946 lines
24 KiB
C
946 lines
24 KiB
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)
|
|
*/
|
|
|
|
|
|
|
|
|
|
/* > \brief \b SLASQ3 checks for deflation, computes a shift and calls dqds. Used by sbdsqr. */
|
|
|
|
/* =========== DOCUMENTATION =========== */
|
|
|
|
/* Online html documentation available at */
|
|
/* http://www.netlib.org/lapack/explore-html/ */
|
|
|
|
/* > \htmlonly */
|
|
/* > Download SLASQ3 + dependencies */
|
|
/* > <a href="http://www.netlib.org/cgi-bin/netlibfiles.tgz?format=tgz&filename=/lapack/lapack_routine/slasq3.
|
|
f"> */
|
|
/* > [TGZ]</a> */
|
|
/* > <a href="http://www.netlib.org/cgi-bin/netlibfiles.zip?format=zip&filename=/lapack/lapack_routine/slasq3.
|
|
f"> */
|
|
/* > [ZIP]</a> */
|
|
/* > <a href="http://www.netlib.org/cgi-bin/netlibfiles.txt?format=txt&filename=/lapack/lapack_routine/slasq3.
|
|
f"> */
|
|
/* > [TXT]</a> */
|
|
/* > \endhtmlonly */
|
|
|
|
/* Definition: */
|
|
/* =========== */
|
|
|
|
/* SUBROUTINE SLASQ3( I0, N0, Z, PP, DMIN, SIGMA, DESIG, QMAX, NFAIL, */
|
|
/* ITER, NDIV, IEEE, TTYPE, DMIN1, DMIN2, DN, DN1, */
|
|
/* DN2, G, TAU ) */
|
|
|
|
/* LOGICAL IEEE */
|
|
/* INTEGER I0, ITER, N0, NDIV, NFAIL, PP */
|
|
/* REAL DESIG, DMIN, DMIN1, DMIN2, DN, DN1, DN2, G, */
|
|
/* $ QMAX, SIGMA, TAU */
|
|
/* REAL Z( * ) */
|
|
|
|
|
|
/* > \par Purpose: */
|
|
/* ============= */
|
|
/* > */
|
|
/* > \verbatim */
|
|
/* > */
|
|
/* > SLASQ3 checks for deflation, computes a shift (TAU) and calls dqds. */
|
|
/* > In case of failure it changes shifts, and tries again until output */
|
|
/* > is positive. */
|
|
/* > \endverbatim */
|
|
|
|
/* Arguments: */
|
|
/* ========== */
|
|
|
|
/* > \param[in] I0 */
|
|
/* > \verbatim */
|
|
/* > I0 is INTEGER */
|
|
/* > First index. */
|
|
/* > \endverbatim */
|
|
/* > */
|
|
/* > \param[in,out] N0 */
|
|
/* > \verbatim */
|
|
/* > N0 is INTEGER */
|
|
/* > Last index. */
|
|
/* > \endverbatim */
|
|
/* > */
|
|
/* > \param[in,out] Z */
|
|
/* > \verbatim */
|
|
/* > Z is REAL array, dimension ( 4*N0 ) */
|
|
/* > Z holds the qd array. */
|
|
/* > \endverbatim */
|
|
/* > */
|
|
/* > \param[in,out] PP */
|
|
/* > \verbatim */
|
|
/* > PP is INTEGER */
|
|
/* > PP=0 for ping, PP=1 for pong. */
|
|
/* > PP=2 indicates that flipping was applied to the Z array */
|
|
/* > and that the initial tests for deflation should not be */
|
|
/* > performed. */
|
|
/* > \endverbatim */
|
|
/* > */
|
|
/* > \param[out] DMIN */
|
|
/* > \verbatim */
|
|
/* > DMIN is REAL */
|
|
/* > Minimum value of d. */
|
|
/* > \endverbatim */
|
|
/* > */
|
|
/* > \param[out] SIGMA */
|
|
/* > \verbatim */
|
|
/* > SIGMA is REAL */
|
|
/* > Sum of shifts used in current segment. */
|
|
/* > \endverbatim */
|
|
/* > */
|
|
/* > \param[in,out] DESIG */
|
|
/* > \verbatim */
|
|
/* > DESIG is REAL */
|
|
/* > Lower order part of SIGMA */
