LAPACK 3.3.1
Linear Algebra PACKage

dgetrf.f

Go to the documentation of this file.
00001       SUBROUTINE DGETRF( M, N, A, LDA, IPIV, INFO )
00002 *
00003 *  -- LAPACK routine (version 3.2) --
00004 *  -- LAPACK is a software package provided by Univ. of Tennessee,    --
00005 *  -- Univ. of California Berkeley, Univ. of Colorado Denver and NAG Ltd..--
00006 *     November 2006
00007 *
00008 *     .. Scalar Arguments ..
00009       INTEGER            INFO, LDA, M, N
00010 *     ..
00011 *     .. Array Arguments ..
00012       INTEGER            IPIV( * )
00013       DOUBLE PRECISION   A( LDA, * )
00014 *     ..
00015 *
00016 *  Purpose
00017 *  =======
00018 *
00019 *  DGETRF computes an LU factorization of a general M-by-N matrix A
00020 *  using partial pivoting with row interchanges.
00021 *
00022 *  The factorization has the form
00023 *     A = P * L * U
00024 *  where P is a permutation matrix, L is lower triangular with unit
00025 *  diagonal elements (lower trapezoidal if m > n), and U is upper
00026 *  triangular (upper trapezoidal if m < n).
00027 *
00028 *  This is the right-looking Level 3 BLAS version of the algorithm.
00029 *
00030 *  Arguments
00031 *  =========
00032 *
00033 *  M       (input) INTEGER
00034 *          The number of rows of the matrix A.  M >= 0.
00035 *
00036 *  N       (input) INTEGER
00037 *          The number of columns of the matrix A.  N >= 0.
00038 *
00039 *  A       (input/output) DOUBLE PRECISION array, dimension (LDA,N)
00040 *          On entry, the M-by-N matrix to be factored.
00041 *          On exit, the factors L and U from the factorization
00042 *          A = P*L*U; the unit diagonal elements of L are not stored.
00043 *
00044 *  LDA     (input) INTEGER
00045 *          The leading dimension of the array A.  LDA >= max(1,M).
00046 *
00047 *  IPIV    (output) INTEGER array, dimension (min(M,N))
00048 *          The pivot indices; for 1 <= i <= min(M,N), row i of the
00049 *          matrix was interchanged with row IPIV(i).
00050 *
00051 *  INFO    (output) INTEGER
00052 *          = 0:  successful exit
00053 *          < 0:  if INFO = -i, the i-th argument had an illegal value
00054 *          > 0:  if INFO = i, U(i,i) is exactly zero. The factorization
00055 *                has been completed, but the factor U is exactly
00056 *                singular, and division by zero will occur if it is used
00057 *                to solve a system of equations.
00058 *
00059 *  =====================================================================
00060 *
00061 *     .. Parameters ..
00062       DOUBLE PRECISION   ONE
00063       PARAMETER          ( ONE = 1.0D+0 )
00064 *     ..
00065 *     .. Local Scalars ..
00066       INTEGER            I, IINFO, J, JB, NB
00067 *     ..
00068 *     .. External Subroutines ..
00069       EXTERNAL           DGEMM, DGETF2, DLASWP, DTRSM, XERBLA
00070 *     ..
00071 *     .. External Functions ..
00072       INTEGER            ILAENV
00073       EXTERNAL           ILAENV
00074 *     ..
00075 *     .. Intrinsic Functions ..
00076       INTRINSIC          MAX, MIN
00077 *     ..
00078 *     .. Executable Statements ..
00079 *
00080 *     Test the input parameters.
00081 *
00082       INFO = 0
00083       IF( M.LT.0 ) THEN
00084          INFO = -1
00085       ELSE IF( N.LT.0 ) THEN
00086          INFO = -2
00087       ELSE IF( LDA.LT.MAX( 1, M ) ) THEN
00088          INFO = -4
00089       END IF
00090       IF( INFO.NE.0 ) THEN
00091          CALL XERBLA( 'DGETRF', -INFO )
00092          RETURN
00093       END IF
00094 *
00095 *     Quick return if possible
00096 *
00097       IF( M.EQ.0 .OR. N.EQ.0 )
00098      $   RETURN
00099 *
00100 *     Determine the block size for this environment.
00101 *
00102       NB = ILAENV( 1, 'DGETRF', ' ', M, N, -1, -1 )
00103       IF( NB.LE.1 .OR. NB.GE.MIN( M, N ) ) THEN
00104 *
00105 *        Use unblocked code.
00106 *
00107          CALL DGETF2( M, N, A, LDA, IPIV, INFO )
00108       ELSE
00109 *
00110 *        Use blocked code.
00111 *
00112          DO 20 J = 1, MIN( M, N ), NB
00113             JB = MIN( MIN( M, N )-J+1, NB )
00114 *
00115 *           Factor diagonal and subdiagonal blocks and test for exact
00116 *           singularity.
00117 *
00118             CALL DGETF2( M-J+1, JB, A( J, J ), LDA, IPIV( J ), IINFO )
00119 *
00120 *           Adjust INFO and the pivot indices.
00121 *
00122             IF( INFO.EQ.0 .AND. IINFO.GT.0 )
00123      $         INFO = IINFO + J - 1
00124             DO 10 I = J, MIN( M, J+JB-1 )
00125                IPIV( I ) = J - 1 + IPIV( I )
00126    10       CONTINUE
00127 *
00128 *           Apply interchanges to columns 1:J-1.
00129 *
00130             CALL DLASWP( J-1, A, LDA, J, J+JB-1, IPIV, 1 )
00131 *
00132             IF( J+JB.LE.N ) THEN
00133 *
00134 *              Apply interchanges to columns J+JB:N.
00135 *
00136                CALL DLASWP( N-J-JB+1, A( 1, J+JB ), LDA, J, J+JB-1,
00137      $                      IPIV, 1 )
00138 *
00139 *              Compute block row of U.
00140 *
00141                CALL DTRSM( 'Left', 'Lower', 'No transpose', 'Unit', JB,
00142      $                     N-J-JB+1, ONE, A( J, J ), LDA, A( J, J+JB ),
00143      $                     LDA )
00144                IF( J+JB.LE.M ) THEN
00145 *
00146 *                 Update trailing submatrix.
00147 *
00148                   CALL DGEMM( 'No transpose', 'No transpose', M-J-JB+1,
00149      $                        N-J-JB+1, JB, -ONE, A( J+JB, J ), LDA,
00150      $                        A( J, J+JB ), LDA, ONE, A( J+JB, J+JB ),
00151      $                        LDA )
00152                END IF
00153             END IF
00154    20    CONTINUE
00155       END IF
00156       RETURN
00157 *
00158 *     End of DGETRF
00159 *
00160       END
 All Files Functions