1013 lines
30 KiB
C
1013 lines
30 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 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 doublecomplex c_b1 = {-1.,0.};
|
|
static doublecomplex c_b2 = {1.,0.};
|
|
static integer c__1 = 1;
|
|
|
|
/* > \brief <b> ZCPOSV computes the solution to system of linear equations A * X = B for PO matrices</b> */
|
|
|
|
/* =========== DOCUMENTATION =========== */
|
|
|
|
/* Online html documentation available at */
|
|
/* http://www.netlib.org/lapack/explore-html/ */
|
|
|
|
/* > \htmlonly */
|
|
/* > Download ZCPOSV + dependencies */
|
|
/* > <a href="http://www.netlib.org/cgi-bin/netlibfiles.tgz?format=tgz&filename=/lapack/lapack_routine/zcposv.
|
|
f"> */
|
|
/* > [TGZ]</a> */
|
|
/* > <a href="http://www.netlib.org/cgi-bin/netlibfiles.zip?format=zip&filename=/lapack/lapack_routine/zcposv.
|
|
f"> */
|
|
/* > [ZIP]</a> */
|
|
/* > <a href="http://www.netlib.org/cgi-bin/netlibfiles.txt?format=txt&filename=/lapack/lapack_routine/zcposv.
|
|
f"> */
|
|
/* > [TXT]</a> */
|
|
/* > \endhtmlonly */
|
|
|
|
/* Definition: */
|
|
/* =========== */
|
|
|
|
/* SUBROUTINE ZCPOSV( UPLO, N, NRHS, A, LDA, B, LDB, X, LDX, WORK, */
|
|
/* SWORK, RWORK, ITER, INFO ) */
|
|
|
|
/* CHARACTER UPLO */
|
|
/* INTEGER INFO, ITER, LDA, LDB, LDX, N, NRHS */
|
|
/* DOUBLE PRECISION RWORK( * ) */
|
|
/* COMPLEX SWORK( * ) */
|
|
/* COMPLEX*16 A( LDA, * ), B( LDB, * ), WORK( N, * ), */
|
|
/* $ X( LDX, * ) */
|
|
|
|
|
|
/* > \par Purpose: */
|
|
/* ============= */
|
|
/* > */
|
|
/* > \verbatim */
|
|
/* > */
|
|
/* > ZCPOSV computes the solution to a complex system of linear equations */
|
|
/* > A * X = B, */
|
|
/* > where A is an N-by-N Hermitian positive definite matrix and X and B */
|
|
/* > are N-by-NRHS matrices. */
|
|
/* > */
|
|
/* > ZCPOSV first attempts to factorize the matrix in COMPLEX and use this */
|
|
/* > factorization within an iterative refinement procedure to produce a */
|
|
/* > solution with COMPLEX*16 normwise backward error quality (see below). */
|
|
/* > If the approach fails the method switches to a COMPLEX*16 */
|
|
/* > factorization and solve. */
|
|
/* > */
|
|
/* > The iterative refinement is not going to be a winning strategy if */
|
|
/* > the ratio COMPLEX performance over COMPLEX*16 performance is too */
|
|
/* > small. A reasonable strategy should take the number of right-hand */
|
|
/* > sides and the size of the matrix into account. This might be done */
|
|
/* > with a call to ILAENV in the future. Up to now, we always try */
|
|
/* > iterative refinement. */
|
|
/* > */
|
|
/* > The iterative refinement process is stopped if */
|
|
/* > ITER > ITERMAX */
|
|
/* > or for all the RHS we have: */
|
|
/* > RNRM < SQRT(N)*XNRM*ANRM*EPS*BWDMAX */
|
|
/* > where */
|
|
/* > o ITER is the number of the current iteration in the iterative */
|
|
/* > refinement process */
|
|
/* > o RNRM is the infinity-norm of the residual */
|
|
/* > o XNRM is the infinity-norm of the solution */
|
|
/* > o ANRM is the infinity-operator-norm of the matrix A */
|
|
/* > o EPS is the machine epsilon returned by DLAMCH('Epsilon') */
