floraachy
/
OpenBLAS
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
OpenBLAS/lapack-netlib/SRC/sgels.c

828 lines
23 KiB
C
Raw Blame History

#include <math.h>
#include <stdlib.h>
#include <string.h>
#include <stdio.h>
#include <complex.h>
#ifdef complex
#undef complex
#endif
#ifdef I
#undef I
#endif
#if defined(_WIN64)
typedef long long BLASLONG;
typedef unsigned long long BLASULONG;
#else
typedef long BLASLONG;
typedef unsigned long BLASULONG;
#endif
#ifdef LAPACK_ILP64
typedef BLASLONG blasint;
#if defined(_WIN64)
#define blasabs(x) llabs(x)
#else
#define blasabs(x) labs(x)
#endif
#else
typedef int blasint;
#define blasabs(x) abs(x)
#endif
typedef blasint integer;
typedef unsigned int uinteger;
typedef char *address;
typedef short int shortint;
typedef float real;
typedef double doublereal;
typedef struct { real r, i; } complex;
typedef struct { doublereal r, i; } doublecomplex;
#ifdef _MSC_VER
static inline _Fcomplex Cf(complex *z) {_Fcomplex zz={z->r , z->i}; return zz;}
static inline _Dcomplex Cd(doublecomplex *z) {_Dcomplex zz={z->r , z->i};return zz;}
static inline _Fcomplex * _pCf(complex *z) {return (_Fcomplex*)z;}
static inline _Dcomplex * _pCd(doublecomplex *z) {return (_Dcomplex*)z;}
#else
static inline _Complex float Cf(complex *z) {return z->r + z->i*_Complex_I;}
static inline _Complex double Cd(doublecomplex *z) {return z->r + z->i*_Complex_I;}
static inline _Complex float * _pCf(complex *z) {return (_Complex float*)z;}
static inline _Complex double * _pCd(doublecomplex *z) {return (_Complex double*)z;}
#endif
#define pCf(z) (*_pCf(z))
#define pCd(z) (*_pCd(z))
typedef blasint logical;
typedef char logical1;
typedef char integer1;
#define TRUE_ (1)
#define FALSE_ (0)
/* Extern is for use with -E */
#ifndef Extern
#define Extern extern
#endif
/* I/O stuff */
typedef int flag;
typedef int ftnlen;
typedef int ftnint;
/*external read, write*/
typedef struct
{ flag cierr;
ftnint ciunit;
flag ciend;
char *cifmt;
ftnint cirec;
} cilist;
/*internal read, write*/
typedef struct
{ flag icierr;
char *iciunit;
flag iciend;
char *icifmt;
ftnint icirlen;
ftnint icirnum;
} icilist;
/*open*/
typedef struct
{ flag oerr;
ftnint ounit;
char *ofnm;
ftnlen ofnmlen;
char *osta;
char *oacc;
char *ofm;
ftnint orl;
char *oblnk;
} olist;
/*close*/
typedef struct
{ flag cerr;
ftnint cunit;
char *csta;
} cllist;
/*rewind, backspace, endfile*/
typedef struct
{ flag aerr;
ftnint aunit;
} alist;
/* inquire */
typedef struct
{ flag inerr;
ftnint inunit;
char *infile;
ftnlen infilen;
ftnint *inex; /*parameters in standard's order*/
ftnint *inopen;
ftnint *innum;
ftnint *innamed;
char *inname;
ftnlen innamlen;
char *inacc;
ftnlen inacclen;
char *inseq;
ftnlen inseqlen;
char *indir;
ftnlen indirlen;
char *infmt;
ftnlen infmtlen;
char *inform;
ftnint informlen;
char *inunf;
ftnlen inunflen;
ftnint *inrecl;
ftnint *innrec;
char *inblank;
ftnlen inblanklen;
} inlist;
#define VOID void
union Multitype { /* for multiple entry points */
integer1 g;
shortint h;
integer i;
/* longint j; */
