OpenBLAS/lapack-netlib/SRC/cpbtrf.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 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 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++ */


#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 complex c_b1 = {1.f,0.f};
static integer c__1 = 1;
static integer c_n1 = -1;
static real c_b21 = -1.f;
static real c_b22 = 1.f;
static integer c__33 = 33;

/* > \brief \b CPBTRF */

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

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

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

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

/*       SUBROUTINE CPBTRF( UPLO, N, KD, AB, LDAB, INFO ) */

/*       CHARACTER          UPLO */
/*       INTEGER            INFO, KD, LDAB, N */
/*       COMPLEX            AB( LDAB, * ) */


/* > \par Purpose: */
/*  ============= */
/* > */
/* > \verbatim */
/* > */
/* > CPBTRF computes the Cholesky factorization of a complex Hermitian */
/* > positive definite band matrix A. */
/* > */
/* > The factorization has the form */
/* >    A = U**H * U,  if UPLO = 'U', or */
/* >    A = L  * L**H,  if UPLO = 'L', */
/* > where U is an upper triangular matrix and L is lower triangular. */
/* > \endverbatim */

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

/* > \param[in] UPLO */
/* > \verbatim */
/* >          UPLO is CHARACTER*1 */
/* >          = 'U':  Upper triangle of A is stored; */
/* >          = 'L':  Lower triangle of A is stored. */
/* > \endverbatim */
/* > */
/* > \param[in] N */
/* > \verbatim */
/* >          N is INTEGER */
/* >          The order of the matrix A.  N >= 0. */
/* > \endverbatim */
/* > */
/* > \param[in] KD */
/* > \verbatim */
/* >          KD is INTEGER */
/* >          The number of superdiagonals of the matrix A if UPLO = 'U', */
/* >          or the number of subdiagonals if UPLO = 'L'.  KD >= 0. */
/* > \endverbatim */
/* > */
/* > \param[in,out] AB */
/* > \verbatim */
/* >          AB is COMPLEX array, dimension (LDAB,N) */
/* >          On entry, the upper or lower triangle of the Hermitian band */
/* >          matrix A, stored in the first KD+1 rows of the array.  The */
/* >          j-th column of A is stored in the j-th column of the array AB */
/* >          as follows: */
/* >          if UPLO = 'U', AB(kd+1+i-j,j) = A(i,j) for f2cmax(1,j-kd)<=i<=j; */
/* >          if UPLO = 'L', AB(1+i-j,j)    = A(i,j) for j<=i<=f2cmin(n,j+kd). */
/* > */
/* >          On exit, if INFO = 0, the triangular factor U or L from the */
/* >          Cholesky factorization A = U**H*U or A = L*L**H of the band */
/* >          matrix A, in the same storage format as A. */
/* > \endverbatim */
/* > */
/* > \param[in] LDAB */
/* > \verbatim */
/* >          LDAB is INTEGER */
/* >          The leading dimension of the array AB.  LDAB >= KD+1. */
/* > \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 leading minor of order i is not */
/* >                positive definite, and the factorization could not be */
/* >                completed. */
/* > \endverbatim */

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

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

/* > \date December 2016 */

/* > \ingroup complexOTHERcomputational */

/* > \par Further Details: */
/*  ===================== */
/* > */
/* > \verbatim */
/* > */
/* >  The band storage scheme is illustrated by the following example, when */
/* >  N = 6, KD = 2, and UPLO = 'U': */
/* > */
/* >  On entry:                       On exit: */
/* > */
/* >      *    *   a13  a24  a35  a46      *    *   u13  u24  u35  u46 */
/* >      *   a12  a23  a34  a45  a56      *   u12  u23  u34  u45  u56 */
/* >     a11  a22  a33  a44  a55  a66     u11  u22  u33  u44  u55  u66 */
/* > */
/* >  Similarly, if UPLO = 'L' the format of A is as follows: */
/* > */
/* >  On entry:                       On exit: */
/* > */
/* >     a11  a22  a33  a44  a55  a66     l11  l22  l33  l44  l55  l66 */
/* >     a21  a32  a43  a54  a65   *      l21  l32  l43  l54  l65   * */
/* >     a31  a42  a53  a64   *    *      l31  l42  l53  l64   *    * */
/* > */
/* >  Array elements marked * are not used by the routine. */
/* > \endverbatim */

/* > \par Contributors: */
/*  ================== */
/* > */
/* >  Peter Mayes and Giuseppe Radicati, IBM ECSEC, Rome, March 23, 1989 */

