Where are the algorithms behind these impressive speed-ups documented -- or is it done by clever implementation? Either way they should be published for posterity, or at least documented !
John On 9/12/07, William Stein <[EMAIL PROTECTED]> wrote: > > On 9/12/07, mabshoff <[EMAIL PROTECTED]> wrote: > > > > > > Kate Minola wrote: > > > FYI - the following have recently > > > been released: > > > > > > gmp-4.2.2 > > > mpfr-2.3.0 > > > ecm-6.1.3 > > > > > > Kate > > > > Hello Kate, > > > > #541, #542 and #642 in sagetrac respectively. I have to admit that I > > added #642 only two hours ago. I expect that I will spend some time on > > those tickets during Bug Day 3 (probably the 20th of September) > > Yep. However, one thing I want to add is that my understanding > is that the GPL-only patches that make xgcd/gcd *vastly* faster > for large numbers might only work with gmp-4.2.1, at least > without a lot more work to port it, and if so we should wait > for that patch to be updated (especially if gmp-4.2.2 is merely > a build improvement over gmp-4.2.1). This sort of thing is > just really cool: > > BEFORE the patch: > sage: k=2000000; n = QQ.random_element(2^k); m=QQ.random_element(2^k) > sage: time a=n+m > CPU times: user 5.69 s, sys: 0.24 s, total: 5.93 s > > AFTER the patch: > sage: k=2000000; n = QQ.random_element(2^k); m=QQ.random_element(2^k) > sage: sage: time a=n+m > CPU times: user 0.99 s, sys: 0.00 s, total: 0.99 s > sage: len(str(n)) > 1204120 > > Having fast arithmetic with million digits rational numbers can be really > important... > > Presumably mpfr and ecm would both work with gmp-4.2.1 if we have > to stay with it for some amount of time. > > -- William > > > > -- John Cremona --~--~---------~--~----~------------~-------~--~----~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~----------~----~----~----~------~----~------~--~---
