What about homogeneous cyclic-8? I'm not sure it will be any better; I'm just curious.
I do know Singular is working on improving aspects of the sba() implementation, and I'm a bit surprised it's that slow. On Saturday, May 7, 2016 at 1:32:36 AM UTC-5, parisse wrote: > > > > Le samedi 7 mai 2016 07:30:42 UTC+2, john_perry_usm a écrit : >> >> I'm sorry. I got the name mixed up; the function you want to look at is >> sba(), not dstd() (which is something experimental of mine that never saw >> the light of day). >> >> http://www.singular.uni-kl.de/Manual/latest/sing_391.htm#SEC430 >> >> Because my copy of Singular is a little... "tinkered with" I'd have to >> redownload and recompile Singular, but replace std() with sba() in your >> example and see how the timings compare, but for example when I run your >> example the timing my version reports falls from 169 to 114. -- Again, I >> tinkered with mine, so the timings won't be reflective of actual >> performance. >> >> > sba is indeed a little faster for cyclic7 in Z/pZ (25%) but not for > cyclic8 in Z/pZ (10 minutes here compared to groebner 40s, giac 3.3s) and > it eats a lot of memory (1.5G, compared to 45M for giac) > //timer=1; > system("--ticks-per-sec",100); > //option(prot); > ring r=16777213,(x1,x2,x3,x4,x5,x6,x7,x8),dp; > ideal k=x1 + x2 + x3 + x4 + x5 + x6 + x7 + x8, x1*x2 + x2*x3 + x3*x4 + > x4*x5 + x5*x6 + x6*x7 + x1*x8 + x7*x8, x1*x2*x3 + x2*x3*x4 + x3*x4*x5 + > x4*x5*x6 + x5*x6*x7 + x1*x2*x8 + x1*x7*x8 + x6*x7*x8, x1*x2*x3*x4 + > x2*x3*x4*x5 + x3*x4*x5*x6 + x4*x5*x6*x7 + x1*x2*x3*x8 + x1*x2*x7*x8 + > x1*x6*x7*x8 + x5*x6*x7*x8, x1*x2*x3*x4*x5 + x2*x3*x4*x5*x6 + x3*x4*x5*x6*x7 > + x1*x2*x3*x4*x8 + x1*x2*x3*x7*x8 + x1*x2*x6*x7*x8 + x1*x5*x6*x7*x8 + > x4*x5*x6*x7*x8, x1*x2*x3*x4*x5*x6 + x2*x3*x4*x5*x6*x7 + x1*x2*x3*x4*x5*x8 + > x1*x2*x3*x4*x7*x8 + x1*x2*x3*x6*x7*x8 + x1*x2*x5*x6*x7*x8 + > x1*x4*x5*x6*x7*x8 + x3*x4*x5*x6*x7*x8, x1*x2*x3*x4*x5*x6*x7 + > x1*x2*x3*x4*x5*x6*x8 + x1*x2*x3*x4*x5*x7*x8 + x1*x2*x3*x4*x6*x7*x8 + > x1*x2*x3*x5*x6*x7*x8 + x1*x2*x4*x5*x6*x7*x8 + x1*x3*x4*x5*x6*x7*x8 + > x2*x3*x4*x5*x6*x7*x8, x1*x2*x3*x4*x5*x6*x7*x8 - 1; > //int RT = rtimer; int T=timer; size(groebner(k)); rtimer-RT; timer-T; > int RT = rtimer; int T=timer; size(sba(k)); rtimer-RT; timer-T; > > This does not change the comparison (especially if you consider that giac > is now a little faster than in Roman reported timings and this might also > be the case for mgb and magma), singular is still much slower than the 3 > others for dense inputs. > -- 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.