/*  ===================================================================== */
/* Subroutine */ void cpbtrf_(char *uplo, integer *n, integer *kd, complex *ab,
	 integer *ldab, integer *info)
{
    /* System generated locals */
    integer ab_dim1, ab_offset, i__1, i__2, i__3, i__4, i__5, i__6;
    complex q__1;

    /* Local variables */
    complex work[1056]	/* was [33][32] */;
    integer i__, j;
    extern /* Subroutine */ void cgemm_(char *, char *, integer *, integer *, 
	    integer *, complex *, complex *, integer *, complex *, integer *, 
	    complex *, complex *, integer *), cherk_(char *, 
	    char *, integer *, integer *, real *, complex *, integer *, real *
	    , complex *, integer *);
    extern logical lsame_(char *, char *);
    extern /* Subroutine */ void ctrsm_(char *, char *, char *, char *, 
	    integer *, integer *, complex *, complex *, integer *, complex *, 
	    integer *);
    integer i2, i3;
    extern /* Subroutine */ void cpbtf2_(char *, integer *, integer *, complex 
	    *, integer *, integer *);
    extern int cpotf2_(char *, integer *, 
	    complex *, integer *, integer *);
    integer ib, nb, ii, jj;
    extern /* Subroutine */ int xerbla_(char *, integer *, ftnlen);
    extern integer ilaenv_(integer *, char *, char *, integer *, integer *, 
	    integer *, integer *, ftnlen, ftnlen);


/*  -- LAPACK computational 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 parameters. */

    /* Parameter adjustments */
    ab_dim1 = *ldab;
    ab_offset = 1 + ab_dim1 * 1;
    ab -= ab_offset;

    /* Function Body */
    *info = 0;
    if (! lsame_(uplo, "U") && ! lsame_(uplo, "L")) {
	*info = -1;
    } else if (*n < 0) {
	*info = -2;
    } else if (*kd < 0) {
	*info = -3;
    } else if (*ldab < *kd + 1) {
	*info = -5;
    }
    if (*info != 0) {
	i__1 = -(*info);
	xerbla_("CPBTRF", &i__1, (ftnlen)6);
	return;
    }

/*     Quick return if possible */

    if (*n == 0) {
	return;
    }

/*     Determine the block size for this environment */

    nb = ilaenv_(&c__1, "CPBTRF", uplo, n, kd, &c_n1, &c_n1, (ftnlen)6, (
	    ftnlen)1);

/*     The block size must not exceed the semi-bandwidth KD, and must not */
/*     exceed the limit set by the size of the local array WORK. */

    nb = f2cmin(nb,32);

    if (nb <= 1 || nb > *kd) {

/*        Use unblocked code */

	cpbtf2_(uplo, n, kd, &ab[ab_offset], ldab, info);
    } else {

/*        Use blocked code */

	if (lsame_(uplo, "U")) {

/*           Compute the Cholesky factorization of a Hermitian band */
/*           matrix, given the upper triangle of the matrix in band */
/*           storage. */

/*           Zero the upper triangle of the work array. */

	    i__1 = nb;
	    for (j = 1; j <= i__1; ++j) {
		i__2 = j - 1;
		for (i__ = 1; i__ <= i__2; ++i__) {
		    i__3 = i__ + j * 33 - 34;
		    work[i__3].r = 0.f, work[i__3].i = 0.f;
/* L10: */
		}
/* L20: */
	    }

/*           Process the band matrix one diagonal block at a time. */

	    i__1 = *n;
	    i__2 = nb;
	    for (i__ = 1; i__2 < 0 ? i__ >= i__1 : i__ <= i__1; i__ += i__2) {
/* Computing MIN */
		i__3 = nb, i__4 = *n - i__ + 1;
		ib = f2cmin(i__3,i__4);

/*              Factorize the diagonal block */

		i__3 = *ldab - 1;
		cpotf2_(uplo, &ib, &ab[*kd + 1 + i__ * ab_dim1], &i__3, &ii);
		if (ii != 0) {
		    *info = i__ + ii - 1;
		    goto L150;
		}
		if (i__ + ib <= *n) {

/*                 Update the relevant part of the trailing submatrix. */
/*                 If A11 denotes the diagonal block which has just been */
/*                 factorized, then we need to update the remaining */
/*                 blocks in the diagram: */

/*                    A11   A12   A13 */
/*                          A22   A23 */
/*                                A33 */

