OpenBLAS/lapack-netlib/SRC/dbdsdc.c

#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 int logical;
typedef short int shortlogical;
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 pow_dd(ap, bp) ( pow(*(ap), *(bp)))
#define pow_si(B,E) spow_ui(*(B),*(E))
#define pow_ri(B,E) spow_ui(*(B),*(E))
#define pow_di(B,E) dpow_ui(*(B),*(E))
#define pow_zi(p, a, b) {pCd(p) = zpow_ui(Cd(a), *(b));}
#define pow_ci(p, a, b) {pCf(p) = cpow_ui(Cf(a), *(b));}
#define pow_zz(R,A,B) {pCd(R) = cpow(Cd(A),*(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);}
static char junk[] = "\n@(#)LIBF77 VERSION 19990503\n";
#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)}

/* procedure parameter types for -A and -C++ */

#define F2C_proc_par_types 1
#ifdef __cplusplus
typedef logical (*L_fp)(...);
#else
typedef logical (*L_fp)();
#endif

static float spow_ui(float x, integer n) {
	float pow=1.0; unsigned long int u;
	if(n != 0) {
		if(n < 0) n = -n, x = 1/x;
		for(u = n; ; ) {
			if(u & 01) pow *= x;
			if(u >>= 1) x *= x;
			else break;
		}
	}
	return pow;
}
static double dpow_ui(double x, integer n) {
	double pow=1.0; unsigned long int u;
	if(n != 0) {
		if(n < 0) n = -n, x = 1/x;
		for(u = n; ; ) {
			if(u & 01) pow *= x;
			if(u >>= 1) x *= x;
			else break;
		}
	}
	return pow;
}
#ifdef _MSC_VER
static _Fcomplex cpow_ui(complex x, integer n) {
	complex pow={1.0,0.0}; unsigned long int u;
		if(n != 0) {
		if(n < 0) n = -n, x.r = 1/x.r, x.i=1/x.i;
		for(u = n; ; ) {
			if(u & 01) pow.r *= x.r, pow.i *= x.i;
			if(u >>= 1) x.r *= x.r, x.i *= x.i;
			else break;
		}
	}
	_Fcomplex p={pow.r, pow.i};
	return p;
}
#else
static _Complex float cpow_ui(_Complex float x, integer n) {
	_Complex float pow=1.0; unsigned long int u;
	if(n != 0) {
		if(n < 0) n = -n, x = 1/x;
		for(u = n; ; ) {
			if(u & 01) pow *= x;
			if(u >>= 1) x *= x;
			else break;
		}
	}
	return pow;
}
#endif
#ifdef _MSC_VER
static _Dcomplex zpow_ui(_Dcomplex x, integer n) {
	_Dcomplex pow={1.0,0.0}; unsigned long int u;
	if(n != 0) {
		if(n < 0) n = -n, x._Val[0] = 1/x._Val[0], x._Val[1] =1/x._Val[1];
		for(u = n; ; ) {
			if(u & 01) pow._Val[0] *= x._Val[0], pow._Val[1] *= x._Val[1];
			if(u >>= 1) x._Val[0] *= x._Val[0], x._Val[1] *= x._Val[1];
			else break;
		}
	}
	_Dcomplex p = {pow._Val[0], pow._Val[1]};
	return p;
}
#else
static _Complex double zpow_ui(_Complex double x, integer n) {
	_Complex double pow=1.0; unsigned long int u;
	if(n != 0) {
		if(n < 0) n = -n, x = 1/x;
		for(u = n; ; ) {
