FLINT is faster of course! (Well for 1d anyway, :-) ) But seriously, Bernard's original contention that giac is within 1.5 times what Magma will do is about right, and for small problems giac is actually faster! I haven't done timings for Magma's Z[x1, ..., xn] GCD yet, only the Z/pZ[x1, ...., xn]. Obviously with asymptotically fast algorithms, giac would be winning outright.
Here I repeat the times: Magma: mod-1d: 0.00067s, 0.00981s, 0.0244s mod-2d: 0.00089s, 0.00374s, 0.0256s, 0.08549s mod-3d: 0.00185s, 0.0069s, 0.05491s mod-4d: 0.006s, 0.028s, 0.071s giac: mod-1d: 0.00038, 0.02, 0.026 mod-2d: 0.00052, 0.0024, 0.03, 0.15 (0.112) mod-3d: 0.004, 0.0085, 0.22 (0.198) mod-4d: 0.012, 0.048, 0.12 (0.12) I'll eventually do timings for polynomials over the integers for Magma as well. I should also give Magma the benefit of running the timings more than once and taking minimum times. I'll get to it eventually and produce a nice table. BTW I just got some kind of multivariable polynomial multiplication working in FLINT. Only another year or two before I get m.v. GCD working. Yay!! Bill. On 25 Jan, 20:33, Ondrej Certik <ond...@certik.cz> wrote: > On Sun, Jan 25, 2009 at 12:27 PM, Bill Hart <goodwillh...@googlemail.com> > wrote: > > > Thanks for the executable. It worked perfectly. Here are the times > > that it reports on the 2.4Ghz Opteron: > > > 1d: 0.00034, 0.06, 0.085 > > 2d: 0.0011, 0.0046, 0.048, 0.2 > > 3d: 0.014, 0.15, 0.63 > > 4d: 0.016, 0.07, 0.18, 1.03 > > > mod-1d: 0.00038, 0.02, 0.026 > > mod-2d: 0.00052, 0.0024, 0.03, 0.15 (0.112) > > mod-3d: 0.004, 0.0085, 0.22 (0.198) > > mod-4d: 0.012, 0.048, 0.12 (0.12) > > > (The ones in brackets are the results from timing that last > > computation five times and dividing the total time by 5). > > > There's quite a lot of timing irregularity on this machine (nothing to > > do with your code), so I ran the timings multiple times and reported > > the best timings. > > So what is faster, magma or giac? I am a bit lost in those numbers. > > Ondrej --~--~---------~--~----~------------~-------~--~----~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to sage-devel-unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://www.sagemath.org -~----------~----~----~----~------~----~------~--~---
