File: dgsvj1.c

package info (click to toggle)
ruby-lapack 1.8.2-1
  • links: PTS, VCS
  • area: main
  • in suites: bookworm, sid, trixie
  • size: 28,572 kB
  • sloc: ansic: 191,612; ruby: 3,937; makefile: 6
file content (186 lines) | stat: -rw-r--r-- 13,327 bytes parent folder | download | duplicates (3)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
#include "rb_lapack.h"

extern VOID dgsvj1_(char* jobv, integer* m, integer* n, integer* n1, doublereal* a, integer* lda, doublereal* d, doublereal* sva, integer* mv, doublereal* v, integer* ldv, doublereal* eps, doublereal* sfmin, doublereal* tol, integer* nsweep, doublereal* work, integer* lwork, integer* info);


static VALUE
rblapack_dgsvj1(int argc, VALUE *argv, VALUE self){
  VALUE rblapack_jobv;
  char jobv; 
  VALUE rblapack_m;
  integer m; 
  VALUE rblapack_n1;
  integer n1; 
  VALUE rblapack_a;
  doublereal *a; 
  VALUE rblapack_d;
  doublereal *d; 
  VALUE rblapack_sva;
  doublereal *sva; 
  VALUE rblapack_mv;
  integer mv; 
  VALUE rblapack_v;
  doublereal *v; 
  VALUE rblapack_eps;
  doublereal eps; 
  VALUE rblapack_sfmin;
  doublereal sfmin; 
  VALUE rblapack_tol;
  doublereal tol; 
  VALUE rblapack_nsweep;
  integer nsweep; 
  VALUE rblapack_lwork;
  integer lwork; 
  VALUE rblapack_info;
  integer info; 
  VALUE rblapack_a_out__;
  doublereal *a_out__;
  VALUE rblapack_d_out__;
  doublereal *d_out__;
  VALUE rblapack_sva_out__;
  doublereal *sva_out__;
  VALUE rblapack_v_out__;
  doublereal *v_out__;
  doublereal *work;

  integer lda;
  integer n;
  integer ldv;