			if(u & 01) pow *= x;
			if(u >>= 1) x *= x;
			else break;
		}
	}
	return pow;
}
#endif
static integer pow_ii(integer x, integer n) {
	integer pow; unsigned long int u;
	if (n <= 0) {
		if (n == 0 || x == 1) pow = 1;
		else if (x != -1) pow = x == 0 ? 1/x : 0;
		else n = -n;
	}
	if ((n > 0) || !(n == 0 || x == 1 || x != -1)) {
		u = n;
		for(pow = 1; ; ) {
			if(u & 01) pow *= x;
			if(u >>= 1) x *= x;
			else break;
		}
	}
	return pow;
}
static integer dmaxloc_(double *w, integer s, integer e, integer *n)
{
	double m; integer i, mi;
	for(m=w[s-1], mi=s, i=s+1; i<=e; i++)
		if (w[i-1]>m) mi=i ,m=w[i-1];
	return mi-s+1;
}
static integer smaxloc_(float *w, integer s, integer e, integer *n)
{
	float m; integer i, mi;
	for(m=w[s-1], mi=s, i=s+1; i<=e; i++)
		if (w[i-1]>m) mi=i ,m=w[i-1];
	return mi-s+1;
}
static inline void cdotc_(complex *z, integer *n_, complex *x, integer *incx_, complex *y, integer *incy_) {
	integer n = *n_, incx = *incx_, incy = *incy_, i;
#ifdef _MSC_VER
	_Fcomplex zdotc = {0.0, 0.0};
	if (incx == 1 && incy == 1) {
		for (i=0;i<n;i++) { /* zdotc = zdotc + dconjg(x(i))* y(i) */
			zdotc._Val[0] += conjf(Cf(&x[i]))._Val[0] * Cf(&y[i])._Val[0];
			zdotc._Val[1] += conjf(Cf(&x[i]))._Val[1] * Cf(&y[i])._Val[1];
		}
	} else {
		for (i=0;i<n;i++) { /* zdotc = zdotc + dconjg(x(i))* y(i) */
			zdotc._Val[0] += conjf(Cf(&x[i*incx]))._Val[0] * Cf(&y[i*incy])._Val[0];
			zdotc._Val[1] += conjf(Cf(&x[i*incx]))._Val[1] * Cf(&y[i*incy])._Val[1];
		}
	}
	pCf(z) = zdotc;
}
#else
	_Complex float zdotc = 0.0;
	if (incx == 1 && incy == 1) {
		for (i=0;i<n;i++) { /* zdotc = zdotc + dconjg(x(i))* y(i) */
			zdotc += conjf(Cf(&x[i])) * Cf(&y[i]);
		}
	} else {
		for (i=0;i<n;i++) { /* zdotc = zdotc + dconjg(x(i))* y(i) */
			zdotc += conjf(Cf(&x[i*incx])) * Cf(&y[i*incy]);
		}
	}
	pCf(z) = zdotc;
}
#endif
static inline void zdotc_(doublecomplex *z, integer *n_, doublecomplex *x, integer *incx_, doublecomplex *y, integer *incy_) {
	integer n = *n_, incx = *incx_, incy = *incy_, i;
#ifdef _MSC_VER
	_Dcomplex zdotc = {0.0, 0.0};
	if (incx == 1 && incy == 1) {
		for (i=0;i<n;i++) { /* zdotc = zdotc + dconjg(x(i))* y(i) */
			zdotc._Val[0] += conj(Cd(&x[i]))._Val[0] * Cd(&y[i])._Val[0];
			zdotc._Val[1] += conj(Cd(&x[i]))._Val[1] * Cd(&y[i])._Val[1];
		}
	} else {
		for (i=0;i<n;i++) { /* zdotc = zdotc + dconjg(x(i))* y(i) */
			zdotc._Val[0] += conj(Cd(&x[i*incx]))._Val[0] * Cd(&y[i*incy])._Val[0];
			zdotc._Val[1] += conj(Cd(&x[i*incx]))._Val[1] * Cd(&y[i*incy])._Val[1];
		}
	}