/*                 The numbers of rows and columns in the partitioning */
/*                 are IB, I2, I3 respectively. The blocks A12, A22 and */
/*                 A23 are empty if IB = KD. The upper triangle of A13 */
/*                 lies outside the band. */

/* Computing MIN */
		    i__3 = *kd - ib, i__4 = *n - i__ - ib + 1;
		    i2 = f2cmin(i__3,i__4);
/* Computing MIN */
		    i__3 = ib, i__4 = *n - i__ - *kd + 1;
		    i3 = f2cmin(i__3,i__4);

		    if (i2 > 0) {

/*                    Update A12 */

			i__3 = *ldab - 1;
			i__4 = *ldab - 1;
			ctrsm_("Left", "Upper", "Conjugate transpose", "Non-"
				"unit", &ib, &i2, &c_b1, &ab[*kd + 1 + i__ * 
				ab_dim1], &i__3, &ab[*kd + 1 - ib + (i__ + ib)
				 * ab_dim1], &i__4);

/*                    Update A22 */

			i__3 = *ldab - 1;
			i__4 = *ldab - 1;
			cherk_("Upper", "Conjugate transpose", &i2, &ib, &
				c_b21, &ab[*kd + 1 - ib + (i__ + ib) * 
				ab_dim1], &i__3, &c_b22, &ab[*kd + 1 + (i__ + 
				ib) * ab_dim1], &i__4);
		    }

		    if (i3 > 0) {

/*                    Copy the lower triangle of A13 into the work array. */

			i__3 = i3;
			for (jj = 1; jj <= i__3; ++jj) {
			    i__4 = ib;
			    for (ii = jj; ii <= i__4; ++ii) {
				i__5 = ii + jj * 33 - 34;
				i__6 = ii - jj + 1 + (jj + i__ + *kd - 1) * 
					ab_dim1;
				work[i__5].r = ab[i__6].r, work[i__5].i = ab[
					i__6].i;
/* L30: */
			    }
/* L40: */
			}

/*                    Update A13 (in the work array). */

			i__3 = *ldab - 1;
			ctrsm_("Left", "Upper", "Conjugate transpose", "Non-"
				"unit", &ib, &i3, &c_b1, &ab[*kd + 1 + i__ * 
				ab_dim1], &i__3, work, &c__33);

/*                    Update A23 */

			if (i2 > 0) {
			    q__1.r = -1.f, q__1.i = 0.f;
			    i__3 = *ldab - 1;
			    i__4 = *ldab - 1;
			    cgemm_("Conjugate transpose", "No transpose", &i2,
				     &i3, &ib, &q__1, &ab[*kd + 1 - ib + (i__ 
				    + ib) * ab_dim1], &i__3, work, &c__33, &
				    c_b1, &ab[ib + 1 + (i__ + *kd) * ab_dim1],
				     &i__4);
			}

/*                    Update A33 */

			i__3 = *ldab - 1;
			cherk_("Upper", "Conjugate transpose", &i3, &ib, &
				c_b21, work, &c__33, &c_b22, &ab[*kd + 1 + (
				i__ + *kd) * ab_dim1], &i__3);

/*                    Copy the lower triangle of A13 back into place. */

			i__3 = i3;
			for (jj = 1; jj <= i__3; ++jj) {
			    i__4 = ib;
			    for (ii = jj; ii <= i__4; ++ii) {
				i__5 = ii - jj + 1 + (jj + i__ + *kd - 1) * 
					ab_dim1;
				i__6 = ii + jj * 33 - 34;
				ab[i__5].r = work[i__6].r, ab[i__5].i = work[
					i__6].i;
/* L50: */
			    }
/* L60: */
			}
		    }
		}
/* L70: */
	    }
	} else {

/*           Compute the Cholesky factorization of a Hermitian band */
/*           matrix, given the lower triangle of the matrix in band */
/*           storage. */

/*           Zero the lower triangle of the work array. */

	    i__2 = nb;
	    for (j = 1; j <= i__2; ++j) {
		i__1 = nb;
		for (i__ = j + 1; i__ <= i__1; ++i__) {
		    i__3 = i__ + j * 33 - 34;
		    work[i__3].r = 0.f, work[i__3].i = 0.f;
/* L80: */
		}
/* L90: */
	    }

/*           Process the band matrix one diagonal block at a time. */

	    i__2 = *n;
	    i__1 = nb;
	    for (i__ = 1; i__1 < 0 ? i__ >= i__2 : i__ <= i__2; i__ += i__1) {
/* Computing MIN */
		i__3 = nb, i__4 = *n - i__ + 1;
		ib = f2cmin(i__3,i__4);

