On Thursday 07 April 2011, mario wrote: > > From: Mateusz Paprocki <> > > > > > > In case people are curious, Sage (because of Singular!) takes 0.07 > > > > seconds to do the benchmark that Sympy takes 11 seconds to do at the > > > > end of the Sympy talk: http://flask.sagenb.org/home/pub/16/ > > > > > > Seems like some room for improvement. > > > > > > Mateusz, do you think this is just because Python is slower than C, or > > > because Singular implements a better algorithm? > > > > I'm sure that Singular implements much better algorithm(s) in this > > case. The implementation (of Buchberger's algorithm) that we have > > currently in SymPy is something more than a toy, but much less than a > > tool for solving real life problems. Implementing more (or better) > > reduction criteria would definitively help here. Groebner walk, F4 or > > F5 would be also a huge improvement (we have a GSoC prospective > > student willing to work on F5B, among other things). Also polynomial > > representation we use in groebner() is suboptimal (we use tuples for > > storing exponents instead of packing exponents into integers). So yes, > > there is a lot of room for improvements. > > In Python, using packed exponents and Buchberger's algorithm this > example takes 0.37s in rmpoly, see > example in http://code.google.com/p/rmpoly/wiki/Tutorial
Hi, first: nice slides! However, it might be worth mentioning that none of the packages benchmarked in the slides is generally considered to be very good at Gröbner basis computations. To get a real sense of performance you should take a much much bigger example + compare with Magma, Maple, Singular (and Macaulay2, CoCoa) all of which have decent GB implementations. Cheers, Martin -- name: Martin Albrecht _pgp: http://pgp.mit.edu:11371/pks/lookup?op=get&search=0x8EF0DC99 _otr: 47F43D1A 5D68C36F 468BAEBA 640E8856 D7951CCF _www: http://martinralbrecht.wordpress.com/ _jab: [email protected] -- To post to this group, send an email to [email protected] To unsubscribe from this group, send an email to [email protected] For more options, visit this group at http://groups.google.com/group/sage-devel URL: http://www.sagemath.org