|
|
/* > \endverbatim */
|
|
/* > */
|
|
/* > \param[in] QMAX */
|
|
/* > \verbatim */
|
|
/* > QMAX is REAL */
|
|
/* > Maximum value of q. */
|
|
/* > \endverbatim */
|
|
/* > */
|
|
/* > \param[in,out] NFAIL */
|
|
/* > \verbatim */
|
|
/* > NFAIL is INTEGER */
|
|
/* > Increment NFAIL by 1 each time the shift was too big. */
|
|
/* > \endverbatim */
|
|
/* > */
|
|
/* > \param[in,out] ITER */
|
|
/* > \verbatim */
|
|
/* > ITER is INTEGER */
|
|
/* > Increment ITER by 1 for each iteration. */
|
|
/* > \endverbatim */
|
|
/* > */
|
|
/* > \param[in,out] NDIV */
|
|
/* > \verbatim */
|
|
/* > NDIV is INTEGER */
|
|
/* > Increment NDIV by 1 for each division. */
|
|
/* > \endverbatim */
|
|
/* > */
|
|
/* > \param[in] IEEE */
|
|
/* > \verbatim */
|
|
/* > IEEE is LOGICAL */
|
|
/* > Flag for IEEE or non IEEE arithmetic (passed to SLASQ5). */
|
|
/* > \endverbatim */
|
|
/* > */
|
|
/* > \param[in,out] TTYPE */
|
|
/* > \verbatim */
|
|
/* > TTYPE is INTEGER */
|
|
/* > Shift type. */
|
|
/* > \endverbatim */
|
|
/* > */
|
|
/* > \param[in,out] DMIN1 */
|
|
/* > \verbatim */
|
|
/* > DMIN1 is REAL */
|
|
/* > \endverbatim */
|
|
/* > */
|
|
/* > \param[in,out] DMIN2 */
|
|
/* > \verbatim */
|
|
/* > DMIN2 is REAL */
|
|
/* > \endverbatim */
|
|
/* > */
|
|
/* > \param[in,out] DN */
|
|
/* > \verbatim */
|
|
/* > DN is REAL */
|
|
/* > \endverbatim */
|
|
/* > */
|
|
/* > \param[in,out] DN1 */
|
|
/* > \verbatim */
|
|
/* > DN1 is REAL */
|
|
/* > \endverbatim */
|
|
/* > */
|
|
/* > \param[in,out] DN2 */
|
|
/* > \verbatim */
|
|
/* > DN2 is REAL */
|
|
/* > \endverbatim */
|
|
/* > */
|
|
/* > \param[in,out] G */
|
|
/* > \verbatim */
|
|
/* > G is REAL */
|
|
/* > \endverbatim */
|
|
/* > */
|
|
/* > \param[in,out] TAU */
|
|
/* > \verbatim */
|
|
/* > TAU is REAL */
|
|
/* > */
|
|
/* > These are passed as arguments in order to save their values */
|
|
/* > between calls to SLASQ3. */
|
|
/* > \endverbatim */
|
|
|
|
/* Authors: */
|
|
/* ======== */
|
|
|
|
/* > \author Univ. of Tennessee */
|
|
/* > \author Univ. of California Berkeley */
|
|
/* > \author Univ. of Colorado Denver */
|
|
/* > \author NAG Ltd. */
|
|
|
|
/* > \date June 2016 */
|
|
|
|
/* > \ingroup auxOTHERcomputational */
|
|
|
|
/* ===================================================================== */
|
|
/* Subroutine */ void slasq3_(integer *i0, integer *n0, real *z__, integer *pp,
|
|
real *dmin__, real *sigma, real *desig, real *qmax, integer *nfail,
|
|
integer *iter, integer *ndiv, logical *ieee, integer *ttype, real *
|
|
dmin1, real *dmin2, real *dn, real *dn1, real *dn2, real *g, real *
|
|
tau)
|
|
{
|
|
/* System generated locals */
|
|
integer i__1;
|
|
real r__1, r__2;
|
|
|
|
/* Local variables */
|
|
real temp, s, t;
|
|
integer j4;
|
|
extern /* Subroutine */ void slasq4_(integer *, integer *, real *, integer
|
|
