> > This is interesting. One thing we discovered with Maple, which I think > is known by others, is that when degree drops occur in the modular > computations, you can stop F4 and output the new polynomials that have > lower than expected degree, together with a flag saying the computation is > incomplete. An outer loop computing over Q can focus on reconstructing > those polynomials, which also tend to be small, before restarting with them > in the generating set. This strategy was effective for cyclic-n, however, > we found other systems where the degree fluctuates rapidly over hundreds of > steps. >
On Friday, March 29, 2019 at 4:56:41 AM UTC-7, parisse wrote: > > Thanks to W. Stein, I could run Giac on cocalc server during the last 3 > months, improve the Groebner basis source code, and eventually I could > solve the huge cyclic10 benchmark on Q : 2225 primes were required for > Chinese remaindering, 217G of memory, and more than 200 sequential days > computation (10 days real time). > For more details, see this report > <https://hal.archives-ouvertes.fr/hal-02081648> > Giac source code: > https://www-fourier.ujf-grenoble.fr/~parisse/debian/dists/stable/main/source/giac_1.5.0-49.tar.gz > -- 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.