	pCd(z) = zdotc;
}
#else
	_Complex double zdotc = 0.0;
	if (incx == 1 && incy == 1) {
		for (i=0;i<n;i++) { /* zdotc = zdotc + dconjg(x(i))* y(i) */
			zdotc += conj(Cd(&x[i])) * Cd(&y[i]);
		}
	} else {
		for (i=0;i<n;i++) { /* zdotc = zdotc + dconjg(x(i))* y(i) */
			zdotc += conj(Cd(&x[i*incx])) * Cd(&y[i*incy]);
		}
	}
	pCd(z) = zdotc;
}
#endif	
static inline void cdotu_(complex *z, integer *n_, complex *x, integer *incx_, complex *y, integer *incy_) {
	integer n = *n_, incx = *incx_, incy = *incy_, i;
#ifdef _MSC_VER
	_Fcomplex zdotc = {0.0, 0.0};
	if (incx == 1 && incy == 1) {
		for (i=0;i<n;i++) { /* zdotc = zdotc + dconjg(x(i))* y(i) */
			zdotc._Val[0] += Cf(&x[i])._Val[0] * Cf(&y[i])._Val[0];
			zdotc._Val[1] += Cf(&x[i])._Val[1] * Cf(&y[i])._Val[1];
		}
	} else {
		for (i=0;i<n;i++) { /* zdotc = zdotc + dconjg(x(i))* y(i) */
			zdotc._Val[0] += Cf(&x[i*incx])._Val[0] * Cf(&y[i*incy])._Val[0];
			zdotc._Val[1] += Cf(&x[i*incx])._Val[1] * Cf(&y[i*incy])._Val[1];
		}
	}
	pCf(z) = zdotc;
}
#else
	_Complex float zdotc = 0.0;
	if (incx == 1 && incy == 1) {
		for (i=0;i<n;i++) { /* zdotc = zdotc + dconjg(x(i))* y(i) */
			zdotc += Cf(&x[i]) * Cf(&y[i]);
		}
	} else {
		for (i=0;i<n;i++) { /* zdotc = zdotc + dconjg(x(i))* y(i) */
			zdotc += Cf(&x[i*incx]) * Cf(&y[i*incy]);
		}
	}
	pCf(z) = zdotc;
}
#endif
static inline void zdotu_(doublecomplex *z, integer *n_, doublecomplex *x, integer *incx_, doublecomplex *y, integer *incy_) {
	integer n = *n_, incx = *incx_, incy = *incy_, i;
#ifdef _MSC_VER
	_Dcomplex zdotc = {0.0, 0.0};
	if (incx == 1 && incy == 1) {
		for (i=0;i<n;i++) { /* zdotc = zdotc + dconjg(x(i))* y(i) */
			zdotc._Val[0] += Cd(&x[i])._Val[0] * Cd(&y[i])._Val[0];
			zdotc._Val[1] += Cd(&x[i])._Val[1] * Cd(&y[i])._Val[1];
		}
	} else {
		for (i=0;i<n;i++) { /* zdotc = zdotc + dconjg(x(i))* y(i) */
			zdotc._Val[0] += Cd(&x[i*incx])._Val[0] * Cd(&y[i*incy])._Val[0];
			zdotc._Val[1] += Cd(&x[i*incx])._Val[1] * Cd(&y[i*incy])._Val[1];
		}
	}
	pCd(z) = zdotc;
}
#else
	_Complex double zdotc = 0.0;
	if (incx == 1 && incy == 1) {
		for (i=0;i<n;i++) { /* zdotc = zdotc + dconjg(x(i))* y(i) */
			zdotc += Cd(&x[i]) * Cd(&y[i]);
		}
	} else {
		for (i=0;i<n;i++) { /* zdotc = zdotc + dconjg(x(i))* y(i) */
			zdotc += Cd(&x[i*incx]) * Cd(&y[i*incy]);
		}
	}
	pCd(z) = zdotc;
}
#endif
/*  -- 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__9 = 9;
static integer c__0 = 0;
static doublereal c_b15 = 1.;
static integer c__1 = 1;
static doublereal c_b29 = 0.;