|
|
/* > The value ITERMAX and BWDMAX are fixed to 30 and 1.0D+00 */
|
|
/* > respectively. */
|
|
/* > \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 number of linear equations, i.e., the order 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 matrix B. NRHS >= 0. */
|
|
/* > \endverbatim */
|
|
/* > */
|
|
/* > \param[in,out] A */
|
|
/* > \verbatim */
|
|
/* > A is COMPLEX*16 array, */
|
|
/* > dimension (LDA,N) */
|
|
/* > On entry, the Hermitian matrix A. If UPLO = 'U', the leading */
|
|
/* > N-by-N upper triangular part of A contains the upper */
|
|
/* > triangular part of the matrix A, and the strictly lower */
|
|
/* > triangular part of A is not referenced. If UPLO = 'L', the */
|
|
/* > leading N-by-N lower triangular part of A contains the lower */
|
|
/* > triangular part of the matrix A, and the strictly upper */
|
|
/* > triangular part of A is not referenced. */
|
|
/* > */
|
|
/* > Note that the imaginary parts of the diagonal */
|
|
/* > elements need not be set and are assumed to be zero. */
|
|
/* > */
|
|
/* > On exit, if iterative refinement has been successfully used */
|
|
/* > (INFO = 0 and ITER >= 0, see description below), then A is */
|
|
/* > unchanged, if double precision factorization has been used */
|
|
/* > (INFO = 0 and ITER < 0, see description below), then the */
|
|
/* > array A contains the factor U or L from the Cholesky */
|
|
/* > factorization A = U**H*U or A = L*L**H. */
|
|
/* > \endverbatim */
|
|
/* > */
|
|
/* > \param[in] LDA */
|
|
/* > \verbatim */
|
|
/* > LDA is INTEGER */
|
|
/* > The leading dimension of the array A. LDA >= f2cmax(1,N). */
|
|
/* > \endverbatim */
|
|
/* > */
|
|
/* > \param[in] B */
|
|
/* > \verbatim */
|
|
/* > B is COMPLEX*16 array, dimension (LDB,NRHS) */
|
|
/* > The N-by-NRHS right hand side matrix B. */
|
|
/* > \endverbatim */
|
|
/* > */
|
|
/* > \param[in] LDB */
|
|
/* > \verbatim */
|
|
/* > LDB is INTEGER */
|
|
/* > The leading dimension of the array B. LDB >= f2cmax(1,N). */
|
|
/* > \endverbatim */
|
|
/* > */
|
|
/* > \param[out] X */
|
|
/* > \verbatim */
|
|
/* > X is COMPLEX*16 array, dimension (LDX,NRHS) */
|
|
/* > If INFO = 0, the N-by-NRHS solution matrix X. */
|
|
/* > \endverbatim */
|
|
/* > */
|
|
/* > \param[in] LDX */
|
|
/* > \verbatim */
|
|
/* > LDX is INTEGER */
|
|
/* > The leading dimension of the array X. LDX >= f2cmax(1,N). */
|
|
/* > \endverbatim */
|
|
/* > */
|
|
/* > \param[out] WORK */
|
|
/* > \verbatim */
|
|
/* > WORK is COMPLEX*16 array, dimension (N,NRHS) */
|
|
/* > This array is used to hold the residual vectors. */
|
|
/* > \endverbatim */
|
|
/* > */
|
|
/* > \param[out] SWORK */
|
|
/* > \verbatim */
|
|
/* > SWORK is COMPLEX array, dimension (N*(N+NRHS)) */
|
|
/* > This array is used to use the single precision matrix and the */
|
|
/* > right-hand sides or solutions in single precision. */
|
|
/* > \endverbatim */
|
|
/* > */
|
|
/* > \param[out] RWORK */
|
|
/* > \verbatim */
|
|
/* > RWORK is DOUBLE PRECISION array, dimension (N) */
|
|
/* > \endverbatim */
|
|
/* > */
|
|
/* > \param[out] ITER */
|
|
/* > \verbatim */
|
|