real r;
doublereal d;
complex c;
doublecomplex z;
};
typedef union Multitype Multitype;
struct Vardesc { /* for Namelist */
char *name;
char *addr;
ftnlen *dims;
int type;
};
typedef struct Vardesc Vardesc;
struct Namelist {
char *name;
Vardesc **vars;
int nvars;
};
typedef struct Namelist Namelist;
#define abs(x) ((x) >= 0 ? (x) : -(x))
#define dabs(x) (fabs(x))
#define f2cmin(a,b) ((a) <= (b) ? (a) : (b))
#define f2cmax(a,b) ((a) >= (b) ? (a) : (b))
#define dmin(a,b) (f2cmin(a,b))
#define dmax(a,b) (f2cmax(a,b))
#define bit_test(a,b) ((a) >> (b) & 1)
#define bit_clear(a,b) ((a) & ~((uinteger)1 << (b)))
#define bit_set(a,b) ((a) | ((uinteger)1 << (b)))
#define abort_() { sig_die("Fortran abort routine called", 1); }
#define c_abs(z) (cabsf(Cf(z)))
#define c_cos(R,Z) { pCf(R)=ccos(Cf(Z)); }
#ifdef _MSC_VER
#define c_div(c, a, b) {Cf(c)._Val[0] = (Cf(a)._Val[0]/Cf(b)._Val[0]); Cf(c)._Val[1]=(Cf(a)._Val[1]/Cf(b)._Val[1]);}
#define z_div(c, a, b) {Cd(c)._Val[0] = (Cd(a)._Val[0]/Cd(b)._Val[0]); Cd(c)._Val[1]=(Cd(a)._Val[1]/df(b)._Val[1]);}
#else
#define c_div(c, a, b) {pCf(c) = Cf(a)/Cf(b);}
#define z_div(c, a, b) {pCd(c) = Cd(a)/Cd(b);}
#endif
#define c_exp(R, Z) {pCf(R) = cexpf(Cf(Z));}
#define c_log(R, Z) {pCf(R) = clogf(Cf(Z));}
#define c_sin(R, Z) {pCf(R) = csinf(Cf(Z));}
//#define c_sqrt(R, Z) {*(R) = csqrtf(Cf(Z));}
#define c_sqrt(R, Z) {pCf(R) = csqrtf(Cf(Z));}
#define d_abs(x) (fabs(*(x)))
#define d_acos(x) (acos(*(x)))
#define d_asin(x) (asin(*(x)))
#define d_atan(x) (atan(*(x)))
#define d_atn2(x, y) (atan2(*(x),*(y)))
#define d_cnjg(R, Z) { pCd(R) = conj(Cd(Z)); }
#define r_cnjg(R, Z) { pCf(R) = conjf(Cf(Z)); }
#define d_cos(x) (cos(*(x)))
#define d_cosh(x) (cosh(*(x)))
#define d_dim(__a, __b) ( *(__a) > *(__b) ? *(__a) - *(__b) : 0.0 )
#define d_exp(x) (exp(*(x)))
#define d_imag(z) (cimag(Cd(z)))
#define r_imag(z) (cimagf(Cf(z)))
#define d_int(__x) (*(__x)>0 ? floor(*(__x)) : -floor(- *(__x)))
#define r_int(__x) (*(__x)>0 ? floor(*(__x)) : -floor(- *(__x)))
#define d_lg10(x) ( 0.43429448190325182765 * log(*(x)) )
#define r_lg10(x) ( 0.43429448190325182765 * log(*(x)) )
#define d_log(x) (log(*(x)))
#define d_mod(x, y) (fmod(*(x), *(y)))
#define u_nint(__x) ((__x)>=0 ? floor((__x) + .5) : -floor(.5 - (__x)))
#define d_nint(x) u_nint(*(x))
#define u_sign(__a,__b) ((__b) >= 0 ? ((__a) >= 0 ? (__a) : -(__a)) : -((__a) >= 0 ? (__a) : -(__a)))
#define d_sign(a,b) u_sign(*(a),*(b))
#define r_sign(a,b) u_sign(*(a),*(b))
#define d_sin(x) (sin(*(x)))
#define d_sinh(x) (sinh(*(x)))
#define d_sqrt(x) (sqrt(*(x)))
#define d_tan(x) (tan(*(x)))
#define d_tanh(x) (tanh(*(x)))
#define i_abs(x) abs(*(x))
#define i_dnnt(x) ((integer)u_nint(*(x)))
#define i_len(s, n) (n)
#define i_nint(x) ((integer)u_nint(*(x)))
#define i_sign(a,b) ((integer)u_sign((integer)*(a),(integer)*(b)))
#define s_cat(lpp, rpp, rnp, np, llp) { ftnlen i, nc, ll; char *f__rp, *lp; ll = (llp); lp = (lpp); for(i=0; i < (int)*(np); ++i) { nc = ll; if((rnp)[i] < nc) nc = (rnp)[i]; ll -= nc; f__rp = (rpp)[i]; while(--nc >= 0) *lp++ = *(f__rp)++; } while(--ll >= 0) *lp++ = ' '; }
