Mark Dickinson <dicki...@gmail.com> added the comment: One more lot of timings, from Trevor's pow_benchmark.txt:
Unpatched --------- 1024 bits: 0.008256 2048 bits: 0.052324 3072 bits: 0.159689 4096 bits: 0.357264 Patched (percent speedup) ------- 1024 bits: 0.006576 (+25.5%) 2048 bits: 0.045878 (+14.1%) 3072 bits: 0.135740 (+17.6%) 4096 bits: 0.310756 (+15.0%) I'm not quite sure why I'm not seeing the same level of speedup that Trevor originally reported. Perhaps this a result of some of the other optimizations that have been applied to the long implementation (30-bit digits, a reworked division algorithm), or perhaps the same level of speedup isn't available on 64-bit machines for some reason. I hope I haven't somehow negated the effects of the patch in my refactoring. ---------- _______________________________________ Python tracker <rep...@bugs.python.org> <http://bugs.python.org/issue936813> _______________________________________ _______________________________________________ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
