On Wed, Aug 13, 2014 at 3:16 PM, Marc Masdeu <marc.mas...@gmail.com> wrote: > > On Wed, Aug 13, 2014 at 2:08 PM, William A Stein <wst...@uw.edu> wrote: >> >> >> >> On Wednesday, August 13, 2014, Marc Masdeu <marc.mas...@gmail.com> wrote: >>> >>> >>> >>> On Tuesday, August 12, 2014 11:17:55 PM UTC+1, Fredrik Johansson wrote: >>>> >>>> On Tuesday, August 12, 2014 8:06:02 PM UTC+2, wstein wrote: >>>>> >>>>> Hi -- Another question. You just deleted this [1] below -- does flint >>>>> really solidly beat it? >>>> >>>> >>>> FLINT uses the same formula for 4x4 determinants, so the difference >>>> should be negligible (just the difference in overhead between the mpz and >>>> fmpz types). >>>> >>> I'm afraid that the overhead cannot be avoided in the current >>> implementation, one would need repeated calls to fmpz_mat_entry() anyways. >>> By the way, the new proposed implementation (purely FLINT) solidly beats >>> Magma (I am trying random matrices with entries of having 10^6 decimal >>> digits). And for those matrices it takes 0.8s to compute one such >>> determinant in my laptop... It takes the same in the Sage current. >>> >>> -- >> >> >> Million digit entries is one case that is prob not so important in the >> context of quaternion algebras - another case to test would be smaller >> entries, 30 digits... > > > They take the same time (~50us in my laptop).
That's probably all timer overhead. Try using "%timeit mat.det(); mat._clear_cache()". Fredrik -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