*, integer *, real *, real *, real *, real *, real *, real *,
|
|
real *, integer *, real *), slasq5_(integer *, integer *, real *,
|
|
integer *, real *, real *, real *, real *, real *, real *, real *,
|
|
real *, logical *, real *), slasq6_(integer *, integer *, real *,
|
|
integer *, real *, real *, real *, real *, real *, real *);
|
|
integer nn;
|
|
extern real slamch_(char *);
|
|
extern logical sisnan_(real *);
|
|
real eps, tol;
|
|
integer n0in, ipn4;
|
|
real tol2;
|
|
|
|
|
|
/* -- 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..-- */
|
|
/* June 2016 */
|
|
|
|
|
|
/* ===================================================================== */
|
|
|
|
|
|
/* Parameter adjustments */
|
|
--z__;
|
|
|
|
/* Function Body */
|
|
n0in = *n0;
|
|
eps = slamch_("Precision");
|
|
tol = eps * 100.f;
|
|
/* Computing 2nd power */
|
|
r__1 = tol;
|
|
tol2 = r__1 * r__1;
|
|
|
|
/* Check for deflation. */
|
|
|
|
L10:
|
|
|
|
if (*n0 < *i0) {
|
|
return;
|
|
}
|
|
if (*n0 == *i0) {
|
|
goto L20;
|
|
}
|
|
nn = (*n0 << 2) + *pp;
|
|
if (*n0 == *i0 + 1) {
|
|
goto L40;
|
|
}
|
|
|
|
/* Check whether E(N0-1) is negligible, 1 eigenvalue. */
|
|
|
|
if (z__[nn - 5] > tol2 * (*sigma + z__[nn - 3]) && z__[nn - (*pp << 1) -
|
|
4] > tol2 * z__[nn - 7]) {
|
|
goto L30;
|
|
}
|
|
|
|
L20:
|
|
|
|
z__[(*n0 << 2) - 3] = z__[(*n0 << 2) + *pp - 3] + *sigma;
|
|
--(*n0);
|
|
goto L10;
|
|
|
|
/* Check whether E(N0-2) is negligible, 2 eigenvalues. */
|
|
|
|
L30:
|
|
|
|
if (z__[nn - 9] > tol2 * *sigma && z__[nn - (*pp << 1) - 8] > tol2 * z__[
|
|
nn - 11]) {
|
|
goto L50;
|
|
}
|
|
|
|
L40:
|
|
|
|
if (z__[nn - 3] > z__[nn - 7]) {
|
|
s = z__[nn - 3];
|
|
z__[nn - 3] = z__[nn - 7];
|
|
z__[nn - 7] = s;
|
|
}
|
|
t = (z__[nn - 7] - z__[nn - 3] + z__[nn - 5]) * .5f;
|
|
if (z__[nn - 5] > z__[nn - 3] * tol2 && t != 0.f) {
|
|
s = z__[nn - 3] * (z__[nn - 5] / t);
|
|
if (s <= t) {
|
|
s = z__[nn - 3] * (z__[nn - 5] / (t * (sqrt(s / t + 1.f) + 1.f)));
|
|
} else {
|
|
s = z__[nn - 3] * (z__[nn - 5] / (t + sqrt(t) * sqrt(t + s)));
|
|
}
|
|
t = z__[nn - 7] + (s + z__[nn - 5]);
|
|
z__[nn - 3] *= z__[nn - 7] / t;
|
|
z__[nn - 7] = t;
|
|
}
|
|
z__[(*n0 << 2) - 7] = z__[nn - 7] + *sigma;
|
|
z__[(*n0 << 2) - 3] = z__[nn - 3] + *sigma;
|
|
*n0 += -2;
|
|
goto L10;
|
|
|
|
L50:
|
|
if (*pp == 2) {
|
|
*pp = 0;
|
|
}
|
|
|
|
/* Reverse the qd-array, if warranted. */
|
|
|
|
if (*dmin__ <= 0.f || *n0 < n0in) {
|
|
if (z__[(*i0 << 2) + *pp - 3] * 1.5f < z__[(*n0 << 2) + *pp - 3]) {
|
|
ipn4 = *i0 + *n0 << 2;
|
|
i__1 = *i0 + *n0 - 1 << 1;
|
|
for (j4 = *i0 << 2; j4 <= i__1; j4 += 4) {
|
|
temp = z__[j4 - 3];
|
|
z__[j4 - 3] = z__[ipn4 - j4 - 3];
|
|
z__[ipn4 - j4 - 3] = temp;
|
|
temp = z__[j4 - 2];
|
|
z__[j4 - 2] = z__[ipn4 - j4 - 2];
|
|
z__[ipn4 - j4 - 2] = temp;
|
|
temp = z__[j4 - 1];
|
|
z__[j4 - 1] = z__[ipn4 - j4 - 5];
|
|