/* > ITER is INTEGER */
|
|
/* > < 0: iterative refinement has failed, COMPLEX*16 */
|
|
/* > factorization has been performed */
|
|
/* > -1 : the routine fell back to full precision for */
|
|
/* > implementation- or machine-specific reasons */
|
|
/* > -2 : narrowing the precision induced an overflow, */
|
|
/* > the routine fell back to full precision */
|
|
/* > -3 : failure of CPOTRF */
|
|
/* > -31: stop the iterative refinement after the 30th */
|
|
/* > iterations */
|
|
/* > > 0: iterative refinement has been successfully used. */
|
|
/* > Returns the number of iterations */
|
|
/* > \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 of */
|
|
/* > (COMPLEX*16) A is not positive definite, so the */
|
|
/* > factorization could not be completed, and the solution */
|
|
/* > has not been computed. */
|
|
/* > \endverbatim */
|
|
|
|
/* Authors: */
|
|
/* ======== */
|
|
|
|
/* > \author Univ. of Tennessee */
|
|
/* > \author Univ. of California Berkeley */
|
|
/* > \author Univ. of Colorado Denver */
|
|
/* > \author NAG Ltd. */
|
|
|
|
/* > \date June 2016 */
|
|
|
|
/* > \ingroup complex16POsolve */
|
|
|
|
/* ===================================================================== */
|
|
/* Subroutine */ void zcposv_(char *uplo, integer *n, integer *nrhs,
|
|
doublecomplex *a, integer *lda, doublecomplex *b, integer *ldb,
|
|
doublecomplex *x, integer *ldx, doublecomplex *work, complex *swork,
|
|
doublereal *rwork, integer *iter, integer *info)
|
|
{
|
|
/* System generated locals */
|
|
integer a_dim1, a_offset, b_dim1, b_offset, work_dim1, work_offset,
|
|
x_dim1, x_offset, i__1, i__2;
|
|
doublereal d__1, d__2;
|
|
|
|
/* Local variables */
|
|
doublereal anrm;
|
|
integer ptsa;
|
|
doublereal rnrm, xnrm;
|
|
integer ptsx, i__;
|
|
extern logical lsame_(char *, char *);
|
|
integer iiter;
|
|
extern /* Subroutine */ void zhemm_(char *, char *, integer *, integer *,
|
|
doublecomplex *, doublecomplex *, integer *, doublecomplex *,
|
|
integer *, doublecomplex *, doublecomplex *, integer *), zaxpy_(integer *, doublecomplex *, doublecomplex *,
|
|
integer *, doublecomplex *, integer *), zlag2c_(integer *,
|
|
integer *, doublecomplex *, integer *, complex *, integer *,
|
|
integer *), clag2z_(integer *, integer *, complex *, integer *,
|
|
doublecomplex *, integer *, integer *), zlat2c_(char *, integer *,
|
|
doublecomplex *, integer *, complex *, integer *, integer *);
|
|
extern doublereal dlamch_(char *);
|
|
extern /* Subroutine */ int xerbla_(char *, integer *, ftnlen);
|
|
extern doublereal zlanhe_(char *, char *, integer *, doublecomplex *,
|
|
integer *, doublereal *);
|
|
extern integer izamax_(integer *, doublecomplex *, integer *);
|
|
extern /* Subroutine */ int cpotrf_(char *, integer *, complex *, integer
|
|
*, integer *);
|
|
extern void zlacpy_(char *, integer *, integer *,
|
|
doublecomplex *, integer *, doublecomplex *, integer *),
|
|
cpotrs_(char *, integer *, integer *, complex *, integer *,
|
|
complex *, integer *, integer *);
|
|
extern int zpotrf_(char *, integer
|
|
*, doublecomplex *, integer *, integer *);
|
|
extern void zpotrs_(char *,