#define s_cmp(a,b,c,d) ((integer)strncmp((a),(b),f2cmin((c),(d))))
#define s_copy(A,B,C,D) { int __i,__m; for (__i=0, __m=f2cmin((C),(D)); __i<__m && (B)[__i] != 0; ++__i) (A)[__i] = (B)[__i]; }
#define sig_die(s, kill) { exit(1); }
#define s_stop(s, n) {exit(0);}
#define z_abs(z) (cabs(Cd(z)))
#define z_exp(R, Z) {pCd(R) = cexp(Cd(Z));}
#define z_sqrt(R, Z) {pCd(R) = csqrt(Cd(Z));}
#define myexit_() break;
#define mycycle() continue;
#define myceiling(w) {ceil(w)}
#define myhuge(w) {HUGE_VAL}
//#define mymaxloc_(w,s,e,n) {if (sizeof(*(w)) == sizeof(double)) dmaxloc_((w),*(s),*(e),n); else dmaxloc_((w),*(s),*(e),n);}
#define mymaxloc(w,s,e,n) {dmaxloc_(w,*(s),*(e),n)}
/* -- translated by f2c (version 20000121).
You must link the resulting object file with the libraries:
-lf2c -lm (in that order)
*/
/* Table of constant values */
static integer c__1 = 1;
static integer c_n1 = -1;
static real c_b33 = 0.f;
static integer c__0 = 0;
/* > \brief <b> SGELS solves overdetermined or underdetermined systems for GE matrices</b> */
/* =========== DOCUMENTATION =========== */
/* Online html documentation available at */
/* http://www.netlib.org/lapack/explore-html/ */
/* > \htmlonly */
/* > Download SGELS + dependencies */
/* > <a href="http://www.netlib.org/cgi-bin/netlibfiles.tgz?format=tgz&filename=/lapack/lapack_routine/sgels.f
"> */
/* > [TGZ]</a> */
/* > <a href="http://www.netlib.org/cgi-bin/netlibfiles.zip?format=zip&filename=/lapack/lapack_routine/sgels.f
"> */
/* > [ZIP]</a> */
/* > <a href="http://www.netlib.org/cgi-bin/netlibfiles.txt?format=txt&filename=/lapack/lapack_routine/sgels.f
"> */
/* > [TXT]</a> */
/* > \endhtmlonly */
/* Definition: */
/* =========== */
/* SUBROUTINE SGELS( TRANS, M, N, NRHS, A, LDA, B, LDB, WORK, LWORK, */
/* INFO ) */
/* CHARACTER TRANS */
/* INTEGER INFO, LDA, LDB, LWORK, M, N, NRHS */
/* REAL A( LDA, * ), B( LDB, * ), WORK( * ) */
/* > \par Purpose: */
/* ============= */
/* > */
/* > \verbatim */
/* > */
/* > SGELS solves overdetermined or underdetermined real linear systems */
/* > involving an M-by-N matrix A, or its transpose, using a QR or LQ */
/* > factorization of A. It is assumed that A has full rank. */
/* > */
/* > The following options are provided: */
/* > */
/* > 1. If TRANS = 'N' and m >= n: find the least squares solution of */
/* > an overdetermined system, i.e., solve the least squares problem */
/* > minimize || B - A*X ||. */
/* > */
/* > 2. If TRANS = 'N' and m < n: find the minimum norm solution of */
/* > an underdetermined system A * X = B. */
/* > */
/* > 3. If TRANS = 'T' and m >= n: find the minimum norm solution of */
/* > an underdetermined system A**T * X = B. */
/* > */
/* > 4. If TRANS = 'T' and m < n: find the least squares solution of */
/* > an overdetermined system, i.e., solve the least squares problem */
/* > minimize || B - A**T * X ||. */
/* > */
/* > 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. */
/* > \endverbatim */
/* Arguments: */
