LAPACK 3.3.0

zlqt02.f

Go to the documentation of this file.
00001       SUBROUTINE ZLQT02( M, N, K, A, AF, Q, L, LDA, TAU, WORK, LWORK,
00002      $                   RWORK, RESULT )
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       INTEGER            K, LDA, LWORK, M, N
00010 *     ..
00011 *     .. Array Arguments ..
00012       DOUBLE PRECISION   RESULT( * ), RWORK( * )
00013       COMPLEX*16         A( LDA, * ), AF( LDA, * ), L( LDA, * ),
00014      $                   Q( LDA, * ), TAU( * ), WORK( LWORK )
00015 *     ..
00016 *
00017 *  Purpose
00018 *  =======
00019 *
00020 *  ZLQT02 tests ZUNGLQ, which generates an m-by-n matrix Q with
00021 *  orthonornmal rows that is defined as the product of k elementary
00022 *  reflectors.
00023 *
00024 *  Given the LQ factorization of an m-by-n matrix A, ZLQT02 generates
00025 *  the orthogonal matrix Q defined by the factorization of the first k
00026 *  rows of A; it compares L(1:k,1:m) with A(1:k,1:n)*Q(1:m,1:n)', and
00027 *  checks that the rows of Q are orthonormal.
00028 *
00029 *  Arguments
00030 *  =========
00031 *
00032 *  M       (input) INTEGER
00033 *          The number of rows of the matrix Q to be generated.  M >= 0.
00034 *
00035 *  N       (input) INTEGER
00036 *          The number of columns of the matrix Q to be generated.
00037 *          N >= M >= 0.
00038 *
00039 *  K       (input) INTEGER
00040 *          The number of elementary reflectors whose product defines the
00041 *          matrix Q. M >= K >= 0.
00042 *
00043 *  A       (input) COMPLEX*16 array, dimension (LDA,N)
00044 *          The m-by-n matrix A which was factorized by ZLQT01.
00045 *
00046 *  AF      (input) COMPLEX*16 array, dimension (LDA,N)
00047 *          Details of the LQ factorization of A, as returned by ZGELQF.
00048 *          See ZGELQF for further details.
00049 *
00050 *  Q       (workspace) COMPLEX*16 array, dimension (LDA,N)
00051 *
00052 *  L       (workspace) COMPLEX*16 array, dimension (LDA,M)
00053 *
00054 *  LDA     (input) INTEGER
00055 *          The leading dimension of the arrays A, AF, Q and L. LDA >= N.
00056 *
00057 *  TAU     (input) COMPLEX*16 array, dimension (M)
00058 *          The scalar factors of the elementary reflectors corresponding
00059 *          to the LQ factorization in AF.
00060 *
00061 *  WORK    (workspace) COMPLEX*16 array, dimension (LWORK)
00062 *
00063 *  LWORK   (input) INTEGER
00064 *          The dimension of the array WORK.
00065 *
00066 *  RWORK   (workspace) DOUBLE PRECISION array, dimension (M)
00067 *
00068 *  RESULT  (output) DOUBLE PRECISION array, dimension (2)
00069 *          The test ratios:
00070 *          RESULT(1) = norm( L - A*Q' ) / ( N * norm(A) * EPS )
00071 *          RESULT(2) = norm( I - Q*Q' ) / ( N * EPS )
00072 *
00073 *  =====================================================================
00074 *
00075 *     .. Parameters ..
00076       DOUBLE PRECISION   ZERO, ONE
00077       PARAMETER          ( ZERO = 0.0D+0, ONE = 1.0D+0 )
00078       COMPLEX*16         ROGUE
00079       PARAMETER          ( ROGUE = ( -1.0D+10, -1.0D+10 ) )
00080 *     ..
00081 *     .. Local Scalars ..
00082       INTEGER            INFO
00083       DOUBLE PRECISION   ANORM, EPS, RESID
00084 *     ..
00085 *     .. External Functions ..
00086       DOUBLE PRECISION   DLAMCH, ZLANGE, ZLANSY
00087       EXTERNAL           DLAMCH, ZLANGE, ZLANSY
00088 *     ..
00089 *     .. External Subroutines ..
00090       EXTERNAL           ZGEMM, ZHERK, ZLACPY, ZLASET, ZUNGLQ
00091 *     ..
00092 *     .. Intrinsic Functions ..
00093       INTRINSIC          DBLE, DCMPLX, MAX
00094 *     ..
00095 *     .. Scalars in Common ..
00096       CHARACTER*32       SRNAMT
00097 *     ..
00098 *     .. Common blocks ..
00099       COMMON             / SRNAMC / SRNAMT
00100 *     ..
00101 *     .. Executable Statements ..
00102 *
00103       EPS = DLAMCH( 'Epsilon' )
00104 *
00105 *     Copy the first k rows of the factorization to the array Q
00106 *
00107       CALL ZLASET( 'Full', M, N, ROGUE, ROGUE, Q, LDA )
00108       CALL ZLACPY( 'Upper', K, N-1, AF( 1, 2 ), LDA, Q( 1, 2 ), LDA )
00109 *
00110 *     Generate the first n columns of the matrix Q
00111 *
00112       SRNAMT = 'ZUNGLQ'
00113       CALL ZUNGLQ( M, N, K, Q, LDA, TAU, WORK, LWORK, INFO )
00114 *
00115 *     Copy L(1:k,1:m)
00116 *
00117       CALL ZLASET( 'Full', K, M, DCMPLX( ZERO ), DCMPLX( ZERO ), L,
00118      $             LDA )
00119       CALL ZLACPY( 'Lower', K, M, AF, LDA, L, LDA )
00120 *
00121 *     Compute L(1:k,1:m) - A(1:k,1:n) * Q(1:m,1:n)'
00122 *
00123       CALL ZGEMM( 'No transpose', 'Conjugate transpose', K, M, N,
00124      $            DCMPLX( -ONE ), A, LDA, Q, LDA, DCMPLX( ONE ), L,
00125      $            LDA )
00126 *
00127 *     Compute norm( L - A*Q' ) / ( N * norm(A) * EPS ) .
00128 *
00129       ANORM = ZLANGE( '1', K, N, A, LDA, RWORK )
00130       RESID = ZLANGE( '1', K, M, L, LDA, RWORK )
00131       IF( ANORM.GT.ZERO ) THEN
00132          RESULT( 1 ) = ( ( RESID / DBLE( MAX( 1, N ) ) ) / ANORM ) / EPS
00133       ELSE
00134          RESULT( 1 ) = ZERO
00135       END IF
00136 *
00137 *     Compute I - Q*Q'
00138 *
00139       CALL ZLASET( 'Full', M, M, DCMPLX( ZERO ), DCMPLX( ONE ), L, LDA )
00140       CALL ZHERK( 'Upper', 'No transpose', M, N, -ONE, Q, LDA, ONE, L,
00141      $            LDA )
00142 *
00143 *     Compute norm( I - Q*Q' ) / ( N * EPS ) .
00144 *
00145       RESID = ZLANSY( '1', 'Upper', M, L, LDA, RWORK )
00146 *
00147       RESULT( 2 ) = ( RESID / DBLE( MAX( 1, N ) ) ) / EPS
00148 *
00149       RETURN
00150 *
00151 *     End of ZLQT02
00152 *
00153       END
 All Files Functions