On Fri, Mar 2, 2012 at 11:59, Martin Albrecht <martinralbre...@googlemail.com> wrote: > Hi there, > > I am going to give a talk about Sage at the ECrypt PhD Summer School on Tools > > https://www.cosic.esat.kuleuven.be/ecrypt/courses/mykonos12/index.shtml > > For that I tried to compile an overview on where we stand in terms of linear > algebra. I figured this should be of more general interest and perhaps > eventually go into the reference manual. Hence, I started a wiki page > > http://wiki.sagemath.org/DenseLinearAlgebra > > which collects this information. I am not done yet filling it in, but I have > to catch a train now. Meaning: knock yourselves out adding/filling in your > favourite base ring :)
I might as well mention here that FLINT 2.3 provides direct support for matrices over Z[x] and (Z/pZ)[x] (for p < 2^64), as well as (Z/pZ) for 2^22 < p < 2^64. Most of this isn't really fully optimized (FLINT does not even use BLAS yet), but it should be very easy to wrap and will provide significant speedups over whatever generic things Sage does right now. The last time I did some benchmarking, FLINT was also typically faster than Sage on various matrix operations over Z and Q, including multiplication, determinants, and nonsingular solving (until IML took the lead at n = 1000 or so...), but Sage probably performs better now that it makes proper use of Linbox for the base operations over finite fields. Still, FLINT is currently likely to be faster than Sage or any of its libraries for smallish matrices over Z and Q (or asymptotically in rare places where Sage happens to do something stupid), so it will probably be worth at least wrapping as an alternative algorithm for a few operations. Fredrik -- 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
