概要
本サンプルはFortran言語によりLAPACKルーチンDSPSVXを利用するサンプルプログラムです。
以下の式を解きます。は対称行列です。
解のエラー推定値、行列の条件数の逆数の推定値も合わせて出力されます。
入力データ
(本ルーチンの詳細はDSPSVX のマニュアルページを参照)このデータをダウンロード |
DSPSVX Example Program Data 4 2 :Values of N and NRHS -1.81 2.06 0.63 -1.15 1.15 1.87 4.20 -0.21 3.87 2.07 :End of matrix A 0.96 3.93 6.07 19.25 8.38 9.90 9.50 27.85 :End of matrix B
出力結果
(本ルーチンの詳細はDSPSVX のマニュアルページを参照)この出力例をダウンロード |
DSPSVX Example Program Results Solution(s) 1 2 1 -5.0000 2.0000 2 -2.0000 3.0000 3 1.0000 4.0000 4 4.0000 1.0000 Backward errors (machine-dependent) 1.4E-16 1.0E-16 Estimated forward error bounds (machine-dependent) 2.5E-14 3.2E-14 Estimate of reciprocal condition number 1.3E-02
ソースコード
(本ルーチンの詳細はDSPSVX のマニュアルページを参照)※本サンプルソースコードのご利用手順は「サンプルのコンパイル及び実行方法」をご参照下さい。
このソースコードをダウンロード |
Program dspsvx_example ! DSPSVX Example Program Text ! Copyright 2017, Numerical Algorithms Group Ltd. http://www.nag.com ! .. Use Statements .. Use lapack_example_aux, Only: nagf_file_print_matrix_real_gen Use lapack_interfaces, Only: dspsvx Use lapack_precision, Only: dp ! .. Implicit None Statement .. Implicit None ! .. Parameters .. Integer, Parameter :: nin = 5, nout = 6 Character (1), Parameter :: uplo = 'U' ! .. Local Scalars .. Real (Kind=dp) :: rcond Integer :: i, ifail, info, j, ldb, ldx, n, nrhs ! .. Local Arrays .. Real (Kind=dp), Allocatable :: afp(:), ap(:), b(:, :), berr(:), ferr(:), & work(:), x(:, :) Integer, Allocatable :: ipiv(:), iwork(:) ! .. Executable Statements .. Write (nout, *) 'DSPSVX Example Program Results' Write (nout, *) Flush (nout) ! Skip heading in data file Read (nin, *) Read (nin, *) n, nrhs ldb = n ldx = n Allocate (afp((n*(n+1))/2), ap((n*(n+1))/2), b(ldb,nrhs), berr(nrhs), & ferr(nrhs), work(3*n), x(ldx,nrhs), ipiv(n), iwork(n)) ! Read the upper or lower triangular part of the matrix A from ! data file If (uplo=='U') Then Read (nin, *)((ap(i+(j*(j-1))/2),j=i,n), i=1, n) Else If (uplo=='L') Then Read (nin, *)((ap(i+((2*n-j)*(j-1))/2),j=1,i), i=1, n) End If ! Read B from data file Read (nin, *)(b(i,1:nrhs), i=1, n) ! Solve the equations AX = B for X Call dspsvx('Not factored', uplo, n, nrhs, ap, afp, ipiv, b, ldb, x, & ldx, rcond, ferr, berr, work, iwork, info) If ((info==0) .Or. (info==n+1)) Then ! Print solution, error bounds and condition number ! ifail: behaviour on error exit ! =0 for hard exit, =1 for quiet-soft, =-1 for noisy-soft ifail = 0 Call nagf_file_print_matrix_real_gen('General', ' ', n, nrhs, x, ldx, & 'Solution(s)', ifail) Write (nout, *) Write (nout, *) 'Backward errors (machine-dependent)' Write (nout, 100) berr(1:nrhs) Write (nout, *) Write (nout, *) 'Estimated forward error bounds (machine-dependent)' Write (nout, 100) ferr(1:nrhs) Write (nout, *) Write (nout, *) 'Estimate of reciprocal condition number' Write (nout, 100) rcond Write (nout, *) If (info==n+1) Then Write (nout, *) Write (nout, *) 'The matrix A is singular to working precision' End If Else Write (nout, 110) 'The diagonal block ', info, ' of D is zero' End If 100 Format ((3X,1P,7E11.1)) 110 Format (1X, A, I3, A) End Program