On Mon, 2008-11-10 at 16:27 +0000, Martin Albrecht wrote:
> On Monday 10 November 2008, Thomas Kahle wrote:
> > Hi,
> >
> > here is another try.
> >
> > sage: R = QQ['a,b,c,w,x,y,z']
> > sage: (a,b,c,w,x,y,z) = R.gens()
> > sage: I = (x^5-a*b*c^3, x^7-w^5*a^5*b^5, b*c^3-a^7, \
> > b^2*a^3*c^5*x-y*z*w^2, x*y*z-w*z^2*a*b, b*x-a*w*z^9)*R
> > sage: time _ = I.radical()
> > CPU times: user 0.21 s, sys: 0.05 s, total: 0.26 s
> > Wall time: 41.03 s
> >
> > And now in Singular:
> > (Sages default termorder is degrevlex which is dp in Singular)
> >
> > So in Singular this looks like this:
> > cat singulartest.sing
> > LIB "primdec.lib";
> > ring R =  0,(a,b,c,w,x,y,z),dp;
> > ideal I = x5-abc3, x7-w5a5b5, bc3-a7, b2a3c5x-yzw2 , xyz-wz2ab, bx-awz9;
> > ideal s = std(I);
> > ideal Ir = radical(s);
> > ideal Sr = std(Ir);
> > quit;
> >
> > time Singular singulartest.sing
> > real        0m31.848s
> > user        0m31.201s
> > sys 0m0.303s
> >
> > I can hardly believe that there are 10 seconds of message passing
> > between singular and python, maybe something is misconfigured here ??
> 
> I still can't reproduce your discrepancy:
> 
> Singular:
> real    0m22.105s
> user    0m21.135s
> sys     0m0.776s
> 
> Sage;
> real    0m23.541s
> user    0m0.929s
> sys     0m0.314s
> 
> where ~ 1 second is startup time.
> 
> > > What is your machine/os? Mine:
> > > Intel Core2Duo 2.33Ghz 3GB RAM
> > > 64-bit Debian/GNU Linux
> >
> > I'm using Core2Duo 1.6Ghz 3GB, 32bit Gentoo Linux, but I'm really
> > interested in the discrepancy being around 25% in my case...
> 
> Maybe its a 32 vs. 64-bit thing? Which Sage version are you running? I'm 
> running 3.1.3 (looks I'm way behind, the current version is 3.1.4 and 3.2 is 
> in preparation)

Well, given that it works the way I want it on your machine I will
consider that a minor problem then :)
Will check 64Bit behavior later ...
The first thing I sent you was with sage 3.1.2 but in the meanwhile I
upgraded to 3.1.4 and it's the same.

Attachment: signature.asc
Description: This is a digitally signed message part

Reply via email to