Post new topic Reply to topic  [ 6 posts ] 
Author Message
 Post subject: Help!! gcd procedure
PostPosted: Fri May 07, 2010 2:14 pm 

Joined: Wed May 02, 2007 12:40 pm
Posts: 5
Hi, I need to know the list of running gcd procedure, someone knows where I can find it? I looked all Singular Liberia, but I can not get it. I would appreciate your help. Thank you.


Report this post
Top
 Profile  
Reply with quote  
 Post subject: Re: Help!! gcd procedure
PostPosted: Fri May 07, 2010 4:11 pm 

Joined: Wed Mar 03, 2010 5:08 pm
Posts: 108
Location: Germany, Münster
ber wrote:
I need to know the list of running gcd procedure, someone knows where I can find it? I looked all Singular Liberies, but I can not get it.


What does 'the list of running gcd procedure' mean?

5.1.41 gcd and 5.1.28 extgcd are commands from the singular kernel.
These are not encoded in líbraries but part of the source code in (C/C++).

But you can also inspect the code of the library crypto.lib (section D.11.3 Teaching)

There you will find implementations for gcd but acting on numbers (not polynommials) only.

Quote:
D.11.3.2 exgcdN compute s,t,d such that d=gcd(a,n)=s*a+t*n
D.11.3.3 eexgcdN T with sum_i L[i]*T[i]=T[n+1]=gcd(L[1],...,L[n])
D.11.3.4 gcdN compute gcd(a,b)


C.Gorzel


Report this post
Top
 Profile  
Reply with quote  
 Post subject: Re: Help!! gcd procedure
PostPosted: Tue May 18, 2010 5:06 pm 
thanks for replying, I found in the library zeroset.lib the gcd calculation procedure called proc EGCDMain, which in turn calls the proc RemainderMain, which executes the function proc SimplifyPoly. From there I could check the calculations you need.
All this is because I want to calculate the GCD of two polynomials defined by
quotient ring. I have calculated using "reduce", but the calculation is complicated and ends with computer memory, apparently Cingular is not foreseen this problem.
Do you know why?


Report this post
Top
  
Reply with quote  
 Post subject: Re: Help!! gcd procedure
PostPosted: Tue May 18, 2010 9:07 pm 

Joined: Wed Mar 03, 2010 5:08 pm
Posts: 108
Location: Germany, Münster
Guest wrote:

All this is because I want to calculate the GCD of two polynomials defined by
quotient ring. I have calculated using "reduce", but the calculation is complicated and ends with computer memory, apparently Singular is not foreseen this problem.
Do you know why?


Note the following, as indicated in the Overwiev of zeroset.lib:
Quote:
Overview:
Algorithms for [...] factorization of univariate polynomials over Q(a)[t] where a is an algebraic number.
This library is meant as a preliminary extension of the functionality of Singular for univariate factorization
of polynomials over simple algebraic extensions in characteristic 0.


For some years, factorization over Q(a)[t] is already implemented in the kernel now.
So some commands of this library are somehow outdated/obsolete.

In any case, post your example to see what went wrong.

C.Gorzel


Report this post
Top
 Profile  
Reply with quote  
 Post subject: Re: Help!! gcd procedure
PostPosted: Wed May 19, 2010 12:32 pm 

Joined: Wed May 02, 2007 12:40 pm
Posts: 5
Hi Gorzel:
This is an example: I have programmed my own algorithm to calculate gcd remaining divisions, which is what I need and Singular that would produce the copy below.

//ejemplo_mcd
ring R=(0,s), (a,r,u(1..4)), dp;

//poly pr=r3-2r2+r-1;
//poly pr=48r3-84r2+42r-(36+s);
poly pr=-r^3+r^2-s*r+u(1)*s+u(2);

poly dividendo;
poly mcd;
poly qr=diff(pr,r);
poly divisor=pr;
poly resto=qr;
poly div_poly;
while (resto<>0)
{
dividendo=divisor;
divisor=resto;
div_poly=reduce(dividendo,std(divisor));
resto=div_poly;
}
mcd=divisor;
"Entrega el mcd entre pr y dpr";
mcd;