  VALUE rblapack_options;
  if (argc > 0 && TYPE(argv[argc-1]) == T_HASH) {
    argc--;
    rblapack_options = argv[argc];
    if (rb_hash_aref(rblapack_options, sHelp) == Qtrue) {
      printf("%s\n", "USAGE:\n  info, a, d, sva, v = NumRu::Lapack.dgsvj1( jobv, m, n1, a, d, sva, mv, v, eps, sfmin, tol, nsweep, [:lwork => lwork, :usage => usage, :help => help])\n\n\nFORTRAN MANUAL\n      SUBROUTINE DGSVJ1( JOBV, M, N, N1, A, LDA, D, SVA, MV, V, LDV, EPS, SFMIN, TOL, NSWEEP, WORK, LWORK, INFO )\n\n*  Purpose\n*  =======\n*\n*  DGSVJ1 is called from SGESVJ as a pre-processor and that is its main\n*  purpose. It applies Jacobi rotations in the same way as SGESVJ does, but\n*  it targets only particular pivots and it does not check convergence\n*  (stopping criterion). Few tunning parameters (marked by [TP]) are\n*  available for the implementer.\n*\n*  Further Details\n*  ~~~~~~~~~~~~~~~\n*  DGSVJ1 applies few sweeps of Jacobi rotations in the column space of\n*  the input M-by-N matrix A. The pivot pairs are taken from the (1,2)\n*  off-diagonal block in the corresponding N-by-N Gram matrix A^T * A. The\n*  block-entries (tiles) of the (1,2) off-diagonal block are marked by the\n*  [x]'s in the following scheme:\n*\n*     | *   *   * [x] [x] [x]|\n*     | *   *   * [x] [x] [x]|    Row-cycling in the nblr-by-nblc [x] blocks.\n*     | *   *   * [x] [x] [x]|    Row-cyclic pivoting inside each [x] block.\n*     |[x] [x] [x] *   *   * |\n*     |[x] [x] [x] *   *   * |\n*     |[x] [x] [x] *   *   * |\n*\n*  In terms of the columns of A, the first N1 columns are rotated 'against'\n*  the remaining N-N1 columns, trying to increase the angle between the\n*  corresponding subspaces. The off-diagonal block is N1-by(N-N1) and it is\n*  tiled using quadratic tiles of side KBL. Here, KBL is a tunning parmeter.\n*  The number of sweeps is given in NSWEEP and the orthogonality threshold\n*  is given in TOL.\n*\n*  Contributors\n*  ~~~~~~~~~~~~\n*  Zlatko Drmac (Zagreb, Croatia) and Kresimir Veselic (Hagen, Germany)\n*\n\n*  Arguments\n*  =========\n*\n*  JOBV    (input) CHARACTER*1\n*          Specifies whether the output from this procedure is used\n*          to compute the matrix V:\n*          = 'V': the product of the Jacobi rotations is accumulated\n*                 by postmulyiplying the N-by-N array V.\n*                (See the description of V.)\n*          = 'A': the product of the Jacobi rotations is accumulated\n*                 by postmulyiplying the MV-by-N array V.\n*                (See the descriptions of MV and V.)\n*          = 'N': the Jacobi rotations are not accumulated.\n*\n*  M       (input) INTEGER\n*          The number of rows of the input matrix A.  M >= 0.\n*\n*  N       (input) INTEGER\n*          The number of columns of the input matrix A.\n*          M >= N >= 0.\n*\n*  N1      (input) INTEGER\n*          N1 specifies the 2 x 2 block partition, the first N1 columns are\n*          rotated 'against' the remaining N-N1 columns of A.\n*\n*  A       (input/output) DOUBLE PRECISION array, dimension (LDA,N)\n*          On entry, M-by-N matrix A, such that A*diag(D) represents\n*          the input matrix.\n*          On exit,\n*          A_onexit * D_onexit represents the input matrix A*diag(D)\n*          post-multiplied by a sequence of Jacobi rotations, where the\n*          rotation threshold and the total number of sweeps are given in\n*          TOL and NSWEEP, respectively.\n*          (See the descriptions of N1, D, TOL and NSWEEP.)\n*\n*  LDA     (input) INTEGER\n*          The leading dimension of the array A.  LDA >= max(1,M).\n*\n*  D       (input/workspace/output) DOUBLE PRECISION array, dimension (N)\n*          The array D accumulates the scaling factors from the fast scaled\n*          Jacobi rotations.\n*          On entry, A*diag(D) represents the input matrix.\n*          On exit, A_onexit*diag(D_onexit) represents the input matrix\n*          post-multiplied by a sequence of Jacobi rotations, where the\n*          rotation threshold and the total number of sweeps are given in\n*          TOL and NSWEEP, respectively.\n*          (See the descriptions of N1, A, TOL and NSWEEP.)