sgelsd.c

提供矩阵类的函数库

#include "blaswrap.h"
/*  -- translated by f2c (version 19990503).
   You must link the resulting object file with the libraries:
	-lf2c -lm   (in that order)
*/

#include "f2c.h"

/* Table of constant values */

static integer c__6 = 6;
static integer c_n1 = -1;
static integer c__9 = 9;
static integer c__0 = 0;
static integer c__1 = 1;
static real c_b82 = 0.f;

/* Subroutine */ int sgelsd_(integer *m, integer *n, integer *nrhs, real *a, 
	integer *lda, real *b, integer *ldb, real *s, real *rcond, integer *
	rank, real *work, integer *lwork, integer *iwork, integer *info)
{
    /* System generated locals */
    integer a_dim1, a_offset, b_dim1, b_offset, i__1, i__2, i__3, i__4;

    /* Builtin functions */
    double log(doublereal);

    /* Local variables */
    static real anrm, bnrm;
    static integer itau, nlvl, iascl, ibscl;
    static real sfmin;
    static integer minmn, maxmn, itaup, itauq, mnthr, nwork, ie, il;
    extern /* Subroutine */ int slabad_(real *, real *);
    static integer mm;
    extern /* Subroutine */ int sgebrd_(integer *, integer *, real *, integer 
	    *, real *, real *, real *, real *, real *, integer *, integer *);
    extern doublereal slamch_(char *), slange_(char *, integer *, 
	    integer *, real *, integer *, real *);
    extern /* Subroutine */ int xerbla_(char *, integer *);
    extern integer ilaenv_(integer *, char *, char *, integer *, integer *, 
	    integer *, integer *, ftnlen, ftnlen);
    static real bignum;
    extern /* Subroutine */ int sgelqf_(integer *, integer *, real *, integer 
	    *, real *, real *, integer *, integer *), slalsd_(char *, integer 
	    *, integer *, integer *, real *, real *, real *, integer *, real *
	    , integer *, real *, integer *, integer *), slascl_(char *
	    , integer *, integer *, real *, real *, integer *, integer *, 
	    real *, integer *, integer *);
    static integer wlalsd;
    extern /* Subroutine */ int sgeqrf_(integer *, integer *, real *, integer 
	    *, real *, real *, integer *, integer *), slacpy_(char *, integer 
	    *, integer *, real *, integer *, real *, integer *), 
	    slaset_(char *, integer *, integer *, real *, real *, real *, 
	    integer *);
    static integer ldwork;
    extern /* Subroutine */ int sormbr_(char *, char *, char *, integer *, 
	    integer *, integer *, real *, integer *, real *, real *, integer *
	    , real *, integer *, integer *);
    static integer minwrk, maxwrk;
    static real smlnum;
    extern /* Subroutine */ int sormlq_(char *, char *, integer *, integer *, 
	    integer *, real *, integer *, real *, real *, integer *, real *, 
	    integer *, integer *);
    static logical lquery;
    static integer smlsiz;
    extern /* Subroutine */ int sormqr_(char *, char *, integer *, integer *, 
	    integer *, real *, integer *, real *, real *, integer *, real *, 
	    integer *, integer *);
    static real eps;


#define a_ref(a_1,a_2) a[(a_2)*a_dim1 + a_1]
#define b_ref(a_1,a_2) b[(a_2)*b_dim1 + a_1]


/*  -- LAPACK driver routine (version 3.0) --   
       Univ. of Tennessee, Univ. of California Berkeley, NAG Ltd.,   
       Courant Institute, Argonne National Lab, and Rice University   
       October 31, 1999   


    Purpose   
    =======   

    SGELSD computes the minimum-norm solution to a real linear least   
    squares problem:   
        minimize 2-norm(| b - A*x |)   
    using the singular value decomposition (SVD) of A. A is an M-by-N   
    matrix which may be rank-deficient.   

    Several right hand side vectors b and solution vectors x can be   
    handled in a single call; they are stored as the columns of the   
    M-by-NRHS right hand side matrix B and the N-by-NRHS solution   
    matrix X.   

    The problem is solved in three steps:   
    (1) Reduce the coefficient matrix A to bidiagonal form with   
        Householder transformations, reducing the original problem   
        into a "bidiagonal least squares problem" (BLS)   
    (2) Solve the BLS using a divide and conquer approach.   
    (3) Apply back all the Householder tranformations to solve   
        the original least squares problem.   

    The effective rank of A is determined by treating as zero those   
    singular values which are less than RCOND times the largest singular   
    value.   

    The divide and conquer algorithm makes very mild assumptions about   
    floating point arithmetic. It will work on machines with a guard   
    digit in add/subtract, or on those binary machines without guard   
    digits which subtract like the Cray X-MP, Cray Y-MP, Cray C-90, or   
    Cray-2. It could conceivably fail on hexadecimal or decimal machines   
    without guard digits, but we know of none.   

    Arguments   
    =========   

    M       (input) INTEGER   
            The number of rows of A. M >= 0.   

