📄 ssor.cc
字号:
#include "mtl/matrix.h"#include "mtl/mtl.h"#include "mtl/utils.h"#include "itl/itl.h"#include "itl/ssor.h"#include "itl/qmr.h"/* In thsi example, we show how to use SSOR, the output should be:iteration 0: resid 1iteration 1: resid 2.52919iteration 2: resid 0.468754iteration 3: resid 0.0612519finished! error code = 04 iterations6.05118e-14 final residual 1 2 0 0 3 x 0.205847 = 1 4 5 0 6 0 x 0.0419363 = 1 0 7 8 0 9 x -0.178049 = 1 0 0 10 11 12 x -0.00551162 = 1 0 0 13 0 14 x 0.23676 = 1*/using namespace mtl;using namespace itl;typedef double Type;//begintypedef matrix< Type, rectangle<>, array< compressed<> >, row_major >::type Matrix;//endint main (int , char* []) { int max_iter = 50; //begin Matrix A(5, 5); //end A(0, 0) = 1.0; A(0, 1) = 2.0; A(0, 4) = 3.0; A(1, 0) = 4.0; A(1, 1) = 5.0; A(1, 3) = 6.0; A(2, 1) = 7.0; A(2, 2) = 8.0; A(2, 4) = 9.0; A(3, 2) = 10.; A(3, 3) = 11.; A(3, 4) = 12.; A(4, 2) = 13.; A(4, 4) = 14.; //begin dense1D<Type> x(A.nrows(), Type(0)); dense1D<Type> b(A.ncols()); for (dense1D<Type>::iterator i=b.begin(); i!=b.end(); i++) *i = 1.; //SSOR preconditioner SSOR<Matrix> precond(A); noisy_iteration<double> iter(b, max_iter, 1e-6); qmr(A, x, b, precond.left(), precond.right(), iter); //end //verify the result dense1D<Type> b1(A.ncols()); mult(A, x, b1); add(b1, scaled(b, -1.), b1); if ( two_norm(b1) < 1.e-6) { //output for (int i=0; i<5; i++) { for (int j=0; j<5; j++) { std::cout.width(6); std::cout << A(i, j) << " "; } std::cout << " x "; std::cout.width(16); std::cout << x[i] << " = "; std::cout.width(6); std::cout << b[i] << std::endl; } } else { std::cout << "Residual " << iter.resid() << std::endl; } return 0; }
⌨️ 快捷键说明
复制代码
Ctrl + C
搜索代码
Ctrl + F
全屏模式
F11
切换主题
Ctrl + Shift + D
显示快捷键
?
增大字号
Ctrl + =
减小字号
Ctrl + -