On Nov 29, 4:03 pm, Bill Hart <[EMAIL PROTECTED]> wrote:
Hi Bill, > I have just released a new bug fix for FLINT, available > athttp://www.flintlib.org/ > > This fixes the following issues: > > * A segfault in the division and pseudo division functions > > * The bound that was being used in fmpz_poly_gcd_modular was as far as > I know, not proven. I have replaced it with a proven bound and cited > the relevant paper. This makes little difference to the timings. > > * A bug in the profiling code for fmpz_poly related to the top bit of > n bit random coefficients always being set has been fixed. > > * I identified an issue that might potentially have caused wrong > results in polynomial multiplication. However I checked that in fact, > it cannot be triggered in FLINT 1.0.17 code (I also ran specifically > constructed tests to verify that this is in fact the case). The code > will be completely rewritten for FLINT 1.1. > > I recommend Sage upgrade to FLINT 1.0.17 as the first bug is critical > and can occur in live code (I actually hit it myself). Ok, Burcin did an updated spkg for FLINT 1.0.16, but it should be trivial for him to upgrade it to 1.0.17. It will miss the 3.2.1 merge window, but a quick 3.2.2 is right on its tail. > FLINT 1.0.17 will likely be the last 1.0.x series release before FLINT > 1.1 which should be available for use in Sage in the next few weeks. > The interface for FLINT 1.1 should be pretty much the same as for > FLINT 1.0.x, however there will be a large quantity of new functions > and a large number of substantial speedups which Sage will be able to > take advantage of. > > I intend to issue beta/release candidates in about 11 days. The > todo.txt file in FLINT 1.0.17 lists the tasks which need to be > completed before the release. > > Over the next few days I will be working on code which will appear in > FLINT 2.0, specifically a matrix module F_mpz_mat over the new F_mpz > integer library I have coded up for FLINT 2.0 and an F_mpz > implementation of fpLLL's "fast" module. > > I am about a third of the way through implementation of F_mpz_poly for > FLINT 2.0. It features a completely rewritten polynomial module over > the new FLINT integer format F_mpz and David Harvey's KS2 algorithm, > which gives *up to* a 30% increase in speed for polynomial > multiplication for lengths between about 45 and 9000. Nice, *really* looking forward to this :) > Bill. Cheers, Michael --~--~---------~--~----~------------~-------~--~----~ 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 -~----------~----~----~----~------~----~------~--~---