/* > \brief \b DBDSDC */

/*  =========== DOCUMENTATION =========== */

/* Online html documentation available at */
/*            http://www.netlib.org/lapack/explore-html/ */

/* > \htmlonly */
/* > Download DBDSDC + dependencies */
/* > <a href="http://www.netlib.org/cgi-bin/netlibfiles.tgz?format=tgz&filename=/lapack/lapack_routine/dbdsdc.
f"> */
/* > [TGZ]</a> */
/* > <a href="http://www.netlib.org/cgi-bin/netlibfiles.zip?format=zip&filename=/lapack/lapack_routine/dbdsdc.
f"> */
/* > [ZIP]</a> */
/* > <a href="http://www.netlib.org/cgi-bin/netlibfiles.txt?format=txt&filename=/lapack/lapack_routine/dbdsdc.
f"> */
/* > [TXT]</a> */
/* > \endhtmlonly */

/*  Definition: */
/*  =========== */

/*       SUBROUTINE DBDSDC( UPLO, COMPQ, N, D, E, U, LDU, VT, LDVT, Q, IQ, */
/*                          WORK, IWORK, INFO ) */

/*       CHARACTER          COMPQ, UPLO */
/*       INTEGER            INFO, LDU, LDVT, N */
/*       INTEGER            IQ( * ), IWORK( * ) */
/*       DOUBLE PRECISION   D( * ), E( * ), Q( * ), U( LDU, * ), */
/*      $                   VT( LDVT, * ), WORK( * ) */


/* > \par Purpose: */
/*  ============= */
/* > */
/* > \verbatim */
/* > */
/* > DBDSDC computes the singular value decomposition (SVD) of a real */
/* > N-by-N (upper or lower) bidiagonal matrix B:  B = U * S * VT, */
/* > using a divide and conquer method, where S is a diagonal matrix */
/* > with non-negative diagonal elements (the singular values of B), and */
/* > U and VT are orthogonal matrices of left and right singular vectors, */
/* > respectively. DBDSDC can be used to compute all singular values, */
/* > and optionally, singular vectors or singular vectors in compact form. */
/* > */
/* > This code 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.  See DLASD3 for details. */
/* > */
/* > The code currently calls DLASDQ if singular values only are desired. */
/* > However, it can be slightly modified to compute singular values */
/* > using the divide and conquer method. */
/* > \endverbatim */

/*  Arguments: */
/*  ========== */

/* > \param[in] UPLO */
/* > \verbatim */
/* >          UPLO is CHARACTER*1 */
/* >          = 'U':  B is upper bidiagonal. */
/* >          = 'L':  B is lower bidiagonal. */
/* > \endverbatim */
/* > */
/* > \param[in] COMPQ */
/* > \verbatim */
/* >          COMPQ is CHARACTER*1 */
/* >          Specifies whether singular vectors are to be computed */
/* >          as follows: */
/* >          = 'N':  Compute singular values only; */
/* >          = 'P':  Compute singular values and compute singular */
/* >                  vectors in compact form; */
/* >          = 'I':  Compute singular values and singular vectors. */
/* > \endverbatim */
/* > */
/* > \param[in] N */
/* > \verbatim */
/* >          N is INTEGER */
/* >          The order of the matrix B.  N >= 0. */
/* > \endverbatim */
/* > */
/* > \param[in,out] D */
/* > \verbatim */
/* >          D is DOUBLE PRECISION array, dimension (N) */
/* >          On entry, the n diagonal elements of the bidiagonal matrix B. */
/* >          On exit, if INFO=0, the singular values of B. */
/* > \endverbatim */
/* > */
/* > \param[in,out] E */
/* > \verbatim */
/* >          E is DOUBLE PRECISION array, dimension (N-1) */
/* >          On entry, the elements of E contain the offdiagonal */
/* >          elements of the bidiagonal matrix whose SVD is desired. */
/* >          On exit, E has been destroyed. */