/*              Factorize the diagonal block */

		i__3 = *ldab - 1;
		cpotf2_(uplo, &ib, &ab[i__ * ab_dim1 + 1], &i__3, &ii);
		if (ii != 0) {
		    *info = i__ + ii - 1;
		    goto L150;
		}
		if (i__ + ib <= *n) {

/*                 Update the relevant part of the trailing submatrix. */
/*                 If A11 denotes the diagonal block which has just been */
/*                 factorized, then we need to update the remaining */
/*                 blocks in the diagram: */

/*                    A11 */
/*                    A21   A22 */
/*                    A31   A32   A33 */

/*                 The numbers of rows and columns in the partitioning */
/*                 are IB, I2, I3 respectively. The blocks A21, A22 and */
/*                 A32 are empty if IB = KD. The lower triangle of A31 */
/*                 lies outside the band. */

/* Computing MIN */
		    i__3 = *kd - ib, i__4 = *n - i__ - ib + 1;
		    i2 = f2cmin(i__3,i__4);
/* Computing MIN */
		    i__3 = ib, i__4 = *n - i__ - *kd + 1;
		    i3 = f2cmin(i__3,i__4);

		    if (i2 > 0) {

/*                    Update A21 */

			i__3 = *ldab - 1;
			i__4 = *ldab - 1;
			ctrsm_("Right", "Lower", "Conjugate transpose", "Non"
				"-unit", &i2, &ib, &c_b1, &ab[i__ * ab_dim1 + 
				1], &i__3, &ab[ib + 1 + i__ * ab_dim1], &i__4);

/*                    Update A22 */

			i__3 = *ldab - 1;
			i__4 = *ldab - 1;
			cherk_("Lower", "No transpose", &i2, &ib, &c_b21, &ab[
				ib + 1 + i__ * ab_dim1], &i__3, &c_b22, &ab[(
				i__ + ib) * ab_dim1 + 1], &i__4);
		    }

		    if (i3 > 0) {

/*                    Copy the upper triangle of A31 into the work array. */

			i__3 = ib;
			for (jj = 1; jj <= i__3; ++jj) {
			    i__4 = f2cmin(jj,i3);
			    for (ii = 1; ii <= i__4; ++ii) {
				i__5 = ii + jj * 33 - 34;
				i__6 = *kd + 1 - jj + ii + (jj + i__ - 1) * 
					ab_dim1;
				work[i__5].r = ab[i__6].r, work[i__5].i = ab[
					i__6].i;
/* L100: */
			    }
/* L110: */
			}

/*                    Update A31 (in the work array). */

			i__3 = *ldab - 1;
			ctrsm_("Right", "Lower", "Conjugate transpose", "Non"
				"-unit", &i3, &ib, &c_b1, &ab[i__ * ab_dim1 + 
				1], &i__3, work, &c__33);

/*                    Update A32 */

			if (i2 > 0) {
			    q__1.r = -1.f, q__1.i = 0.f;
			    i__3 = *ldab - 1;
			    i__4 = *ldab - 1;
			    cgemm_("No transpose", "Conjugate transpose", &i3,
				     &i2, &ib, &q__1, work, &c__33, &ab[ib + 
				    1 + i__ * ab_dim1], &i__3, &c_b1, &ab[*kd 
				    + 1 - ib + (i__ + ib) * ab_dim1], &i__4);
			}

/*                    Update A33 */

			i__3 = *ldab - 1;
			cherk_("Lower", "No transpose", &i3, &ib, &c_b21, 
				work, &c__33, &c_b22, &ab[(i__ + *kd) * 
				ab_dim1 + 1], &i__3);

/*                    Copy the upper triangle of A31 back into place. */

			i__3 = ib;
			for (jj = 1; jj <= i__3; ++jj) {
			    i__4 = f2cmin(jj,i3);
			    for (ii = 1; ii <= i__4; ++ii) {
				i__5 = *kd + 1 - jj + ii + (jj + i__ - 1) * 
					ab_dim1;
				i__6 = ii + jj * 33 - 34;
				ab[i__5].r = work[i__6].r, ab[i__5].i = work[
					i__6].i;
/* L120: */
			    }
/* L130: */
			}
		    }
		}
/* L140: */
	    }
	}
    }
    return;

L150:
    return;

/*     End of CPBTRF */

} /* cpbtrf_ */