    N       (input) INTEGER   
            The number of columns of A. N >= 0.   

    NRHS    (input) INTEGER   
            The number of right hand sides, i.e., the number of columns   
            of the matrices B and X. NRHS >= 0.   

    A       (input) REAL array, dimension (LDA,N)   
            On entry, the M-by-N matrix A.   
            On exit, A has been destroyed.   

    LDA     (input) INTEGER   
            The leading dimension of the array A.  LDA >= max(1,M).   

    B       (input/output) REAL array, dimension (LDB,NRHS)   
            On entry, the M-by-NRHS right hand side matrix B.   
            On exit, B is overwritten by the N-by-NRHS solution   
            matrix X.  If m >= n and RANK = n, the residual   
            sum-of-squares for the solution in the i-th column is given   
            by the sum of squares of elements n+1:m in that column.   

    LDB     (input) INTEGER   
            The leading dimension of the array B. LDB >= max(1,max(M,N)).   

    S       (output) REAL array, dimension (min(M,N))   
            The singular values of A in decreasing order.   
            The condition number of A in the 2-norm = S(1)/S(min(m,n)).   

    RCOND   (input) REAL   
            RCOND is used to determine the effective rank of A.   
            Singular values S(i) <= RCOND*S(1) are treated as zero.   
            If RCOND < 0, machine precision is used instead.   

    RANK    (output) INTEGER   
            The effective rank of A, i.e., the number of singular values   
            which are greater than RCOND*S(1).   

    WORK    (workspace/output) REAL array, dimension (LWORK)   
            On exit, if INFO = 0, WORK(1) returns the optimal LWORK.   

    LWORK   (input) INTEGER   
            The dimension of the array WORK. LWORK must be at least 1.   
            The exact minimum amount of workspace needed depends on M,   
            N and NRHS. As long as LWORK is at least   
                12*N + 2*N*SMLSIZ + 8*N*NLVL + N*NRHS + (SMLSIZ+1)**2,   
            if M is greater than or equal to N or   
                12*M + 2*M*SMLSIZ + 8*M*NLVL + M*NRHS + (SMLSIZ+1)**2,   
            if M is less than N, the code will execute correctly.   
            SMLSIZ is returned by ILAENV and is equal to the maximum   
            size of the subproblems at the bottom of the computation   
            tree (usually about 25), and   
               NLVL = MAX( 0, INT( LOG_2( MIN( M,N )/(SMLSIZ+1) ) ) + 1 )   
            For good performance, LWORK should generally be larger.   

            If LWORK = -1, then a workspace query is assumed; the routine   
            only calculates the optimal size of the WORK array, returns   
            this value as the first entry of the WORK array, and no error   
            message related to LWORK is issued by XERBLA.   

    IWORK   (workspace) INTEGER array, dimension (LIWORK)   
            LIWORK >= 3 * MINMN * NLVL + 11 * MINMN,   
            where MINMN = MIN( M,N ).   

    INFO    (output) INTEGER   
            = 0:  successful exit   
            < 0:  if INFO = -i, the i-th argument had an illegal value.   
            > 0:  the algorithm for computing the SVD failed to converge;   
                  if INFO = i, i off-diagonal elements of an intermediate   
                  bidiagonal form did not converge to zero.   

    Further Details   
    ===============   

    Based on contributions by   
       Ming Gu and Ren-Cang Li, Computer Science Division, University of   
         California at Berkeley, USA   
       Osni Marques, LBNL/NERSC, USA   

    =====================================================================   


       Test the input arguments.   

       Parameter adjustments */
    a_dim1 = *lda;
    a_offset = 1 + a_dim1 * 1;
    a -= a_offset;
    b_dim1 = *ldb;
    b_offset = 1 + b_dim1 * 1;
    b -= b_offset;
    --s;
    --work;
    --iwork;

    /* Function Body */
    *info = 0;
    minmn = min(*m,*n);
    maxmn = max(*m,*n);
    mnthr = ilaenv_(&c__6, "SGELSD", " ", m, n, nrhs, &c_n1, (ftnlen)6, (
	    ftnlen)1);
    lquery = *lwork == -1;
    if (*m < 0) {
	*info = -1;
    } else if (*n < 0) {
	*info = -2;
    } else if (*nrhs < 0) {
	*info = -3;
    } else if (*lda < max(1,*m)) {
	*info = -5;
    } else if (*ldb < max(1,maxmn)) {
	*info = -7;
    }

    smlsiz = ilaenv_(&c__9, "SGELSD", " ", &c__0, &c__0, &c__0, &c__0, (
	    ftnlen)6, (ftnlen)1);