/* > \endverbatim */
/* > */
/* > \param[out] U */
/* > \verbatim */
/* >          U is DOUBLE PRECISION array, dimension (LDU,N) */
/* >          If  COMPQ = 'I', then: */
/* >             On exit, if INFO = 0, U contains the left singular vectors */
/* >             of the bidiagonal matrix. */
/* >          For other values of COMPQ, U is not referenced. */
/* > \endverbatim */
/* > */
/* > \param[in] LDU */
/* > \verbatim */
/* >          LDU is INTEGER */
/* >          The leading dimension of the array U.  LDU >= 1. */
/* >          If singular vectors are desired, then LDU >= f2cmax( 1, N ). */
/* > \endverbatim */
/* > */
/* > \param[out] VT */
/* > \verbatim */
/* >          VT is DOUBLE PRECISION array, dimension (LDVT,N) */
/* >          If  COMPQ = 'I', then: */
/* >             On exit, if INFO = 0, VT**T contains the right singular */
/* >             vectors of the bidiagonal matrix. */
/* >          For other values of COMPQ, VT is not referenced. */
/* > \endverbatim */
/* > */
/* > \param[in] LDVT */
/* > \verbatim */
/* >          LDVT is INTEGER */
/* >          The leading dimension of the array VT.  LDVT >= 1. */
/* >          If singular vectors are desired, then LDVT >= f2cmax( 1, N ). */
/* > \endverbatim */
/* > */
/* > \param[out] Q */
/* > \verbatim */
/* >          Q is DOUBLE PRECISION array, dimension (LDQ) */
/* >          If  COMPQ = 'P', then: */
/* >             On exit, if INFO = 0, Q and IQ contain the left */
/* >             and right singular vectors in a compact form, */
/* >             requiring O(N log N) space instead of 2*N**2. */
/* >             In particular, Q contains all the DOUBLE PRECISION data in */
/* >             LDQ >= N*(11 + 2*SMLSIZ + 8*INT(LOG_2(N/(SMLSIZ+1)))) */
/* >             words of memory, where 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). */
/* >          For other values of COMPQ, Q is not referenced. */
/* > \endverbatim */
/* > */
/* > \param[out] IQ */
/* > \verbatim */
/* >          IQ is INTEGER array, dimension (LDIQ) */
/* >          If  COMPQ = 'P', then: */
/* >             On exit, if INFO = 0, Q and IQ contain the left */
/* >             and right singular vectors in a compact form, */
/* >             requiring O(N log N) space instead of 2*N**2. */
/* >             In particular, IQ contains all INTEGER data in */
/* >             LDIQ >= N*(3 + 3*INT(LOG_2(N/(SMLSIZ+1)))) */
/* >             words of memory, where 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). */
/* >          For other values of COMPQ, IQ is not referenced. */
/* > \endverbatim */
/* > */
/* > \param[out] WORK */
/* > \verbatim */
/* >          WORK is DOUBLE PRECISION array, dimension (MAX(1,LWORK)) */
/* >          If COMPQ = 'N' then LWORK >= (4 * N). */
/* >          If COMPQ = 'P' then LWORK >= (6 * N). */
/* >          If COMPQ = 'I' then LWORK >= (3 * N**2 + 4 * N). */
/* > \endverbatim */
/* > */
/* > \param[out] IWORK */
/* > \verbatim */
/* >          IWORK is INTEGER array, dimension (8*N) */
/* > \endverbatim */
/* > */
/* > \param[out] INFO */
/* > \verbatim */
/* >          INFO is INTEGER */
/* >          = 0:  successful exit. */
/* >          < 0:  if INFO = -i, the i-th argument had an illegal value. */
/* >          > 0:  The algorithm failed to compute a singular value. */
/* >                The update process of divide and conquer failed. */
/* > \endverbatim */

