xref: /petsc/src/mat/graphops/color/impls/minpack/dsm.c (revision 53673ba54f5aaba04b9d49ab22cf56c7a7461fe9)
1 /* dsm.f -- translated by f2c (version of 25 March 1992  12:58:56). */
2 
3 #include <../src/mat/graphops/color/impls/minpack/color.h>
4 
5 static PetscInt c_n1 = -1;
6 
MINPACKdsm(PetscInt * m,PetscInt * n,PetscInt * npairs,PetscInt * indrow,PetscInt * indcol,PetscInt * ngrp,PetscInt * maxgrp,PetscInt * mingrp,PetscInt * info,PetscInt * ipntr,PetscInt * jpntr,PetscInt * iwa,PetscInt * liwa)7 PetscErrorCode MINPACKdsm(PetscInt *m, PetscInt *n, PetscInt *npairs, PetscInt *indrow, PetscInt *indcol, PetscInt *ngrp, PetscInt *maxgrp, PetscInt *mingrp, PetscInt *info, PetscInt *ipntr, PetscInt *jpntr, PetscInt *iwa, PetscInt *liwa)
8 {
9   /* System generated locals */
10   PetscInt i__1, i__2, i__3;
11 
12   /* Local variables */
13   PetscInt i, j, maxclq, numgrp;
14 
15   /*     Given the sparsity pattern of an m by n matrix A, this */
16   /*     subroutine determines a partition of the columns of A */
17   /*     consistent with the direct determination of A. */
18   /*     The sparsity pattern of the matrix A is specified by */
19   /*     the arrays indrow and indcol. On input the indices */
20   /*     for the non-zero elements of A are */
21   /*           indrow(k),indcol(k), k = 1,2,...,npairs. */
22   /*     The (indrow,indcol) pairs may be specified in any order. */
23   /*     Duplicate input pairs are permitted, but the subroutine */
24   /*     eliminates them. */
25   /*     The subroutine partitions the columns of A into groups */
26   /*     such that columns in the same group do not have a */
27   /*     non-zero in the same row position. A partition of the */
28   /*     columns of A with this property is consistent with the */
29   /*     direct determination of A. */
30   /*     The subroutine statement is */
31   /*       subroutine dsm(m,n,npairs,indrow,indcol,ngrp,maxgrp,mingrp, */
32   /*                      info,ipntr,jpntr,iwa,liwa) */
33   /*     where */
34   /*       m is a positive integer input variable set to the number */
35   /*         of rows of A. */
36   /*       n is a positive integer input variable set to the number */
37   /*         of columns of A. */
38   /*       npairs is a positive integer input variable set to the */
39   /*         number of (indrow,indcol) pairs used to describe the */
40   /*         sparsity pattern of A. */
41   /*       indrow is an integer array of length npairs. On input indrow */
42   /*         must contain the row indices of the non-zero elements of A. */
43   /*         On output indrow is permuted so that the corresponding */
44   /*         column indices are in non-decreasing order. The column */
45   /*         indices can be recovered from the array jpntr. */
46   /*       indcol is an integer array of length npairs. On input indcol */
47   /*         must contain the column indices of the non-zero elements of */
48   /*         A. On output indcol is permuted so that the corresponding */
49   /*         row indices are in non-decreasing order. The row indices */
50   /*         can be recovered from the array ipntr. */
51   /*       ngrp is an integer output array of length n which specifies */
52   /*         the partition of the columns of A. Column jcol belongs */
53   /*         to group ngrp(jcol). */
54   /*       maxgrp is an integer output variable which specifies the */
55   /*         number of groups in the partition of the columns of A. */
56   /*       mingrp is an integer output variable which specifies a lower */
57   /*         bound for the number of groups in any consistent partition */
58   /*         of the columns of A. */