/* ========== */
/* > \param[in] TRANS */
/* > \verbatim */
/* > TRANS is CHARACTER*1 */
/* > = 'N': the linear system involves A; */
/* > = 'T': the linear system involves A**T. */
/* > \endverbatim */
/* > */
/* > \param[in] M */
/* > \verbatim */
/* > M is INTEGER */
/* > The number of rows of the matrix A. M >= 0. */
/* > \endverbatim */
/* > */
/* > \param[in] N */
/* > \verbatim */
/* > N is INTEGER */
/* > The number of columns of the matrix A. N >= 0. */
/* > \endverbatim */
/* > */
/* > \param[in] NRHS */
/* > \verbatim */
/* > NRHS is INTEGER */
/* > The number of right hand sides, i.e., the number of */
/* > columns of the matrices B and X. NRHS >=0. */
/* > \endverbatim */
/* > */
/* > \param[in,out] A */
/* > \verbatim */
/* > A is REAL array, dimension (LDA,N) */
/* > On entry, the M-by-N matrix A. */
/* > On exit, */
/* > if M >= N, A is overwritten by details of its QR */
/* > factorization as returned by SGEQRF; */
/* > if M < N, A is overwritten by details of its LQ */
/* > factorization as returned by SGELQF. */
/* > \endverbatim */
/* > */
/* > \param[in] LDA */
/* > \verbatim */
/* > LDA is INTEGER */
/* > The leading dimension of the array A. LDA >= f2cmax(1,M). */
/* > \endverbatim */
/* > */
/* > \param[in,out] B */
/* > \verbatim */
/* > B is REAL array, dimension (LDB,NRHS) */
/* > On entry, the matrix B of right hand side vectors, stored */
/* > columnwise; B is M-by-NRHS if TRANS = 'N', or N-by-NRHS */
/* > if TRANS = 'T'. */
/* > On exit, if INFO = 0, B is overwritten by the solution */
/* > vectors, stored columnwise: */
/* > if TRANS = 'N' and m >= n, rows 1 to n of B contain the least */
/* > squares solution vectors; the residual sum of squares for the */
/* > solution in each column is given by the sum of squares of */
/* > elements N+1 to M in that column; */
/* > if TRANS = 'N' and m < n, rows 1 to N of B contain the */
/* > minimum norm solution vectors; */
/* > if TRANS = 'T' and m >= n, rows 1 to M of B contain the */
/* > minimum norm solution vectors; */
/* > if TRANS = 'T' and m < n, rows 1 to M of B contain the */
/* > least squares solution vectors; the residual sum of squares */
/* > for the solution in each column is given by the sum of */
/* > squares of elements M+1 to N in that column. */
/* > \endverbatim */
/* > */
/* > \param[in] LDB */
/* > \verbatim */
/* > LDB is INTEGER */
/* > The leading dimension of the array B. LDB >= MAX(1,M,N). */
/* > \endverbatim */
/* > */
/* > \param[out] WORK */
/* > \verbatim */
/* > WORK is REAL array, dimension (MAX(1,LWORK)) */
/* > On exit, if INFO = 0, WORK(1) returns the optimal LWORK. */
/* > \endverbatim */
/* > */
/* > \param[in] LWORK */
/* > \verbatim */
/* > LWORK is INTEGER */
/* > The dimension of the array WORK. */
/* > LWORK >= f2cmax( 1, MN + f2cmax( MN, NRHS ) ). */
/* > For optimal performance, */
/* > LWORK >= f2cmax( 1, MN + f2cmax( MN, NRHS )*NB ). */
/* > where MN = f2cmin(M,N) and NB is the optimum block size. */
/* > */
/* > 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. */