|
|
integer *, integer *, doublecomplex *, integer *, doublecomplex *
|
|
, integer *, integer *);
|
|
doublereal cte, eps;
|
|
|
|
|
|
/* -- LAPACK driver routine (version 3.8.0) -- */
|
|
/* -- 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 */
|
|
work_dim1 = *n;
|
|
work_offset = 1 + work_dim1 * 1;
|
|
work -= work_offset;
|
|
a_dim1 = *lda;
|
|
a_offset = 1 + a_dim1 * 1;
|
|
a -= a_offset;
|
|
b_dim1 = *ldb;
|
|
b_offset = 1 + b_dim1 * 1;
|
|
b -= b_offset;
|
|
x_dim1 = *ldx;
|
|
x_offset = 1 + x_dim1 * 1;
|
|
x -= x_offset;
|
|
--swork;
|
|
--rwork;
|
|
|
|
/* Function Body */
|
|
*info = 0;
|
|
*iter = 0;
|
|
|
|
/* Test the input parameters. */
|
|
|
|
if (! lsame_(uplo, "U") && ! lsame_(uplo, "L")) {
|
|
*info = -1;
|
|
} else if (*n < 0) {
|
|
*info = -2;
|
|
} else if (*nrhs < 0) {
|
|
*info = -3;
|
|
} else if (*lda < f2cmax(1,*n)) {
|
|
*info = -5;
|
|
} else if (*ldb < f2cmax(1,*n)) {
|
|
*info = -7;
|
|
} else if (*ldx < f2cmax(1,*n)) {
|
|
*info = -9;
|
|
}
|
|
if (*info != 0) {
|
|
i__1 = -(*info);
|
|
xerbla_("ZCPOSV", &i__1, (ftnlen)6);
|
|
return;
|
|
}
|
|
|
|
/* Quick return if (N.EQ.0). */
|
|
|
|
if (*n == 0) {
|
|
return;
|
|
}
|
|
|
|
/* Skip single precision iterative refinement if a priori slower */
|
|
/* than double precision factorization. */
|
|
|
|
if (FALSE_) {
|
|
*iter = -1;
|
|
goto L40;
|
|
}
|
|
|
|
/* Compute some constants. */
|
|
|
|
anrm = zlanhe_("I", uplo, n, &a[a_offset], lda, &rwork[1]);
|
|
eps = dlamch_("Epsilon");
|
|
cte = anrm * eps * sqrt((doublereal) (*n)) * 1.;
|
|
|
|
/* Set the indices PTSA, PTSX for referencing SA and SX in SWORK. */
|
|
|
|
ptsa = 1;
|
|
ptsx = ptsa + *n * *n;
|
|
|
|
/* Convert B from double precision to single precision and store the */
|
|
/* result in SX. */
|
|
|
|
zlag2c_(n, nrhs, &b[b_offset], ldb, &swork[ptsx], n, info);
|
|
|
|
if (*info != 0) {
|
|
*iter = -2;
|
|
goto L40;
|
|
}
|
|
|
|
/* Convert A from double precision to single precision and store the */
|
|
/* result in SA. */
|
|
|
|
zlat2c_(uplo, n, &a[a_offset], lda, &swork[ptsa], n, info);
|
|
|
|
if (*info != 0) {
|
|
*iter = -2;
|
|
goto L40;
|
|
}
|
|
|
|
/* Compute the Cholesky factorization of SA. */
|
|
|
|
cpotrf_(uplo, n, &swork[ptsa], n, info);
|
|
|
|
if (*info != 0) {
|
|
*iter = -3;
|
|
goto L40;
|
|
}
|
|
|
|
/* Solve the system SA*SX = SB. */
|
|
|
|
cpotrs_(uplo, n, nrhs, &swork[ptsa], n, &swork[ptsx], n, info);
|
|
|
|
/* Convert SX back to COMPLEX*16 */
|
|
|
|
clag2z_(n, nrhs, &swork[ptsx], n, &x[x_offset], ldx, info);
|
|
|
|
/* Compute R = B - AX (R is WORK). */
|
|
|
|
zlacpy_("All", n, nrhs, &b[b_offset], ldb, &work[work_offset], n);
|
|
|
|
zhemm_("Left", uplo, n, nrhs, &c_b1, &a[a_offset], lda, &x[x_offset], ldx,
|
|
&c_b2, &work[work_offset], n);
|
|
|
|
/* Check whether the NRHS normwise backward errors satisfy the */
|
|
/* stopping criterion. If yes, set ITER=0 and return. */
|
|
|
|
i__1 = *nrhs;
|
|
for (i__ = 1; i__ <= i__1; ++i__) {
|
|
i__2 = izamax_(n, &x[i__ * x_dim1 + 1], &c__1) + i__ * x_dim1;
|
|
xnrm = (d__1 = x[i__2].r, abs(d__1)) + (d__2 = d_imag(&x[izamax_(n, &
|
|
