On Wed, Oct 22, 2008 at 3:47 PM, john_perry_usm <[EMAIL PROTECTED]> wrote: > >> Is it still true that Perry is working on putting a version in Singular? > > I personally am not writing the code. I did offer, but Christian Eder, > a student at the University of Kaiserslautern, has the primary > responsibility. (I had worked with him on the original toy > implementation as an interpreted library in Singular.) I understand > that the Singular team is very interested in getting this going. I am > also very interested in the completion of the basic F5, because I > would like to see how some improvements that Chris & I developed will > work in a compiled environment instead of our toy implementation. > > Every now and again I ask Chris about the progress and he says it's > coming along. I understand that he has a regular job as a > schoolteacher, and cannot work on it full time. > > regards > john perry > > PS: Martin & Simon did not mention that their toy implementations in > Sage were faster than the toy implementation in Singular.
That's what I was about to ask. Interesting! How much faster? William > >> Even so, if someone improves the cython version it seems >> possible that it could become very competitive. >> >> -M. Hampton >> >> On Oct 22, 11:26 am, "David Joyner" <[EMAIL PROTECTED]> wrote: >> >> > On Wed, Oct 22, 2008 at 12:53 PM, Martin Albrecht >> >> > <[EMAIL PROTECTED]> wrote: >> >> > > On Wednesday 22 October 2008, Simon King wrote: >> > >> Dear Team, >> >> > >> at SD 10, Martin Albrecht and I implemented the F5 algorithm according >> > >> to John Perry's pseudocode. The two implementations are at >> > >>http://wiki.sagemath.org/days10/CodingSprint >> > >> attachment f5.py (Martin's pure python implementation) respectively >> > >> f5.pyx (my cython implementation). >> >> > >> These are only toy implementations that clearly can't compete with >> > >> Singular: >> > >> f5.pyx f5.py >> > >> Cyclic-6 8.78s 22.44s >> > >> Katsura-5 lex 93.85s 428.52s >> > >> Katsura-7 degrevlex 4.21s 7.86s >> >> > > btw. Singular: 0.3, 0.02, 0.35 >> >> > >> Nevertheless: There already is a toy implementation of Buchberger's >> > >> algorithm in Sage. So, do you think the toy-F5 shall be included as >> > >> well? >> >> > > The question is: What purpose would such an implementation have: >> > > (a) educational (i.e. quite read-able/hack-able code) >> > > (b) coverage (i.e. provide GB calculations for fields Singular doesn't >> > > support) >> >> > > The current toy Buchberger provides both (which is probably >> > > unfortunate). If >> > > (a) is the focus then I guess my code is more suitable while the above >> > > timing >> > > suggest to use Simon's code if (b) is desired. >> >> > My vote would be for both to be included and SImon's code to be attached >> > (ie, the default). But maybe f5.py could use a bit more documentation >> > in some parts >> > (since it is for educational uses)? >> >> > > Cheers, >> > > Martin >> >> > > PS: I suspect that there is some memleak in my code, contributing to the >> > > exceptionally poor performance. I never got around checking this. >> >> > > -- >> > > name: Martin Albrecht >> > > _pgp:http://pgp.mit.edu:11371/pks/lookup?op=get&search=0x8EF0DC99 >> > > _www:http://www.informatik.uni-bremen.de/~malb >> > > _jab: [EMAIL PROTECTED] > > > -- William Stein Associate Professor of Mathematics University of Washington http://wstein.org --~--~---------~--~----~------------~-------~--~----~ 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://www.sagemath.org -~----------~----~----~----~------~----~------~--~---