/*  Authors: */
/*  ======== */

/* > \author Univ. of Tennessee */
/* > \author Univ. of California Berkeley */
/* > \author Univ. of Colorado Denver */
/* > \author NAG Ltd. */

/* > \date June 2016 */

/* > \ingroup auxOTHERcomputational */

/* > \par Contributors: */
/*  ================== */
/* > */
/* >     Ming Gu and Huan Ren, Computer Science Division, University of */
/* >     California at Berkeley, USA */
/* > */
/*  ===================================================================== */
/* Subroutine */ void dbdsdc_(char *uplo, char *compq, integer *n, doublereal *
	d__, doublereal *e, doublereal *u, integer *ldu, doublereal *vt, 
	integer *ldvt, doublereal *q, integer *iq, doublereal *work, integer *
	iwork, integer *info)
{
    /* System generated locals */
    integer u_dim1, u_offset, vt_dim1, vt_offset, i__1, i__2;
    doublereal d__1;

    /* Local variables */
    integer difl, difr, ierr, perm, mlvl, sqre, i__, j, k;
    doublereal p, r__;
    integer z__;
    extern logical lsame_(char *, char *);
    extern /* Subroutine */ void dlasr_(char *, char *, char *, integer *, 
	    integer *, doublereal *, doublereal *, doublereal *, integer *), dcopy_(integer *, doublereal *, integer *
	    , doublereal *, integer *), dswap_(integer *, doublereal *, 
	    integer *, doublereal *, integer *);
    integer poles, iuplo, nsize, start;
    extern /* Subroutine */ void dlasd0_(integer *, integer *, doublereal *, 
	    doublereal *, doublereal *, integer *, doublereal *, integer *, 
	    integer *, integer *, doublereal *, integer *);
    integer ic, ii, kk;
    doublereal cs;
    extern doublereal dlamch_(char *);
    extern /* Subroutine */ void dlasda_(integer *, integer *, integer *, 
	    integer *, doublereal *, doublereal *, doublereal *, integer *, 
	    doublereal *, integer *, doublereal *, doublereal *, doublereal *,
	     doublereal *, integer *, integer *, integer *, integer *, 
	    doublereal *, doublereal *, doublereal *, doublereal *, integer *,
	     integer *);
    integer is, iu;
    doublereal sn;
    extern /* Subroutine */ void dlascl_(char *, integer *, integer *, 
	    doublereal *, doublereal *, integer *, integer *, doublereal *, 
	    integer *, integer *), dlasdq_(char *, integer *, integer 
	    *, integer *, integer *, integer *, doublereal *, doublereal *, 
	    doublereal *, integer *, doublereal *, integer *, doublereal *, 
	    integer *, doublereal *, integer *), dlaset_(char *, 
	    integer *, integer *, doublereal *, doublereal *, doublereal *, 
	    integer *), dlartg_(doublereal *, doublereal *, 
	    doublereal *, doublereal *, doublereal *);
    extern integer ilaenv_(integer *, char *, char *, integer *, integer *, 
	    integer *, integer *, ftnlen, ftnlen);
    extern /* Subroutine */ int xerbla_(char *, integer *, ftnlen);
    integer givcol;
    extern doublereal dlanst_(char *, integer *, doublereal *, doublereal *);
    integer icompq;
    doublereal orgnrm;
    integer givnum, givptr, nm1, qstart, smlsiz, wstart, smlszp;
    doublereal eps;
    integer ivt;


/*  -- LAPACK computational routine (version 3.7.1) -- */
/*  -- LAPACK is a software package provided by Univ. of Tennessee,    -- */
/*  -- Univ. of California Berkeley, Univ. of Colorado Denver and NAG Ltd..-- */
/*     June 2016 */


/*  ===================================================================== */
/*  Changed dimension statement in comment describing E from (N) to */
/*  (N-1).  Sven, 17 Feb 05. */
/*  ===================================================================== */


/*     Test the input parameters. */

    /* Parameter adjustments */
    --d__;
    --e;
    u_dim1 = *ldu;
    u_offset = 1 + u_dim1 * 1;
    u -= u_offset;
    vt_dim1 = *ldvt;
    vt_offset = 1 + vt_dim1 * 1;
    vt -= vt_offset;
    --q;
    --iq;
    --work;
    --iwork;

    /* Function Body */
    *info = 0;

    iuplo = 0;
    if (lsame_(uplo, "U")) {
	iuplo = 1;
    }
    if (lsame_(uplo, "L")) {
	iuplo = 2;
    }
    if (lsame_(compq, "N")) {
	icompq = 0;
    } else if (lsame_(compq, "P")) {
	icompq = 1;
    } else if (lsame_(compq, "I")) {
	icompq = 2;
    } else {
	icompq = -1;
    }
    if (iuplo == 0) {
	*info = -1;
    } else if (icompq < 0) {
	*info = -2;
    } else if (*n < 0) {
	*info = -3;
    } else if (*ldu < 1 || icompq == 2 && *ldu < *n) {
	*info = -7;
    } else if (*ldvt < 1 || icompq == 2 && *ldvt < *n) {
	*info = -9;
    }
    if (*info != 0) {
	i__1 = -(*info);
	xerbla_("DBDSDC", &i__1, (ftnlen)6);
	return;
    }

/*     Quick return if possible */

    if (*n == 0) {
	return;
    }
    smlsiz = ilaenv_(&c__9, "DBDSDC", " ", &c__0, &c__0, &c__0, &c__0, (
	    ftnlen)6, (ftnlen)1);
    if (*n == 1) {
	if (icompq == 1) {
	    q[1] = d_sign(&c_b15, &d__[1]);
	    q[smlsiz * *n + 1] = 1.;
	} else if (icompq == 2) {
	    u[u_dim1 + 1] = d_sign(&c_b15, &d__[1]);
	    vt[vt_dim1 + 1] = 1.;
	}
	d__[1] = abs(d__[1]);
	return;
    }
    nm1 = *n - 1;

