.plus.com> wrote:
> This could make a nice "real world" benchmark for MPIR , where we have cache > misses and branch mispredictions. The current mpirbench and speed don't have > these features. For example if we pipeline the entire mul_basecase to > increase the speed , it turns out to be 1.5x larger in the code cache. Now > speed and mpirbench would probably not show any indication of this, whereas > in the real world it may be a slowdown. Sounds like a good idea to me. I finally got around to doing some tests with MPIR 0.9 on Windows, both with cygwin and MSVC so I will post my results. It seems that what I really want to see for Windows is the MPIR 1.0 release with all of Brian's additions since the current version is essentially the same speed as GMP. Benchmark: time echo "1606274791359742683462193665835026393961358862962351944720191458360082027218282749671391672280676891140068379759963605046949616963145972154002637700442143987397361301827559875507425065995179598763956235845311986100972998806344397600124996519181" | ./ecm -sigma 2956027510 11e6 Results: AMD Opteron 248 @ 2.2GHz (Linux 64bit) MPIR 0.9 (gcc 3.4.6) = 2m32.406s GMP 4.2.4 (gcc 3.4.6) = 3m20.908s Intel Xeon E5405 @ 2.00GHz (Linux 64bit) MPIR 0.9 (gcc 4.1.2) = 3m45.447s GMP 4.2.4 (gcc 4.1.2) = 4m40.356s Intel Core2 Q9550 @ 3.4GHz (Vista 64bit) GMP 4.2.4 (MSVC 64bit) = 2m06.904s MPIR 0.9 (MSVC 64bit) = 2m07.032s MPIR 0.9 (cyg gcc 3.4.4) = 5m45.800s GMP 4.2.4 (cyg gcc 3.4.4) = 6m25.568s GMP 4.2.4 (MSVC 32bit) = 6m45.073s MPIR 0.9 (MSVC 32bit) = 6m50.230s One thing I noticed is that building MPIR in Linux you get libgmp.a and gmp.h files out. With Windows however, MSVC is setup to create mpir.lib and mpir.h which isn't consistent and I had to rename those files to get them to work with GMP-ECM or any other application that uses GMP. So whatever is decided is it possible to have both use the same naming convention? Regards, Jeff. --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "mpir-devel" group. To post to this group, send email to mpir-devel@googlegroups.com To unsubscribe from this group, send email to mpir-devel+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/mpir-devel?hl=en -~----------~----~----~----~------~----~------~--~---