/* > \endverbatim */
/* > */
/* > \param[out] INFO */
/* > \verbatim */
/* > INFO is INTEGER */
/* > = 0: successful exit */
/* > < 0: if INFO = -i, the i-th argument had an illegal value */
/* > > 0: if INFO = i, the i-th diagonal element of the */
/* > triangular factor of A is zero, so that A does not have */
/* > full rank; the least squares solution could not be */
/* > computed. */
/* > \endverbatim */
/* Authors: */
/* ======== */
/* > \author Univ. of Tennessee */
/* > \author Univ. of California Berkeley */
/* > \author Univ. of Colorado Denver */
/* > \author NAG Ltd. */
/* > \date December 2016 */
/* > \ingroup realGEsolve */
/* ===================================================================== */
/* Subroutine */ void sgels_(char *trans, integer *m, integer *n, integer *
nrhs, real *a, integer *lda, real *b, integer *ldb, real *work,
integer *lwork, integer *info)
{
/* System generated locals */
integer a_dim1, a_offset, b_dim1, b_offset, i__1, i__2;
/* Local variables */
real anrm, bnrm;
integer brow;
logical tpsd;
integer i__, j, iascl, ibscl;
extern logical lsame_(char *, char *);
integer wsize;
real rwork[1];
integer nb;
extern /* Subroutine */ void slabad_(real *, real *);
integer mn;
extern real slamch_(char *), slange_(char *, integer *, integer *,
real *, integer *, real *);
extern /* Subroutine */ int xerbla_(char *, integer *, ftnlen);
extern integer ilaenv_(integer *, char *, char *, integer *, integer *,
integer *, integer *, ftnlen, ftnlen);
integer scllen;
real bignum;
extern /* Subroutine */ void sgelqf_(integer *, integer *, real *, integer
*, real *, real *, integer *, integer *), slascl_(char *, integer
*, integer *, real *, real *, integer *, integer *, real *,
integer *, integer *), sgeqrf_(integer *, integer *, real
*, integer *, real *, real *, integer *, integer *), slaset_(char
*, integer *, integer *, real *, real *, real *, integer *);
real smlnum;
extern /* Subroutine */ void sormlq_(char *, char *, integer *, integer *,
integer *, real *, integer *, real *, real *, integer *, real *,
integer *, integer *);
logical lquery;
extern /* Subroutine */ void sormqr_(char *, char *, integer *, integer *,
integer *, real *, integer *, real *, real *, integer *, real *,
integer *, integer *);
extern void strtrs_(char *, char *,
char *, integer *, integer *, real *, integer *, real *, integer *
, integer *);
/* -- LAPACK driver routine (version 3.7.0) -- */
/* -- LAPACK is a software package provided by Univ. of Tennessee, -- */
/* -- Univ. of California Berkeley, Univ. of Colorado Denver and NAG Ltd..-- */
/* December 2016 */
/* ===================================================================== */
/* 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;
--work;
/* Function Body */
*info = 0;
mn = f2cmin(*m,*n);
lquery = *lwork == -1;
if (! (lsame_(trans, "N") || lsame_(trans, "T"))) {
*info = -1;
} else if (*m < 0) {
*info = -2;
} else if (*n < 0) {
*info = -3;
} else if (*nrhs < 0) {
*info = -4;
