LAPACK 3.3.1
Linear Algebra PACKage

ztpt02.f

Go to the documentation of this file.
00001       SUBROUTINE ZTPT02( UPLO, TRANS, DIAG, N, NRHS, AP, X, LDX, B, LDB,
00002      $                   WORK, RWORK, RESID )
00003 *
00004 *  -- LAPACK test routine (version 3.1) --
00005 *     Univ. of Tennessee, Univ. of California Berkeley and NAG Ltd..
00006 *     November 2006
00007 *
00008 *     .. Scalar Arguments ..
00009       CHARACTER          DIAG, TRANS, UPLO
00010       INTEGER            LDB, LDX, N, NRHS
00011       DOUBLE PRECISION   RESID
00012 *     ..
00013 *     .. Array Arguments ..
00014       DOUBLE PRECISION   RWORK( * )
00015       COMPLEX*16         AP( * ), B( LDB, * ), WORK( * ), X( LDX, * )
00016 *     ..
00017 *
00018 *  Purpose
00019 *  =======
00020 *
00021 *  ZTPT02 computes the residual for the computed solution to a
00022 *  triangular system of linear equations  A*x = b,  A**T *x = b,  or
00023 *  A**H *x = b, when the triangular matrix A is stored in packed format.
00024 *  Here A**T denotes the transpose of A, A**H denotes the conjugate
00025 *  transpose of A, and x and b are N by NRHS matrices.  The test ratio
00026 *  is the maximum over the number of right hand sides of
00027 *  the maximum over the number of right hand sides of
00028 *     norm(b - op(A)*x) / ( norm(op(A)) * norm(x) * EPS ),
00029 *  where op(A) denotes A, A**T, or A**H, and EPS is the machine epsilon.
00030 *
00031 *  Arguments
00032 *  =========
00033 *
00034 *  UPLO    (input) CHARACTER*1
00035 *          Specifies whether the matrix A is upper or lower triangular.
00036 *          = 'U':  Upper triangular
00037 *          = 'L':  Lower triangular
00038 *
00039 *  TRANS   (input) CHARACTER*1
00040 *          Specifies the operation applied to A.
00041 *          = 'N':  A *x = b     (No transpose)
00042 *          = 'T':  A**T *x = b  (Transpose)
00043 *          = 'C':  A**H *x = b  (Conjugate transpose)
00044 *
00045 *  DIAG    (input) CHARACTER*1
00046 *          Specifies whether or not the matrix A is unit triangular.
00047 *          = 'N':  Non-unit triangular
00048 *          = 'U':  Unit triangular
00049 *
00050 *  N       (input) INTEGER
00051 *          The order of the matrix A.  N >= 0.
00052 *
00053 *  NRHS    (input) INTEGER
00054 *          The number of right hand sides, i.e., the number of columns
00055 *          of the matrices X and B.  NRHS >= 0.
00056 *
00057 *  AP      (input) COMPLEX*16 array, dimension (N*(N+1)/2)
00058 *          The upper or lower triangular matrix A, packed columnwise in
00059 *          a linear array.  The j-th column of A is stored in the array
00060 *          AP as follows:
00061 *          if UPLO = 'U', AP((j-1)*j/2 + i) = A(i,j) for 1<=i<=j;
00062 *          if UPLO = 'L',
00063 *             AP((j-1)*(n-j) + j*(j+1)/2 + i-j) = A(i,j) for j<=i<=n.
00064 *
00065 *  X       (input) COMPLEX*16 array, dimension (LDX,NRHS)
00066 *          The computed solution vectors for the system of linear
00067 *          equations.
00068 *
00069 *  LDX     (input) INTEGER
00070 *          The leading dimension of the array X.  LDX >= max(1,N).
00071 *
00072 *  B       (input) COMPLEX*16 array, dimension (LDB,NRHS)
00073 *          The right hand side vectors for the system of linear
00074 *          equations.
00075 *
00076 *  LDB     (input) INTEGER
00077 *          The leading dimension of the array B.  LDB >= max(1,N).
00078 *
00079 *  WORK    (workspace) COMPLEX*16 array, dimension (N)
00080 *
00081 *  RWORK   (workspace) DOUBLE PRECISION array, dimension (N)
00082 *
00083 *  RESID   (output) DOUBLE PRECISION
00084 *          The maximum over the number of right hand sides of
00085 *          norm(op(A)*x - b) / ( norm(op(A)) * norm(x) * EPS ).
00086 *
00087 *  =====================================================================
00088 *
00089 *     .. Parameters ..
00090       DOUBLE PRECISION   ZERO, ONE
00091       PARAMETER          ( ZERO = 0.0D+0, ONE = 1.0D+0 )
00092 *     ..
00093 *     .. Local Scalars ..
00094       INTEGER            J
00095       DOUBLE PRECISION   ANORM, BNORM, EPS, XNORM
00096 *     ..
00097 *     .. External Functions ..
00098       LOGICAL            LSAME
00099       DOUBLE PRECISION   DLAMCH, DZASUM, ZLANTP
00100       EXTERNAL           LSAME, DLAMCH, DZASUM, ZLANTP
00101 *     ..
00102 *     .. External Subroutines ..
00103       EXTERNAL           ZAXPY, ZCOPY, ZTPMV
00104 *     ..
00105 *     .. Intrinsic Functions ..
00106       INTRINSIC          DCMPLX, MAX
00107 *     ..
00108 *     .. Executable Statements ..
00109 *
00110 *     Quick exit if N = 0 or NRHS = 0
00111 *
00112       IF( N.LE.0 .OR. NRHS.LE.0 ) THEN
00113          RESID = ZERO
00114          RETURN
00115       END IF
00116 *
00117 *     Compute the 1-norm of A or A**H.
00118 *
00119       IF( LSAME( TRANS, 'N' ) ) THEN
00120          ANORM = ZLANTP( '1', UPLO, DIAG, N, AP, RWORK )
00121       ELSE
00122          ANORM = ZLANTP( 'I', UPLO, DIAG, N, AP, RWORK )
00123       END IF
00124 *
00125 *     Exit with RESID = 1/EPS if ANORM = 0.
00126 *
00127       EPS = DLAMCH( 'Epsilon' )
00128       IF( ANORM.LE.ZERO ) THEN
00129          RESID = ONE / EPS
00130          RETURN
00131       END IF
00132 *
00133 *     Compute the maximum over the number of right hand sides of
00134 *        norm(op(A)*x - b) / ( norm(op(A)) * norm(x) * EPS ).
00135 *
00136       RESID = ZERO
00137       DO 10 J = 1, NRHS
00138          CALL ZCOPY( N, X( 1, J ), 1, WORK, 1 )
00139          CALL ZTPMV( UPLO, TRANS, DIAG, N, AP, WORK, 1 )
00140          CALL ZAXPY( N, DCMPLX( -ONE ), B( 1, J ), 1, WORK, 1 )
00141          BNORM = DZASUM( N, WORK, 1 )
00142          XNORM = DZASUM( N, X( 1, J ), 1 )
00143          IF( XNORM.LE.ZERO ) THEN
00144             RESID = ONE / EPS
00145          ELSE
00146             RESID = MAX( RESID, ( ( BNORM / ANORM ) / XNORM ) / EPS )
00147          END IF
00148    10 CONTINUE
00149 *
00150       RETURN
00151 *
00152 *     End of ZTPT02
00153 *
00154       END
 All Files Functions