clapll.c
Go to the documentation of this file.
00001 /* clapll.f -- translated by f2c (version 20061008).
00002    You must link the resulting object file with libf2c:
00003         on Microsoft Windows system, link with libf2c.lib;
00004         on Linux or Unix systems, link with .../path/to/libf2c.a -lm
00005         or, if you install libf2c.a in a standard place, with -lf2c -lm
00006         -- in that order, at the end of the command line, as in
00007                 cc *.o -lf2c -lm
00008         Source for libf2c is in /netlib/f2c/libf2c.zip, e.g.,
00009 
00010                 http://www.netlib.org/f2c/libf2c.zip
00011 */
00012 
00013 #include "f2c.h"
00014 #include "blaswrap.h"
00015 
00016 /* Subroutine */ int clapll_(integer *n, complex *x, integer *incx, complex *
00017         y, integer *incy, real *ssmin)
00018 {
00019     /* System generated locals */
00020     integer i__1;
00021     real r__1, r__2, r__3;
00022     complex q__1, q__2, q__3, q__4;
00023 
00024     /* Builtin functions */
00025     void r_cnjg(complex *, complex *);
00026     double c_abs(complex *);
00027 
00028     /* Local variables */
00029     complex c__, a11, a12, a22, tau;
00030     extern /* Subroutine */ int slas2_(real *, real *, real *, real *, real *)
00031             ;
00032     extern /* Complex */ VOID cdotc_(complex *, integer *, complex *, integer 
00033             *, complex *, integer *);
00034     extern /* Subroutine */ int caxpy_(integer *, complex *, complex *, 
00035             integer *, complex *, integer *);
00036     real ssmax;
00037     extern /* Subroutine */ int clarfg_(integer *, complex *, complex *, 
00038             integer *, complex *);
00039 
00040 
00041 /*  -- LAPACK auxiliary routine (version 3.2) -- */
00042 /*     Univ. of Tennessee, Univ. of California Berkeley and NAG Ltd.. */
00043 /*     November 2006 */
00044 
00045 /*     .. Scalar Arguments .. */
00046 /*     .. */
00047 /*     .. Array Arguments .. */
00048 /*     .. */
00049 
00050 /*  Purpose */
00051 /*  ======= */
00052 
00053 /*  Given two column vectors X and Y, let */
00054 
00055 /*                       A = ( X Y ). */
00056 
00057 /*  The subroutine first computes the QR factorization of A = Q*R, */
00058 /*  and then computes the SVD of the 2-by-2 upper triangular matrix R. */
00059 /*  The smaller singular value of R is returned in SSMIN, which is used */
00060 /*  as the measurement of the linear dependency of the vectors X and Y. */
00061 
00062 /*  Arguments */
00063 /*  ========= */
00064 
00065 /*  N       (input) INTEGER */
00066 /*          The length of the vectors X and Y. */
00067 
00068 /*  X       (input/output) COMPLEX array, dimension (1+(N-1)*INCX) */
00069 /*          On entry, X contains the N-vector X. */
00070 /*          On exit, X is overwritten. */
00071 
00072 /*  INCX    (input) INTEGER */
00073 /*          The increment between successive elements of X. INCX > 0. */
00074 
00075 /*  Y       (input/output) COMPLEX array, dimension (1+(N-1)*INCY) */
00076 /*          On entry, Y contains the N-vector Y. */
00077 /*          On exit, Y is overwritten. */
00078 
00079 /*  INCY    (input) INTEGER */
00080 /*          The increment between successive elements of Y. INCY > 0. */
00081 
00082 /*  SSMIN   (output) REAL */
00083 /*          The smallest singular value of the N-by-2 matrix A = ( X Y ). */
00084 
00085 /*  ===================================================================== */
00086 
00087 /*     .. Parameters .. */
00088 /*     .. */
00089 /*     .. Local Scalars .. */
00090 /*     .. */
00091 /*     .. Intrinsic Functions .. */
00092 /*     .. */
00093 /*     .. External Functions .. */
00094 /*     .. */
00095 /*     .. External Subroutines .. */
00096 /*     .. */
00097 /*     .. Executable Statements .. */
00098 
00099 /*     Quick return if possible */
00100 
00101     /* Parameter adjustments */
00102     --y;
00103     --x;
00104 
00105     /* Function Body */
00106     if (*n <= 1) {
00107         *ssmin = 0.f;
00108         return 0;
00109     }
00110 
00111 /*     Compute the QR factorization of the N-by-2 matrix ( X Y ) */
00112 
00113     clarfg_(n, &x[1], &x[*incx + 1], incx, &tau);
00114     a11.r = x[1].r, a11.i = x[1].i;
00115     x[1].r = 1.f, x[1].i = 0.f;
00116 
00117     r_cnjg(&q__3, &tau);
00118     q__2.r = -q__3.r, q__2.i = -q__3.i;
00119     cdotc_(&q__4, n, &x[1], incx, &y[1], incy);
00120     q__1.r = q__2.r * q__4.r - q__2.i * q__4.i, q__1.i = q__2.r * q__4.i + 
00121             q__2.i * q__4.r;
00122     c__.r = q__1.r, c__.i = q__1.i;
00123     caxpy_(n, &c__, &x[1], incx, &y[1], incy);
00124 
00125     i__1 = *n - 1;
00126     clarfg_(&i__1, &y[*incy + 1], &y[(*incy << 1) + 1], incy, &tau);
00127 
00128     a12.r = y[1].r, a12.i = y[1].i;
00129     i__1 = *incy + 1;
00130     a22.r = y[i__1].r, a22.i = y[i__1].i;
00131 
00132 /*     Compute the SVD of 2-by-2 Upper triangular matrix. */
00133 
00134     r__1 = c_abs(&a11);
00135     r__2 = c_abs(&a12);
00136     r__3 = c_abs(&a22);
00137     slas2_(&r__1, &r__2, &r__3, ssmin, &ssmax);
00138 
00139     return 0;
00140 
00141 /*     End of CLAPLL */
00142 
00143 } /* clapll_ */


swiftnav
Author(s):
autogenerated on Sat Jun 8 2019 18:55:31