\n*\n*  SVA     (input/workspace/output) DOUBLE PRECISION array, dimension (N)\n*          On entry, SVA contains the Euclidean norms of the columns of\n*          the matrix A*diag(D).\n*          On exit, SVA contains the Euclidean norms of the columns of\n*          the matrix onexit*diag(D_onexit).\n*\n*  MV      (input) INTEGER\n*          If JOBV .EQ. 'A', then MV rows of V are post-multipled by a\n*                           sequence of Jacobi rotations.\n*          If JOBV = 'N',   then MV is not referenced.\n*\n*  V       (input/output) DOUBLE PRECISION array, dimension (LDV,N)\n*          If JOBV .EQ. 'V' then N rows of V are post-multipled by a\n*                           sequence of Jacobi rotations.\n*          If JOBV .EQ. 'A' then MV rows of V are post-multipled by a\n*                           sequence of Jacobi rotations.\n*          If JOBV = 'N',   then V is not referenced.\n*\n*  LDV     (input) INTEGER\n*          The leading dimension of the array V,  LDV >= 1.\n*          If JOBV = 'V', LDV .GE. N.\n*          If JOBV = 'A', LDV .GE. MV.\n*\n*  EPS     (input) DOUBLE PRECISION\n*          EPS = DLAMCH('Epsilon')\n*\n*  SFMIN   (input) DOUBLE PRECISION\n*          SFMIN = DLAMCH('Safe Minimum')\n*\n*  TOL     (input) DOUBLE PRECISION\n*          TOL is the threshold for Jacobi rotations. For a pair\n*          A(:,p), A(:,q) of pivot columns, the Jacobi rotation is\n*          applied only if DABS(COS(angle(A(:,p),A(:,q)))) .GT. TOL.\n*\n*  NSWEEP  (input) INTEGER\n*          NSWEEP is the number of sweeps of Jacobi rotations to be\n*          performed.\n*\n*  WORK    (workspace) DOUBLE PRECISION array, dimension (LWORK)\n*\n*  LWORK   (input) INTEGER\n*          LWORK is the dimension of WORK. LWORK .GE. M.\n*\n*  INFO    (output) INTEGER\n*          = 0 : successful exit.\n*          < 0 : if INFO = -i, then the i-th argument had an illegal value\n*\n\n*  =====================================================================\n*\n*     .. Local Parameters ..\n      DOUBLE PRECISION   ZERO, HALF, ONE, TWO\n      PARAMETER          ( ZERO = 0.0D0, HALF = 0.5D0, ONE = 1.0D0,\n     +                   TWO = 2.0D0 )\n*     ..\n*     .. Local Scalars ..\n      DOUBLE PRECISION   AAPP, AAPP0, AAPQ, AAQQ, APOAQ, AQOAP, BIG,\n     +                   BIGTHETA, CS, LARGE, MXAAPQ, MXSINJ, ROOTBIG,\n     +                   ROOTEPS, ROOTSFMIN, ROOTTOL, SMALL, SN, T,\n     +                   TEMP1, THETA, THSIGN\n      INTEGER            BLSKIP, EMPTSW, i, ibr, igl, IERR, IJBLSK,\n     +                   ISWROT, jbc, jgl, KBL, MVL, NOTROT, nblc, nblr,\n     +                   p, PSKIPPED, q, ROWSKIP, SWBAND\n      LOGICAL            APPLV, ROTOK, RSVEC\n*     ..\n*     .. Local Arrays ..\n      DOUBLE PRECISION   FASTR( 5 )\n*     ..\n*     .. Intrinsic Functions ..\n      INTRINSIC          DABS, DMAX1, DBLE, MIN0, DSIGN, DSQRT\n*     ..\n*     .. External Functions ..\n      DOUBLE PRECISION   DDOT, DNRM2\n      INTEGER            IDAMAX\n      LOGICAL            LSAME\n      EXTERNAL           IDAMAX, LSAME, DDOT, DNRM2\n*     ..\n*     .. External Subroutines ..\n      EXTERNAL           DAXPY, DCOPY, DLASCL, DLASSQ, DROTM, DSWAP\n*     ..\n\n");
      return Qnil;
    }
    if (rb_hash_aref(rblapack_options, sUsage) == Qtrue) {
      printf("%s\n", "USAGE:\n  info, a, d, sva, v = NumRu::Lapack.dgsvj1( jobv, m, n1, a, d, sva, mv, v, eps, sfmin, tol, nsweep, [:lwork => lwork, :usage => usage, :help => help])\n");
      return Qnil;
    } 
  } else
    rblapack_options = Qnil;
  if (argc != 12 && argc != 13)
    rb_raise(rb_eArgError,"wrong number of arguments (%d for 12)", argc);
  rblapack_jobv = argv[0];
  rblapack_m = argv[1];
  rblapack_n1 = argv[2];
  rblapack_a = argv[3];
  rblapack_d = argv[4];
  rblapack_sva = argv[5];
  rblapack_mv = argv[6];
  rblapack_v = argv[7];
  rblapack_eps = argv[8];
  rblapack_sfmin = argv[9];
  rblapack_tol = argv[10];
  rblapack_nsweep = argv[11];
  if (argc == 13) {
    rblapack_lwork = argv[12];
  } else if (rblapack_options != Qnil) {
    rblapack_lwork = rb_hash_aref(rblapack_options, ID2SYM(rb_intern("lwork")));
  } else {
    rblapack_lwork = Qnil;
  }

