Dear all, I've now attached a patch for this to track ticket #7857. This includes using the Henrici (or Henrici--Brown) algorithms for the arithmetic operations +, -, *, and /, as well as for computing the (univariate) derivative in the case where applicable. If someone could please review this, that would be great.
Sebastian On Jan 5, 12:57 pm, Burcin Erocal <bur...@erocal.org> wrote: > Hi Sebastian, > > On Tue, 5 Jan 2010 04:10:08 -0800 (PST) > > Sebastian Pancratz <s...@pancratz.org> wrote: > > Dear all, > > > When looking at trac ticket #7730 (which essentially ends with saying > > that GCD computations over multivariate polynomial rings are slow), I > > noticed that the current implementation of multiplication in fraction > > fields computes the product (a/b) * (c/d) by first computing a*c and > > b*d, and then dividing out common factors (in the case of exact > > underlying rings). > > > I am wondering whether there is a good reason for doing this. > > I believe the reason is that no one implemented anything else. > > When I read the paragraph above, I couldn't believe we still used naive > multiplication, especially since I implemented the Henrici > multiplication ages ago. Apparently, I never submitted the patch, > waiting to include other improvements with it indefinitely. > > I found my old patch, which had a modification date of "Wed Apr 09 > 09:24:45 2008", and rebased it to 4.3. The new version is here: > > http://sage.math.washington.edu/home/burcin/henrici_mul.patch > > It probably needs more work to add more doctests, etc. > > Note that you can also do a similar trick for derivatives. It would be > great if you implemented that as well. > > Can you handle the rest of the submission & review process? I am very > busy with thesis related work these days. > > Thank you. > > Cheers, > Burcin
-- To post to this group, send an email to sage-devel@googlegroups.com To unsubscribe from this group, send an email to sage-devel+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URL: http://www.sagemath.org