z__[ipn4 - j4 - 5] = temp;
|
|
temp = z__[j4];
|
|
z__[j4] = z__[ipn4 - j4 - 4];
|
|
z__[ipn4 - j4 - 4] = temp;
|
|
/* L60: */
|
|
}
|
|
if (*n0 - *i0 <= 4) {
|
|
z__[(*n0 << 2) + *pp - 1] = z__[(*i0 << 2) + *pp - 1];
|
|
z__[(*n0 << 2) - *pp] = z__[(*i0 << 2) - *pp];
|
|
}
|
|
/* Computing MIN */
|
|
r__1 = *dmin2, r__2 = z__[(*n0 << 2) + *pp - 1];
|
|
*dmin2 = f2cmin(r__1,r__2);
|
|
/* Computing MIN */
|
|
r__1 = z__[(*n0 << 2) + *pp - 1], r__2 = z__[(*i0 << 2) + *pp - 1]
|
|
, r__1 = f2cmin(r__1,r__2), r__2 = z__[(*i0 << 2) + *pp + 3];
|
|
z__[(*n0 << 2) + *pp - 1] = f2cmin(r__1,r__2);
|
|
/* Computing MIN */
|
|
r__1 = z__[(*n0 << 2) - *pp], r__2 = z__[(*i0 << 2) - *pp], r__1 =
|
|
f2cmin(r__1,r__2), r__2 = z__[(*i0 << 2) - *pp + 4];
|
|
z__[(*n0 << 2) - *pp] = f2cmin(r__1,r__2);
|
|
/* Computing MAX */
|
|
r__1 = *qmax, r__2 = z__[(*i0 << 2) + *pp - 3], r__1 = f2cmax(r__1,
|
|
r__2), r__2 = z__[(*i0 << 2) + *pp + 1];
|
|
*qmax = f2cmax(r__1,r__2);
|
|
*dmin__ = 0.f;
|
|
}
|
|
}
|
|
|
|
/* Choose a shift. */
|
|
|
|
slasq4_(i0, n0, &z__[1], pp, &n0in, dmin__, dmin1, dmin2, dn, dn1, dn2,
|
|
tau, ttype, g);
|
|
|
|
/* Call dqds until DMIN > 0. */
|
|
|
|
L70:
|
|
|
|
slasq5_(i0, n0, &z__[1], pp, tau, sigma, dmin__, dmin1, dmin2, dn, dn1,
|
|
dn2, ieee, &eps);
|
|
|
|
*ndiv += *n0 - *i0 + 2;
|
|
++(*iter);
|
|
|
|
/* Check status. */
|
|
|
|
if (*dmin__ >= 0.f && *dmin1 >= 0.f) {
|
|
|
|
/* Success. */
|
|
|
|
goto L90;
|
|
|
|
} else if (*dmin__ < 0.f && *dmin1 > 0.f && z__[(*n0 - 1 << 2) - *pp] <
|
|
tol * (*sigma + *dn1) && abs(*dn) < tol * *sigma) {
|
|
|
|
/* Convergence hidden by negative DN. */
|
|
|
|
z__[(*n0 - 1 << 2) - *pp + 2] = 0.f;
|
|
*dmin__ = 0.f;
|
|
goto L90;
|
|
} else if (*dmin__ < 0.f) {
|
|
|
|
/* TAU too big. Select new TAU and try again. */
|
|
|
|
++(*nfail);
|
|
if (*ttype < -22) {
|
|
|
|
/* Failed twice. Play it safe. */
|
|
|
|
*tau = 0.f;
|
|
} else if (*dmin1 > 0.f) {
|
|
|
|
/* Late failure. Gives excellent shift. */
|
|
|
|
*tau = (*tau + *dmin__) * (1.f - eps * 2.f);
|
|
*ttype += -11;
|
|
} else {
|
|
|
|
/* Early failure. Divide by 4. */
|
|
|
|
*tau *= .25f;
|
|
*ttype += -12;
|
|
}
|
|
goto L70;
|
|
} else if (sisnan_(dmin__)) {
|
|
|
|
/* NaN. */
|
|
|
|
if (*tau == 0.f) {
|
|
goto L80;
|
|
} else {
|
|
*tau = 0.f;
|
|
goto L70;
|
|
}
|
|
} else {
|
|
|
|
/* Possible underflow. Play it safe. */
|
|
|
|
goto L80;
|
|
}
|
|
|
|
/* Risk of underflow. */
|
|
|
|
L80:
|
|
slasq6_(i0, n0, &z__[1], pp, dmin__, dmin1, dmin2, dn, dn1, dn2);
|
|
*ndiv += *n0 - *i0 + 2;
|
|
++(*iter);
|
|
*tau = 0.f;
|
|
|
|
L90:
|
|
if (*tau < *sigma) {
|
|
*desig += *tau;
|
|
t = *sigma + *desig;
|
|
*desig -= t - *sigma;
|
|
} else {
|
|
t = *sigma + *tau;
|
|
*desig = *sigma - (t - *tau) + *desig;
|
|
}
|
|
*sigma = t;
|
|
|
|
return;
|
|
|
|
/* End of SLASQ3 */
|
|
|
|
} /* slasq3_ */
|
|
|