/*     If matrix lower bidiagonal, rotate to be upper bidiagonal */
/*     by applying Givens rotations on the left */

    wstart = 1;
    qstart = 3;
    if (icompq == 1) {
	dcopy_(n, &d__[1], &c__1, &q[1], &c__1);
	i__1 = *n - 1;
	dcopy_(&i__1, &e[1], &c__1, &q[*n + 1], &c__1);
    }
    if (iuplo == 2) {
	qstart = 5;
	if (icompq == 2) {
	    wstart = (*n << 1) - 1;
	}
	i__1 = *n - 1;
	for (i__ = 1; i__ <= i__1; ++i__) {
	    dlartg_(&d__[i__], &e[i__], &cs, &sn, &r__);
	    d__[i__] = r__;
	    e[i__] = sn * d__[i__ + 1];
	    d__[i__ + 1] = cs * d__[i__ + 1];
	    if (icompq == 1) {
		q[i__ + (*n << 1)] = cs;
		q[i__ + *n * 3] = sn;
	    } else if (icompq == 2) {
		work[i__] = cs;
		work[nm1 + i__] = -sn;
	    }
/* L10: */
	}
    }

/*     If ICOMPQ = 0, use DLASDQ to compute the singular values. */

    if (icompq == 0) {
/*        Ignore WSTART, instead using WORK( 1 ), since the two vectors */
/*        for CS and -SN above are added only if ICOMPQ == 2, */
/*        and adding them exceeds documented WORK size of 4*n. */
	dlasdq_("U", &c__0, n, &c__0, &c__0, &c__0, &d__[1], &e[1], &vt[
		vt_offset], ldvt, &u[u_offset], ldu, &u[u_offset], ldu, &work[
		1], info);
	goto L40;
    }

/*     If N is smaller than the minimum divide size SMLSIZ, then solve */
/*     the problem with another solver. */

    if (*n <= smlsiz) {
	if (icompq == 2) {
	    dlaset_("A", n, n, &c_b29, &c_b15, &u[u_offset], ldu);
	    dlaset_("A", n, n, &c_b29, &c_b15, &vt[vt_offset], ldvt);
	    dlasdq_("U", &c__0, n, n, n, &c__0, &d__[1], &e[1], &vt[vt_offset]
		    , ldvt, &u[u_offset], ldu, &u[u_offset], ldu, &work[
		    wstart], info);
	} else if (icompq == 1) {
	    iu = 1;
	    ivt = iu + *n;
	    dlaset_("A", n, n, &c_b29, &c_b15, &q[iu + (qstart - 1) * *n], n);
	    dlaset_("A", n, n, &c_b29, &c_b15, &q[ivt + (qstart - 1) * *n], n);
	    dlasdq_("U", &c__0, n, n, n, &c__0, &d__[1], &e[1], &q[ivt + (
		    qstart - 1) * *n], n, &q[iu + (qstart - 1) * *n], n, &q[
		    iu + (qstart - 1) * *n], n, &work[wstart], info);
	}
	goto L40;
    }

    if (icompq == 2) {
	dlaset_("A", n, n, &c_b29, &c_b15, &u[u_offset], ldu);
	dlaset_("A", n, n, &c_b29, &c_b15, &vt[vt_offset], ldvt);
    }

/*     Scale. */

    orgnrm = dlanst_("M", n, &d__[1], &e[1]);
    if (orgnrm == 0.) {
	return;
    }
    dlascl_("G", &c__0, &c__0, &orgnrm, &c_b15, n, &c__1, &d__[1], n, &ierr);
    dlascl_("G", &c__0, &c__0, &orgnrm, &c_b15, &nm1, &c__1, &e[1], &nm1, &
	    ierr);

    eps = dlamch_("Epsilon") * .9;

    mlvl = (integer) (log((doublereal) (*n) / (doublereal) (smlsiz + 1)) / 
	    log(2.)) + 1;
    smlszp = smlsiz + 1;

    if (icompq == 1) {
	iu = 1;
	ivt = smlsiz + 1;
	difl = ivt + smlszp;
	difr = difl + mlvl;
	z__ = difr + (mlvl << 1);
	ic = z__ + mlvl;
	is = ic + 1;
	poles = is + 1;
	givnum = poles + (mlvl << 1);

	k = 1;
	givptr = 2;
	perm = 3;
	givcol = perm + mlvl;
    }

    i__1 = *n;
    for (i__ = 1; i__ <= i__1; ++i__) {
	if ((d__1 = d__[i__], abs(d__1)) < eps) {
	    d__[i__] = d_sign(&eps, &d__[i__]);
	}
/* L20: */
    }

    start = 1;
    sqre = 0;

    i__1 = nm1;
    for (i__ = 1; i__ <= i__1; ++i__) {
	if ((d__1 = e[i__], abs(d__1)) < eps || i__ == nm1) {