  jobv = StringValueCStr(rblapack_jobv)[0];
  n1 = NUM2INT(rblapack_n1);
  if (!NA_IsNArray(rblapack_d))
    rb_raise(rb_eArgError, "d (5th argument) must be NArray");
  if (NA_RANK(rblapack_d) != 1)
    rb_raise(rb_eArgError, "rank of d (5th argument) must be %d", 1);
  n = NA_SHAPE0(rblapack_d);
  if (NA_TYPE(rblapack_d) != NA_DFLOAT)
    rblapack_d = na_change_type(rblapack_d, NA_DFLOAT);
  d = NA_PTR_TYPE(rblapack_d, doublereal*);
  mv = NUM2INT(rblapack_mv);
  eps = NUM2DBL(rblapack_eps);
  tol = NUM2DBL(rblapack_tol);
  m = NUM2INT(rblapack_m);
  if (!NA_IsNArray(rblapack_sva))
    rb_raise(rb_eArgError, "sva (6th argument) must be NArray");
  if (NA_RANK(rblapack_sva) != 1)
    rb_raise(rb_eArgError, "rank of sva (6th argument) must be %d", 1);
  if (NA_SHAPE0(rblapack_sva) != n)
    rb_raise(rb_eRuntimeError, "shape 0 of sva must be the same as shape 0 of d");
  if (NA_TYPE(rblapack_sva) != NA_DFLOAT)
    rblapack_sva = na_change_type(rblapack_sva, NA_DFLOAT);
  sva = NA_PTR_TYPE(rblapack_sva, doublereal*);
  sfmin = NUM2DBL(rblapack_sfmin);
  lwork = m;
  if (!NA_IsNArray(rblapack_a))
    rb_raise(rb_eArgError, "a (4th argument) must be NArray");
  if (NA_RANK(rblapack_a) != 2)
    rb_raise(rb_eArgError, "rank of a (4th argument) must be %d", 2);
  lda = NA_SHAPE0(rblapack_a);
  if (NA_SHAPE1(rblapack_a) != n)
    rb_raise(rb_eRuntimeError, "shape 1 of a must be the same as shape 0 of d");
  if (NA_TYPE(rblapack_a) != NA_DFLOAT)
    rblapack_a = na_change_type(rblapack_a, NA_DFLOAT);
  a = NA_PTR_TYPE(rblapack_a, doublereal*);
  nsweep = NUM2INT(rblapack_nsweep);
  if (!NA_IsNArray(rblapack_v))
    rb_raise(rb_eArgError, "v (8th argument) must be NArray");
  if (NA_RANK(rblapack_v) != 2)
    rb_raise(rb_eArgError, "rank of v (8th argument) must be %d", 2);
  ldv = NA_SHAPE0(rblapack_v);
  if (NA_SHAPE1(rblapack_v) != n)
    rb_raise(rb_eRuntimeError, "shape 1 of v must be the same as shape 0 of d");
  if (NA_TYPE(rblapack_v) != NA_DFLOAT)
    rblapack_v = na_change_type(rblapack_v, NA_DFLOAT);
  v = NA_PTR_TYPE(rblapack_v, doublereal*);
  {
    na_shape_t shape[2];
    shape[0] = lda;
    shape[1] = n;
    rblapack_a_out__ = na_make_object(NA_DFLOAT, 2, shape, cNArray);
  }
  a_out__ = NA_PTR_TYPE(rblapack_a_out__, doublereal*);
  MEMCPY(a_out__, a, doublereal, NA_TOTAL(rblapack_a));
  rblapack_a = rblapack_a_out__;
  a = a_out__;
  {
    na_shape_t shape[1];
    shape[0] = n;
    rblapack_d_out__ = na_make_object(NA_DFLOAT, 1, shape, cNArray);
  }
  d_out__ = NA_PTR_TYPE(rblapack_d_out__, doublereal*);
  MEMCPY(d_out__, d, doublereal, NA_TOTAL(rblapack_d));
  rblapack_d = rblapack_d_out__;
  d = d_out__;
  {
    na_shape_t shape[1];
    shape[0] = n;
    rblapack_sva_out__ = na_make_object(NA_DFLOAT, 1, shape, cNArray);
  }
  sva_out__ = NA_PTR_TYPE(rblapack_sva_out__, doublereal*);
  MEMCPY(sva_out__, sva, doublereal, NA_TOTAL(rblapack_sva));
  rblapack_sva = rblapack_sva_out__;
  sva = sva_out__;
  {
    na_shape_t shape[2];
    shape[0] = ldv;
    shape[1] = n;
    rblapack_v_out__ = na_make_object(NA_DFLOAT, 2, shape, cNArray);
  }
  v_out__ = NA_PTR_TYPE(rblapack_v_out__, doublereal*);
  MEMCPY(v_out__, v, doublereal, NA_TOTAL(rblapack_v));
  rblapack_v = rblapack_v_out__;
  v = v_out__;
  work = ALLOC_N(doublereal, (lwork));

  dgsvj1_(&jobv, &m, &n, &n1, a, &lda, d, sva, &mv, v, &ldv, &eps, &sfmin, &tol, &nsweep, work, &lwork, &info);

  free(work);
  rblapack_info = INT2NUM(info);
  return rb_ary_new3(5, rblapack_info, rblapack_a, rblapack_d, rblapack_sva, rblapack_v);
}

void
init_lapack_dgsvj1(VALUE mLapack, VALUE sH, VALUE sU, VALUE zero){
  sHelp = sH;
  sUsage = sU;
  rblapack_ZERO = zero;

  rb_define_module_function(mLapack, "dgsvj1", rblapack_dgsvj1, -1);
}