x[i__ * x_dim1 + 1], &c__1) + i__ * x_dim1]), abs(d__2));
|
|
i__2 = izamax_(n, &work[i__ * work_dim1 + 1], &c__1) + i__ *
|
|
work_dim1;
|
|
rnrm = (d__1 = work[i__2].r, abs(d__1)) + (d__2 = d_imag(&work[
|
|
izamax_(n, &work[i__ * work_dim1 + 1], &c__1) + i__ *
|
|
work_dim1]), abs(d__2));
|
|
if (rnrm > xnrm * cte) {
|
|
goto L10;
|
|
}
|
|
}
|
|
|
|
/* If we are here, the NRHS normwise backward errors satisfy the */
|
|
/* stopping criterion. We are good to exit. */
|
|
|
|
*iter = 0;
|
|
return;
|
|
|
|
L10:
|
|
|
|
for (iiter = 1; iiter <= 30; ++iiter) {
|
|
|
|
/* Convert R (in WORK) from double precision to single precision */
|
|
/* and store the result in SX. */
|
|
|
|
zlag2c_(n, nrhs, &work[work_offset], n, &swork[ptsx], n, info);
|
|
|
|
if (*info != 0) {
|
|
*iter = -2;
|
|
goto L40;
|
|
}
|
|
|
|
/* Solve the system SA*SX = SR. */
|
|
|
|
cpotrs_(uplo, n, nrhs, &swork[ptsa], n, &swork[ptsx], n, info);
|
|
|
|
/* Convert SX back to double precision and update the current */
|
|
/* iterate. */
|
|
|
|
clag2z_(n, nrhs, &swork[ptsx], n, &work[work_offset], n, info);
|
|
|
|
i__1 = *nrhs;
|
|
for (i__ = 1; i__ <= i__1; ++i__) {
|
|
zaxpy_(n, &c_b2, &work[i__ * work_dim1 + 1], &c__1, &x[i__ *
|
|
x_dim1 + 1], &c__1);
|
|
}
|
|
|
|
/* Compute R = B - AX (R is WORK). */
|
|
|
|
zlacpy_("All", n, nrhs, &b[b_offset], ldb, &work[work_offset], n);
|
|
|
|
zhemm_("L", uplo, n, nrhs, &c_b1, &a[a_offset], lda, &x[x_offset],
|
|
ldx, &c_b2, &work[work_offset], n);
|
|
|
|
/* Check whether the NRHS normwise backward errors satisfy the */
|
|
/* stopping criterion. If yes, set ITER=IITER>0 and return. */
|
|
|
|
i__1 = *nrhs;
|
|
for (i__ = 1; i__ <= i__1; ++i__) {
|
|
i__2 = izamax_(n, &x[i__ * x_dim1 + 1], &c__1) + i__ * x_dim1;
|
|
xnrm = (d__1 = x[i__2].r, abs(d__1)) + (d__2 = d_imag(&x[izamax_(
|
|
n, &x[i__ * x_dim1 + 1], &c__1) + i__ * x_dim1]), abs(
|
|
d__2));
|
|
i__2 = izamax_(n, &work[i__ * work_dim1 + 1], &c__1) + i__ *
|
|
work_dim1;
|
|
rnrm = (d__1 = work[i__2].r, abs(d__1)) + (d__2 = d_imag(&work[
|
|
izamax_(n, &work[i__ * work_dim1 + 1], &c__1) + i__ *
|
|
work_dim1]), abs(d__2));
|
|
if (rnrm > xnrm * cte) {
|
|
goto L20;
|
|
}
|
|
}
|
|
|
|
/* If we are here, the NRHS normwise backward errors satisfy the */
|
|
/* stopping criterion, we are good to exit. */
|
|
|
|
*iter = iiter;
|
|
|
|
return;
|
|
|
|
L20:
|
|
|
|
/* L30: */
|
|
;
|
|
}
|
|
|
|
/* If we are at this place of the code, this is because we have */
|
|
/* performed ITER=ITERMAX iterations and never satisfied the */
|
|
/* stopping criterion, set up the ITER flag accordingly and follow */
|
|
/* up on double precision routine. */
|
|
|
|
*iter = -31;
|
|
|
|
L40:
|
|
|
|
/* Single-precision iterative refinement failed to converge to a */
|
|
/* satisfactory solution, so we resort to double precision. */
|
|
|
|
zpotrf_(uplo, n, &a[a_offset], lda, info);
|
|
|
|
if (*info != 0) {
|
|
return;
|
|
}
|
|
|
|
zlacpy_("All", n, nrhs, &b[b_offset], ldb, &x[x_offset], ldx);
|
|
zpotrs_(uplo, n, nrhs, &a[a_offset], lda, &x[x_offset], ldx, info);
|
|
|
|
return;
|
|
|
|
/* End of ZCPOSV. */
|
|
|
|
} /* zcposv_ */
|
|
|