This is result:

> <"ejemplo_mcd";
out of memory in vector::SetLength()
Singular : signal 6 (v: 3040/2007112211):
Segment fault/Bus error occurred at 83f8da1 because of 0 (r:1274263428)
please inform the authors
trying to restart...
? error occurred in para_singular line 20: `div_poly=reduce(dividendo,std(divisor));`
? last reserved name was `std`
skipping text from `;` error at token `)`
Singular : signal 11 (v: 3040/2007112211):
Segment fault/Bus error occurred at 8361221 because of 0 (r:1274263428)
please inform the authors
trying to restart...

Thank for you help!!

I. Berna


Report this post
Top
 Profile  
Reply with quote  
 Post subject: Re: Help!! gcd procedure
PostPosted: Wed May 19, 2010 4:20 pm 

Joined: Wed Mar 03, 2010 5:08 pm
Posts: 108
Location: Germany, Münster
Dear I. Berna,

as your output shows,
Code:
out of memory in vector::SetLength()
Singular : signal 6 (v: 3040/2007112211):
Segment fault/Bus error occurred at 83f8da1 because of 0 (r:1274263428)
please inform the authors

you are using an old version Singular 3-0-4 dating from 2007.

So upgrade to Singular-3-1-1; it you are working with LINUX go there:
http://www.mathematik.uni-kl.de/ftp/pub ... ular/UNIX/

The crash, which you report here, does not occur in Singular 3-1-1,
so I will not trace the bug.

But with your code there is the problem that it may run into an infinite loop
if one is calling it with coprime multivariate polynomials.

Code:
ring R=(0,s), (a,r,u(1..4)), dp;
poly pr1=r3-2r2+r-1;                 // univariate in r
poly qr1 =diff(pr1,r);

poly pr2=48r3-84r2+42r-(36+s);  // univariate in r, with parameter s
poly qr2 =diff(pr2,r);

poly pr3=-r^3+r^2-s*r+u(1)*s+u(2);   // mulivar.  variables r, u(1),u(2), param. s
poly qr3 =diff(pr3,r);

proc bergcd(poly divisor,poly resto)
{
  poly dividendo, mcd;
  poly pr,qr = divisor,resto;   // store the input

while (resto<>0)
{
dividendo=divisor;
divisor=resto;
resto=reduce(dividendo,std(divisor));
"resto =",resto;
"deg(resto)= ", deg(resto);
"------------------";
division(dividendo,divisor);
read("");
}
mcd=divisor;
"Entrega el mcd entre", pr, "y", qr ," es ", mcd;
return(mcd);
}


Press continously RETURN, and finally press CTRL-C in the third example.

Code:
> bergcd(pr1,qr1);
Entrega el mcd entre r^3-2*r^2+r-1 y 3*r^2-4*r+1  es  207/4

> bergcd(pr2,qr2);
Entrega el mcd entre 48*r^3-84*r^2+42*r+(-s-36) y 144*r^2-168*r+42  es  (324/49*s^2+19800/49*s+302157/49)

>  bergcd(pr3,qr3);
// ..... periodic loop of length 2.
// CTR-C   a  (abort)


For the polynomials pr3,qr3, the rest is again a polynomial in the
variables u(1),u(2). So degree does not drop and the loop does not termiante.

See also
Code:
> ring rxy =0,(x,y),dp;
> poly f = (x+y+1)^2*(x2-y2);

> gcd(f,x+y);
x+y
> bergcd(f,x+y);  // OK
x+y

> gcd(f,x+y+1);
x+y+1
> bergcd(f,x+y+1);   // OK
x+y+1

// But for coprime multivariate, there is a difference.
> gcd(f,x+y+2);
1
> bergcd(f,x+y+2); // does not terminate
........



Summary:

Use Singular's gcd, if you want to compute the gcd of multivariate polynomials,
and instead reduce you may call divsion

---
Christian Gorzel


Report this post
Top
 Profile  
Reply with quote  
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 6 posts ] 

You can post new topics in this forum
You can reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

It is currently Fri May 13, 2022 11:04 am
Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group