/*           Subproblem found. First determine its size and then */
/*           apply divide and conquer on it. */

	    if (i__ < nm1) {

/*              A subproblem with E(I) small for I < NM1. */

		nsize = i__ - start + 1;
	    } else if ((d__1 = e[i__], abs(d__1)) >= eps) {

/*              A subproblem with E(NM1) not too small but I = NM1. */

		nsize = *n - start + 1;
	    } else {

/*              A subproblem with E(NM1) small. This implies an */
/*              1-by-1 subproblem at D(N). Solve this 1-by-1 problem */
/*              first. */

		nsize = i__ - start + 1;
		if (icompq == 2) {
		    u[*n + *n * u_dim1] = d_sign(&c_b15, &d__[*n]);
		    vt[*n + *n * vt_dim1] = 1.;
		} else if (icompq == 1) {
		    q[*n + (qstart - 1) * *n] = d_sign(&c_b15, &d__[*n]);
		    q[*n + (smlsiz + qstart - 1) * *n] = 1.;
		}
		d__[*n] = (d__1 = d__[*n], abs(d__1));
	    }
	    if (icompq == 2) {
		dlasd0_(&nsize, &sqre, &d__[start], &e[start], &u[start + 
			start * u_dim1], ldu, &vt[start + start * vt_dim1], 
			ldvt, &smlsiz, &iwork[1], &work[wstart], info);
	    } else {
		dlasda_(&icompq, &smlsiz, &nsize, &sqre, &d__[start], &e[
			start], &q[start + (iu + qstart - 2) * *n], n, &q[
			start + (ivt + qstart - 2) * *n], &iq[start + k * *n],
			 &q[start + (difl + qstart - 2) * *n], &q[start + (
			difr + qstart - 2) * *n], &q[start + (z__ + qstart - 
			2) * *n], &q[start + (poles + qstart - 2) * *n], &iq[
			start + givptr * *n], &iq[start + givcol * *n], n, &
			iq[start + perm * *n], &q[start + (givnum + qstart - 
			2) * *n], &q[start + (ic + qstart - 2) * *n], &q[
			start + (is + qstart - 2) * *n], &work[wstart], &
			iwork[1], info);
	    }
	    if (*info != 0) {
		return;
	    }
	    start = i__ + 1;
	}
/* L30: */
    }

/*     Unscale */

    dlascl_("G", &c__0, &c__0, &c_b15, &orgnrm, n, &c__1, &d__[1], n, &ierr);
L40:

/*     Use Selection Sort to minimize swaps of singular vectors */

    i__1 = *n;
    for (ii = 2; ii <= i__1; ++ii) {
	i__ = ii - 1;
	kk = i__;
	p = d__[i__];
	i__2 = *n;
	for (j = ii; j <= i__2; ++j) {
	    if (d__[j] > p) {
		kk = j;
		p = d__[j];
	    }
/* L50: */
	}
	if (kk != i__) {
	    d__[kk] = d__[i__];
	    d__[i__] = p;
	    if (icompq == 1) {
		iq[i__] = kk;
	    } else if (icompq == 2) {
		dswap_(n, &u[i__ * u_dim1 + 1], &c__1, &u[kk * u_dim1 + 1], &
			c__1);
		dswap_(n, &vt[i__ + vt_dim1], ldvt, &vt[kk + vt_dim1], ldvt);
	    }
	} else if (icompq == 1) {
	    iq[i__] = i__;
	}
/* L60: */
    }

/*     If ICOMPQ = 1, use IQ(N,1) as the indicator for UPLO */

    if (icompq == 1) {
	if (iuplo == 1) {
	    iq[*n] = 1;
	} else {
	    iq[*n] = 0;
	}
    }

/*     If B is lower bidiagonal, update U by those Givens rotations */
/*     which rotated B to be upper bidiagonal */

    if (iuplo == 2 && icompq == 2) {
	dlasr_("L", "V", "B", n, n, &work[1], &work[*n], &u[u_offset], ldu);
    }

    return;

/*     End of DBDSDC */

} /* dbdsdc_ */