} else if (*lda < f2cmax(1,*m)) {
*info = -6;
} else /* if(complicated condition) */ {
/* Computing MAX */
i__1 = f2cmax(1,*m);
if (*ldb < f2cmax(i__1,*n)) {
*info = -8;
} else /* if(complicated condition) */ {
/* Computing MAX */
i__1 = 1, i__2 = mn + f2cmax(mn,*nrhs);
if (*lwork < f2cmax(i__1,i__2) && ! lquery) {
*info = -10;
}
}
}
/* Figure out optimal block size */
if (*info == 0 || *info == -10) {
tpsd = TRUE_;
if (lsame_(trans, "N")) {
tpsd = FALSE_;
}
if (*m >= *n) {
nb = ilaenv_(&c__1, "SGEQRF", " ", m, n, &c_n1, &c_n1, (ftnlen)6,
(ftnlen)1);
if (tpsd) {
/* Computing MAX */
i__1 = nb, i__2 = ilaenv_(&c__1, "SORMQR", "LN", m, nrhs, n, &
c_n1, (ftnlen)6, (ftnlen)2);
nb = f2cmax(i__1,i__2);
} else {
/* Computing MAX */
i__1 = nb, i__2 = ilaenv_(&c__1, "SORMQR", "LT", m, nrhs, n, &
c_n1, (ftnlen)6, (ftnlen)2);
nb = f2cmax(i__1,i__2);
}
} else {
nb = ilaenv_(&c__1, "SGELQF", " ", m, n, &c_n1, &c_n1, (ftnlen)6,
(ftnlen)1);
if (tpsd) {
/* Computing MAX */
i__1 = nb, i__2 = ilaenv_(&c__1, "SORMLQ", "LT", n, nrhs, m, &
c_n1, (ftnlen)6, (ftnlen)2);
nb = f2cmax(i__1,i__2);
} else {
/* Computing MAX */
i__1 = nb, i__2 = ilaenv_(&c__1, "SORMLQ", "LN", n, nrhs, m, &
c_n1, (ftnlen)6, (ftnlen)2);
nb = f2cmax(i__1,i__2);
}
}
/* Computing MAX */
i__1 = 1, i__2 = mn + f2cmax(mn,*nrhs) * nb;
wsize = f2cmax(i__1,i__2);
work[1] = (real) wsize;
}
if (*info != 0) {
i__1 = -(*info);
xerbla_("SGELS ", &i__1, (ftnlen)6);
return;
} else if (lquery) {
return;
}
/* Quick return if possible */
/* Computing MIN */
i__1 = f2cmin(*m,*n);
if (f2cmin(i__1,*nrhs) == 0) {
i__1 = f2cmax(*m,*n);
slaset_("Full", &i__1, nrhs, &c_b33, &c_b33, &b[b_offset], ldb);
return;
}
/* Get machine parameters */
smlnum = slamch_("S") / slamch_("P");
bignum = 1.f / smlnum;
slabad_(&smlnum, &bignum);
/* Scale A, B if f2cmax element outside range [SMLNUM,BIGNUM] */
anrm = slange_("M", m, n, &a[a_offset], lda, rwork);
iascl = 0;
if (anrm > 0.f && anrm < smlnum) {
/* Scale matrix norm up to SMLNUM */
slascl_("G", &c__0, &c__0, &anrm, &smlnum, m, n, &a[a_offset], lda,
info);
iascl = 1;
} else if (anrm > bignum) {
/* Scale matrix norm down to BIGNUM */
slascl_("G", &c__0, &c__0, &anrm, &bignum, m, n, &a[a_offset], lda,
info);
iascl = 2;
} else if (anrm == 0.f) {
/* Matrix all zero. Return zero solution. */
i__1 = f2cmax(*m,*n);
slaset_("F", &i__1, nrhs, &c_b33, &c_b33, &b[b_offset], ldb);
goto L50;
}
brow = *m;
if (tpsd) {
brow = *n;
}
bnrm = slange_("M", &brow, nrhs, &b[b_offset], ldb, rwork);
ibscl = 0;
if (bnrm > 0.f && bnrm < smlnum) {
/* Scale matrix norm up to SMLNUM */
slascl_("G", &c__0, &c__0, &bnrm, &smlnum, &brow, nrhs, &b[b_offset],
ldb, info);
ibscl = 1;
} else if (bnrm > bignum) {
/* Scale matrix norm down to BIGNUM */
slascl_("G", &c__0, &c__0, &bnrm, &bignum, &brow, nrhs, &b[b_offset],
ldb, info);
ibscl = 2;
}
if (*m >= *n) {
/* compute QR factorization of A */
i__1 = *lwork - mn;
sgeqrf_(m, n, &a[a_offset], lda, &work[1], &work[mn + 1], &i__1, info)
;
/* workspace at least N, optimally N*NB */
if (! tpsd) {
