TMatrixTSymCramerInv Encapsulate templates of Cramer Inversion routines. The 4x4, 5x5 and 6x6 are adapted from routines written by Mark Fischler and Steven Haywood as part of the CLHEP package Although for sizes <= 6x6 the Cramer Inversion has a gain in speed compared to factorization schemes (like LU) , one pays a price in accuracy . For Example: H * H^-1 = U, where H is a 5x5 Hilbert matrix U is a 5x5 Unity matrix LU : |U_jk| < 10e-13 for j!=k Cramer: |U_jk| < 10e-7 for j!=k however Cramer algorithm is about 10 (!) times faster
Bool_t | Inv2x2(TMatrixTSym<float>& m, Double_t* determ) |
Bool_t | Inv2x2(TMatrixTSym<double>& m, Double_t* determ) |
Bool_t | Inv3x3(TMatrixTSym<float>& m, Double_t* determ) |
Bool_t | Inv3x3(TMatrixTSym<double>& m, Double_t* determ) |
Bool_t | Inv4x4(TMatrixTSym<float>& m, Double_t* determ) |
Bool_t | Inv4x4(TMatrixTSym<double>& m, Double_t* determ) |
Bool_t | Inv5x5(TMatrixTSym<float>& m, Double_t* determ) |
Bool_t | Inv5x5(TMatrixTSym<double>& m, Double_t* determ) |
Bool_t | Inv6x6(TMatrixTSym<float>& m, Double_t* determ) |
Bool_t | Inv6x6(TMatrixTSym<double>& m, Double_t* determ) |