On 2016-02-28, Simon King <simon.k...@uni-koeln.de> wrote: > I had to patch the upstream sources considerably. > - One patch prevents that MeatAxe writes multiplication tables into the > current directory. Upstream claims that setting an appropriate > environment variable would be enough for that purpose, but for that > one first needs to fix a bug. > - One patch does some rather trivial improvements to Gaussian > elimination. > - One patch implements asymptotically fast matrix multiplication > (Strassen-Winograd with a memory efficient schedule). It beats MeatAxe's > school book multiplication even for relatively small matrices. > - One patch makes it so that fast matrix multiplication is used in the > modular representation part of MeatAxe, too.
One more thing: I added proper error handling. MeatAxe upstream sources do provide hooks for callback functions, but error return values are not correctly propagated. That's fixed by one of my patches, too, so that the inversion of a singular matrix doesn't just crash but results in a ZeroDivision error. Cheers, Simon -- 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 https://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
