May 2005 - GiNaC-devel

Patch for products of dummy summations
by Vladimir Kisil 12 Nov '05

12 Nov '05

Here is the patch which was described in my previous message (Fri May 13 17:27:25 CEST 2005). -- Vladimir V. Kisil email: kisilv(a)maths.leeds.ac.uk -- www: http://maths.leeds.ac.uk/~kisilv/

2 1

Problems with dummy summation
by Vladimir Kisil 13 May '05

13 May '05

Dear All, I recently noticed that GiNaC incorrectly expands products with dummy index summations, which is illustrated by the following program: #include <iostream> #include <ginac/ginac.h> using namespace std; using namespace GiNaC; #include <stdexcept> int main(){ realsymbol a("a"), b("b"), c("c"), d("d"); varidx mu(symbol("mu", "\\mu"), 2); try { ex e; e = (indexed(a, mu)*indexed(b, mu.toggle_variance())+1)*indexed(c, mu)*indexed(d, mu.toggle_variance()); cout << e.expand() << endl; // -> d.mu*c~mu*b.mu*a~mu+d.mu*c~mu e = pow(indexed(a, mu)*indexed(b, mu.toggle_variance())+1, 2); cout << e.expand() << endl; // -> 1+2*b.mu*a~mu+(b.mu)^2*(a~mu)^2 } catch (exception &p) { cerr << "Got problem: " << p.what() << endl; } } I made a patch to mul.cpp, ncmul.cpp, power.cpp which _partially_ fix this problem. The bigger test looks now like that: #include <iostream> #include <ginac/ginac.h> using namespace std; using namespace GiNaC; #include <stdexcept> int main(){ realsymbol a("a"), b("b"), c("c"), d("d"); varidx mu(symbol("mu", "\\mu"), 2), nu(symbol("nu", "\\nu"), 2); try { ex e; e = (indexed(a, mu)*indexed(b, mu.toggle_variance())+1)*indexed(c, mu)*indexed(d, mu.toggle_variance()); cout << e.expand() << endl << endl; // -> d.mu*c~mu*b.symbol6*a~symbol6+d.mu*c~mu e = indexed(c, mu)*indexed(d, mu.toggle_variance())*(indexed(a, mu)*indexed(b, mu.toggle_variance())+1); cout << e.expand() << endl << endl; // -> d.mu*c~mu+d.mu*a~symbol7*b.symbol7*c~mu e = (1+indexed(c, mu)*indexed(d, mu.toggle_variance()))*(indexed(a, mu)*indexed(b, mu.toggle_variance())+1); cout << e.expand() << endl << endl; // -> 1+b.mu*a~mu+b.mu*c~symbol8*a~mu*d.symbol8+d.mu*c~mu e = pow(2*indexed(a, mu)*indexed(b, mu.toggle_variance())+1, 2); cout << e.expand() << endl << endl; // -> 1+4*b.mu*a~mu*b.symbol9*a~symbol9+4*b.mu*a~mu e = pow(2*indexed(a, mu)*indexed(b, mu.toggle_variance())+1, 3); cout << e.expand() << endl << endl; // -> 1+6*b.mu*a~mu+12*b.mu*a~mu*a~symbol10*b.symbol10+8*b.symbol12*a~symbol12*b.mu*a~symbol11*b.symbol11*a~mu e = (indexed(a, mu, nu)*dirac_gamma(mu.toggle_variance()) * dirac_gamma(nu.toggle_variance()) + dirac_ONE()) * indexed(b, mu, nu)*dirac_gamma(mu.toggle_variance()) * dirac_gamma(nu.toggle_variance()); cout << e.expand() << endl << endl; // -> (gamma.mu*gamma.nu*gamma.mu*gamma.nu)*a~mu~nu*b~mu~nu+(gamma.mu*gamma.nu)*b~mu~nu e = (indexed(a, mu, nu)*dirac_gamma(mu.toggle_variance()) * dirac_gamma(nu.toggle_variance()) + dirac_ONE()) *( indexed(b, mu, nu)*dirac_gamma(mu.toggle_variance()) * dirac_gamma(nu.toggle_variance()) + dirac_ONE()); cout << e.expand() << endl << endl; // -> b~symbol16~symbol17*a~mu~nu*(gamma.mu*gamma.nu*gamma.symbol16*gamma.symbol17)+ONE+(gamma.mu*gamma.nu)*a~mu~nu+(gamma.mu*gamma.nu)*b~mu~nu e = (indexed(a, mu, nu)*dirac_gamma(mu.toggle_variance()) * dirac_gamma(nu.toggle_variance()) + dirac_ONE()) * indexed(b, mu.toggle_variance(), nu.toggle_variance())*dirac_gamma(mu) * dirac_gamma(nu); cout << e.expand() << endl << endl; // -> (gamma.mu*gamma.nu*gamma~mu*gamma~nu)*a~mu~nu*b.mu.nu+(gamma~mu*gamma~nu)*b.mu.nu e = pow(indexed(a, mu, nu)*dirac_gamma(mu.toggle_variance()) * dirac_gamma(nu.toggle_variance()), 2); cout << e.expand() << endl << endl; // -> (gamma.mu*gamma.nu*gamma.symbol18*gamma.symbol19)*a~mu~nu*a~symbol18~symbol19 e = pow(2*indexed(a, mu, nu)*dirac_gamma(mu.toggle_variance()) * dirac_gamma(nu.toggle_variance())+dirac_ONE(), 2); cout << e.expand() << endl << endl; // -> 4*(gamma.mu*gamma.nu*gamma.symbol18*gamma.symbol19)*a~mu~nu*a~symbol18~symbol19+ONE+4*(gamma.mu*gamma.nu)*a~mu~nu e = pow(2*indexed(a, mu, nu)*dirac_gamma(mu.toggle_variance()) * dirac_gamma(nu.toggle_variance())+dirac_ONE(), 3); cout << e.expand() << endl << endl; // -> ONE+8*a~symbol26~symbol27*a~mu~nu*a~symbol28~symbol29*(gamma.mu*gamma.nu*gamma.symbol26*gamma.symbol27*gamma.symbol28*gamma.symbol29)+6*(gamma.mu*gamma.nu)*a~mu~nu+4*(gamma.mu*gamma.nu*gamma.symbol24*gamma.symbol25)*a~mu~nu*a~symbol24~symbol25+4*a~symbol22~symbol23*a~mu~nu*(gamma.mu*gamma.nu*gamma.symbol22*gamma.symbol23)+4*a~symbol20~symbol21*a~mu~nu*(gamma.mu*gamma.nu*gamma.symbol20*gamma.symbol21) } catch (exception &p) { cerr << "Got problem: " << p.what() << endl; } } So the problem remains in the cases when one dummy index belongs to non-commutative part of product and other---to a commutative part. I do not know how to fix this situation in a simple way. I sent my patch as an attachment to a separate letter to this list (which will be likely hold till moderator's approval). It includes as well a slightly redesigned version of my previous patch for indexed.cpp and clifford.cpp (from the Thu Apr 28 18:17:55 CEST 2005). Best regards, Vladimir -- Vladimir V. Kisil email: kisilv(a)maths.leeds.ac.uk -- www: http://maths.leeds.ac.uk/~kisilv/

1 0

GiNaC 1.3.1
by Jens Vollinga 04 May '05

04 May '05

Hi, GiNaC 1.3.1 is out and available. The changes are: - integral() and eval_integ() can be used from ginsh. - Integrals can be series-expanded. - Fixed a library initialization problem. - GiNaC compiles from tarball even if lex/flex is missing. - Fixed bugs in canonicalize_clifford(), clifford_prime() and clifford_to_lst(). - clifford_moebius_map(), remove_dirac_ONE() and LaTeX output of Clifford objects now care about representation labels. - Fixed bug in gcd. - Better output for slashed expressions. As always, this release can be downloaded from ftp://ftpthep.physik.uni-mainz.de/pub/GiNaC/ Enjoy, Jens

1 0