Brian Silverman | 72890c2 | 2015-09-19 14:37:37 -0400 | [diff] [blame^] | 1 | // Small bench routine for Eigen available in Eigen |
| 2 | // (C) Desire NUENTSA WAKAM, INRIA |
| 3 | |
| 4 | #include <iostream> |
| 5 | #include <fstream> |
| 6 | #include <iomanip> |
| 7 | #include <unsupported/Eigen/SparseExtra> |
| 8 | #include <Eigen/SparseLU> |
| 9 | #include <bench/BenchTimer.h> |
| 10 | #ifdef EIGEN_METIS_SUPPORT |
| 11 | #include <Eigen/MetisSupport> |
| 12 | #endif |
| 13 | |
| 14 | using namespace std; |
| 15 | using namespace Eigen; |
| 16 | |
| 17 | int main(int argc, char **args) |
| 18 | { |
| 19 | // typedef complex<double> scalar; |
| 20 | typedef double scalar; |
| 21 | SparseMatrix<scalar, ColMajor> A; |
| 22 | typedef SparseMatrix<scalar, ColMajor>::Index Index; |
| 23 | typedef Matrix<scalar, Dynamic, Dynamic> DenseMatrix; |
| 24 | typedef Matrix<scalar, Dynamic, 1> DenseRhs; |
| 25 | Matrix<scalar, Dynamic, 1> b, x, tmp; |
| 26 | // SparseLU<SparseMatrix<scalar, ColMajor>, AMDOrdering<int> > solver; |
| 27 | // #ifdef EIGEN_METIS_SUPPORT |
| 28 | // SparseLU<SparseMatrix<scalar, ColMajor>, MetisOrdering<int> > solver; |
| 29 | // std::cout<< "ORDERING : METIS\n"; |
| 30 | // #else |
| 31 | SparseLU<SparseMatrix<scalar, ColMajor>, COLAMDOrdering<int> > solver; |
| 32 | std::cout<< "ORDERING : COLAMD\n"; |
| 33 | // #endif |
| 34 | |
| 35 | ifstream matrix_file; |
| 36 | string line; |
| 37 | int n; |
| 38 | BenchTimer timer; |
| 39 | |
| 40 | // Set parameters |
| 41 | /* Fill the matrix with sparse matrix stored in Matrix-Market coordinate column-oriented format */ |
| 42 | if (argc < 2) assert(false && "please, give the matrix market file "); |
| 43 | loadMarket(A, args[1]); |
| 44 | cout << "End charging matrix " << endl; |
| 45 | bool iscomplex=false, isvector=false; |
| 46 | int sym; |
| 47 | getMarketHeader(args[1], sym, iscomplex, isvector); |
| 48 | // if (iscomplex) { cout<< " Not for complex matrices \n"; return -1; } |
| 49 | if (isvector) { cout << "The provided file is not a matrix file\n"; return -1;} |
| 50 | if (sym != 0) { // symmetric matrices, only the lower part is stored |
| 51 | SparseMatrix<scalar, ColMajor> temp; |
| 52 | temp = A; |
| 53 | A = temp.selfadjointView<Lower>(); |
| 54 | } |
| 55 | n = A.cols(); |
| 56 | /* Fill the right hand side */ |
| 57 | |
| 58 | if (argc > 2) |
| 59 | loadMarketVector(b, args[2]); |
| 60 | else |
| 61 | { |
| 62 | b.resize(n); |
| 63 | tmp.resize(n); |
| 64 | // tmp.setRandom(); |
| 65 | for (int i = 0; i < n; i++) tmp(i) = i; |
| 66 | b = A * tmp ; |
| 67 | } |
| 68 | |
| 69 | /* Compute the factorization */ |
| 70 | // solver.isSymmetric(true); |
| 71 | timer.start(); |
| 72 | // solver.compute(A); |
| 73 | solver.analyzePattern(A); |
| 74 | timer.stop(); |
| 75 | cout << "Time to analyze " << timer.value() << std::endl; |
| 76 | timer.reset(); |
| 77 | timer.start(); |
| 78 | solver.factorize(A); |
| 79 | timer.stop(); |
| 80 | cout << "Factorize Time " << timer.value() << std::endl; |
| 81 | timer.reset(); |
| 82 | timer.start(); |
| 83 | x = solver.solve(b); |
| 84 | timer.stop(); |
| 85 | cout << "solve time " << timer.value() << std::endl; |
| 86 | /* Check the accuracy */ |
| 87 | Matrix<scalar, Dynamic, 1> tmp2 = b - A*x; |
| 88 | scalar tempNorm = tmp2.norm()/b.norm(); |
| 89 | cout << "Relative norm of the computed solution : " << tempNorm <<"\n"; |
| 90 | cout << "Number of nonzeros in the factor : " << solver.nnzL() + solver.nnzU() << std::endl; |
| 91 | |
| 92 | return 0; |
| 93 | } |