/*     Compute workspace.   
       (Note: Comments in the code beginning "Workspace:" describe the   
       minimal amount of workspace needed at that point in the code,   
       as well as the preferred amount for good performance.   
       NB refers to the optimal block size for the immediately   
       following subroutine, as returned by ILAENV.) */

    minwrk = 1;
    minmn = max(1,minmn);
/* Computing MAX */
    i__1 = (integer) (log((real) minmn / (real) (smlsiz + 1)) / log(2.f)) + 1;
    nlvl = max(i__1,0);

    if (*info == 0) {
	maxwrk = 0;
	mm = *m;
	if (*m >= *n && *m >= mnthr) {

/*           Path 1a - overdetermined, with many more rows than columns. */

	    mm = *n;
/* Computing MAX */
	    i__1 = maxwrk, i__2 = *n + *n * ilaenv_(&c__1, "SGEQRF", " ", m, 
		    n, &c_n1, &c_n1, (ftnlen)6, (ftnlen)1);
	    maxwrk = max(i__1,i__2);
/* Computing MAX */
	    i__1 = maxwrk, i__2 = *n + *nrhs * ilaenv_(&c__1, "SORMQR", "LT", 
		    m, nrhs, n, &c_n1, (ftnlen)6, (ftnlen)2);
	    maxwrk = max(i__1,i__2);
	}
	if (*m >= *n) {

/*           Path 1 - overdetermined or exactly determined.   

   Computing MAX */
	    i__1 = maxwrk, i__2 = *n * 3 + (mm + *n) * ilaenv_(&c__1, "SGEBRD"
		    , " ", &mm, n, &c_n1, &c_n1, (ftnlen)6, (ftnlen)1);
	    maxwrk = max(i__1,i__2);
/* Computing MAX */
	    i__1 = maxwrk, i__2 = *n * 3 + *nrhs * ilaenv_(&c__1, "SORMBR", 
		    "QLT", &mm, nrhs, n, &c_n1, (ftnlen)6, (ftnlen)3);
	    maxwrk = max(i__1,i__2);
/* Computing MAX */
	    i__1 = maxwrk, i__2 = *n * 3 + (*n - 1) * ilaenv_(&c__1, "SORMBR",
		     "PLN", n, nrhs, n, &c_n1, (ftnlen)6, (ftnlen)3);
	    maxwrk = max(i__1,i__2);
/* Computing 2nd power */
	    i__1 = smlsiz + 1;
	    wlalsd = *n * 9 + (*n << 1) * smlsiz + (*n << 3) * nlvl + *n * *
		    nrhs + i__1 * i__1;
/* Computing MAX */
	    i__1 = maxwrk, i__2 = *n * 3 + wlalsd;
	    maxwrk = max(i__1,i__2);
/* Computing MAX */
	    i__1 = *n * 3 + mm, i__2 = *n * 3 + *nrhs, i__1 = max(i__1,i__2), 
		    i__2 = *n * 3 + wlalsd;
	    minwrk = max(i__1,i__2);
	}
	if (*n > *m) {
/* Computing 2nd power */
	    i__1 = smlsiz + 1;
	    wlalsd = *m * 9 + (*m << 1) * smlsiz + (*m << 3) * nlvl + *m * *
		    nrhs + i__1 * i__1;
	    if (*n >= mnthr) {

/*              Path 2a - underdetermined, with many more columns   
                than rows. */

		maxwrk = *m + *m * ilaenv_(&c__1, "SGELQF", " ", m, n, &c_n1, 
			&c_n1, (ftnlen)6, (ftnlen)1);
/* Computing MAX */
		i__1 = maxwrk, i__2 = *m * *m + (*m << 2) + (*m << 1) * 
			ilaenv_(&c__1, "SGEBRD", " ", m, m, &c_n1, &c_n1, (
			ftnlen)6, (ftnlen)1);
		maxwrk = max(i__1,i__2);
/* Computing MAX */
		i__1 = maxwrk, i__2 = *m * *m + (*m << 2) + *nrhs * ilaenv_(&
			c__1, "SORMBR", "QLT", m, nrhs, m, &c_n1, (ftnlen)6, (
			ftnlen)3);
		maxwrk = max(i__1,i__2);
/* Computing MAX */
		i__1 = maxwrk, i__2 = *m * *m + (*m << 2) + (*m - 1) * 
			ilaenv_(&c__1, "SORMBR", "PLN", m, nrhs, m, &c_n1, (
			ftnlen)6, (ftnlen)3);
		maxwrk = max(i__1,i__2);
		if (*nrhs > 1) {
/* Computing MAX */
		    i__1 = maxwrk, i__2 = *m * *m + *m + *m * *nrhs;
		    maxwrk = max(i__1,i__2);
		} else {
/* Computing MAX */
		    i__1 = maxwrk, i__2 = *m * *m + (*m << 1);
		    maxwrk = max(i__1,i__2);
		}
/* Computing MAX */
		i__1 = maxwrk, i__2 = *m + *nrhs * ilaenv_(&c__1, "SORMLQ", 
			"LT", n, nrhs, m, &c_n1, (ftnlen)6, (ftnlen)2);
		maxwrk = max(i__1,i__2);
/* Computing MAX */
		i__1 = maxwrk, i__2 = *m * *m + (*m << 2) + wlalsd;
		maxwrk = max(i__1,i__2);