/* Least-Squares Problem f2cmin || A * X - B || */
/* B(1:M,1:NRHS) := Q**T * B(1:M,1:NRHS) */
i__1 = *lwork - mn;
sormqr_("Left", "Transpose", m, nrhs, n, &a[a_offset], lda, &work[
1], &b[b_offset], ldb, &work[mn + 1], &i__1, info);
/* workspace at least NRHS, optimally NRHS*NB */
/* B(1:N,1:NRHS) := inv(R) * B(1:N,1:NRHS) */
strtrs_("Upper", "No transpose", "Non-unit", n, nrhs, &a[a_offset]
, lda, &b[b_offset], ldb, info);
if (*info > 0) {
return;
}
scllen = *n;
} else {
/* Underdetermined system of equations A**T * X = B */
/* B(1:N,1:NRHS) := inv(R**T) * B(1:N,1:NRHS) */
strtrs_("Upper", "Transpose", "Non-unit", n, nrhs, &a[a_offset],
lda, &b[b_offset], ldb, info);
if (*info > 0) {
return;
}
/* B(N+1:M,1:NRHS) = ZERO */
i__1 = *nrhs;
for (j = 1; j <= i__1; ++j) {
i__2 = *m;
for (i__ = *n + 1; i__ <= i__2; ++i__) {
b[i__ + j * b_dim1] = 0.f;
/* L10: */
}
/* L20: */
}
/* B(1:M,1:NRHS) := Q(1:N,:) * B(1:N,1:NRHS) */
i__1 = *lwork - mn;
sormqr_("Left", "No transpose", m, nrhs, n, &a[a_offset], lda, &
work[1], &b[b_offset], ldb, &work[mn + 1], &i__1, info);
/* workspace at least NRHS, optimally NRHS*NB */
scllen = *m;
}
} else {
/* Compute LQ factorization of A */
i__1 = *lwork - mn;
sgelqf_(m, n, &a[a_offset], lda, &work[1], &work[mn + 1], &i__1, info)
;
/* workspace at least M, optimally M*NB. */
if (! tpsd) {
/* underdetermined system of equations A * X = B */
/* B(1:M,1:NRHS) := inv(L) * B(1:M,1:NRHS) */
strtrs_("Lower", "No transpose", "Non-unit", m, nrhs, &a[a_offset]
, lda, &b[b_offset], ldb, info);
if (*info > 0) {
return;
}
/* B(M+1:N,1:NRHS) = 0 */
i__1 = *nrhs;
for (j = 1; j <= i__1; ++j) {
i__2 = *n;
for (i__ = *m + 1; i__ <= i__2; ++i__) {
b[i__ + j * b_dim1] = 0.f;
/* L30: */
}
/* L40: */
}
/* B(1:N,1:NRHS) := Q(1:N,:)**T * B(1:M,1:NRHS) */
i__1 = *lwork - mn;
sormlq_("Left", "Transpose", n, nrhs, m, &a[a_offset], lda, &work[
1], &b[b_offset], ldb, &work[mn + 1], &i__1, info);
/* workspace at least NRHS, optimally NRHS*NB */
scllen = *n;
} else {
/* overdetermined system f2cmin || A**T * X - B || */
/* B(1:N,1:NRHS) := Q * B(1:N,1:NRHS) */
i__1 = *lwork - mn;
sormlq_("Left", "No transpose", n, nrhs, m, &a[a_offset], lda, &
work[1], &b[b_offset], ldb, &work[mn + 1], &i__1, info);
/* workspace at least NRHS, optimally NRHS*NB */
/* B(1:M,1:NRHS) := inv(L**T) * B(1:M,1:NRHS) */
strtrs_("Lower", "Transpose", "Non-unit", m, nrhs, &a[a_offset],
lda, &b[b_offset], ldb, info);
if (*info > 0) {
return;
}
scllen = *m;
}
}
/* Undo scaling */
if (iascl == 1) {
slascl_("G", &c__0, &c__0, &anrm, &smlnum, &scllen, nrhs, &b[b_offset]
, ldb, info);
} else if (iascl == 2) {
slascl_("G", &c__0, &c__0, &anrm, &bignum, &scllen, nrhs, &b[b_offset]
, ldb, info);
}
if (ibscl == 1) {
slascl_("G", &c__0, &c__0, &smlnum, &bnrm, &scllen, nrhs, &b[b_offset]
, ldb, info);
} else if (ibscl == 2) {
slascl_("G", &c__0, &c__0, &bignum, &bnrm, &scllen, nrhs, &b[b_offset]
, ldb, info);
}
L50:
work[1] = (real) wsize;
return;
/* End of SGELS */
} /* sgels_ */
Reference in New Issue View Git Blame Copy Permalink