59   /*       info is an integer output variable set as follows. For */
60   /*         normal termination info = 1. If m, n, or npairs is not */
61   /*         positive or liwa is less than max(m,6*n), then info = 0. */
62   /*         If the k-th element of indrow is not an integer between */
63   /*         1 and m or the k-th element of indcol is not an integer */
64   /*         between 1 and n, then info = -k. */
65   /*       ipntr is an integer output array of length m + 1 which */
66   /*         specifies the locations of the column indices in indcol. */
67   /*         The column indices for row i are */
68   /*               indcol(k), k = ipntr(i),...,ipntr(i+1)-1. */
69   /*         Note that ipntr(m+1)-1 is then the number of non-zero */
70   /*         elements of the matrix A. */
71   /*       jpntr is an integer output array of length n + 1 which */
72   /*         specifies the locations of the row indices in indrow. */
73   /*         The row indices for column j are */
74   /*               indrow(k), k = jpntr(j),...,jpntr(j+1)-1. */
75   /*         Note that jpntr(n+1)-1 is then the number of non-zero */
76   /*         elements of the matrix A. */
77   /*       iwa is an integer work array of length liwa. */
78   /*       liwa is a positive integer input variable not less than */
79   /*         max(m,6*n). */
80   /*     Subprograms called */
81   /*       MINPACK-supplied ... degr,ido,numsrt,seq,setr,slo,srtdat */
82   /*       FORTRAN-supplied ... max */
83   /*     Argonne National Laboratory. MINPACK Project. December 1984. */
84   /*     Thomas F. Coleman, Burton S. Garbow, Jorge J. More' */
85 
86   PetscFunctionBegin;
87   /* Parameter adjustments */
88   --iwa;
89   --jpntr;
90   --ipntr;
91   --ngrp;
92   --indcol;
93   --indrow;
94 
95   *info = 0;
96 
97   /*     Determine a lower bound for the number of groups. */
98 
99   *mingrp = 0;
100   i__1    = *m;
101   for (i = 1; i <= i__1; ++i) {
102     /* Computing MAX */
103     i__2    = *mingrp;
104     i__3    = ipntr[i + 1] - ipntr[i];
105     *mingrp = PetscMax(i__2, i__3);
106   }
107 
108   /*     Determine the degree sequence for the intersection */
109   /*     graph of the columns of A. */
110 
111   PetscCall(MINPACKdegr(n, &indrow[1], &jpntr[1], &indcol[1], &ipntr[1], &iwa[*n * 5 + 1], &iwa[*n + 1]));
112 
113   /*     Color the intersection graph of the columns of A */
114   /*     with the smallest-last (SL) ordering. */
115 
116   PetscCall(MINPACKslo(n, &indrow[1], &jpntr[1], &indcol[1], &ipntr[1], &iwa[*n * 5 + 1], &iwa[(*n << 2) + 1], &maxclq, &iwa[1], &iwa[*n + 1], &iwa[(*n << 1) + 1], &iwa[*n * 3 + 1]));
117   PetscCall(MINPACKseq(n, &indrow[1], &jpntr[1], &indcol[1], &ipntr[1], &iwa[(*n << 2) + 1], &ngrp[1], maxgrp, &iwa[*n + 1]));
118   *mingrp = PetscMax(*mingrp, maxclq);
119 
120   /*     Exit if the smallest-last ordering is optimal. */
121 
122   if (*maxgrp == *mingrp) PetscFunctionReturn(PETSC_SUCCESS);
123 
124   /*     Color the intersection graph of the columns of A */
125   /*     with the incidence-degree (ID) ordering. */
126 
127   PetscCall(MINPACKido(m, n, &indrow[1], &jpntr[1], &indcol[1], &ipntr[1], &iwa[*n * 5 + 1], &iwa[(*n << 2) + 1], &maxclq, &iwa[1], &iwa[*n + 1], &iwa[(*n << 1) + 1], &iwa[*n * 3 + 1]));
128   PetscCall(MINPACKseq(n, &indrow[1], &jpntr[1], &indcol[1], &ipntr[1], &iwa[(*n << 2) + 1], &iwa[1], &numgrp, &iwa[*n + 1]));
129   *mingrp = PetscMax(*mingrp, maxclq);
130 
131   /*     Retain the better of the two orderings so far. */
132 
133   if (numgrp < *maxgrp) {
134     *maxgrp = numgrp;
135     i__1    = *n;
136     for (j = 1; j <= i__1; ++j) ngrp[j] = iwa[j];
137 
138     /*        Exit if the incidence-degree ordering is optimal. */
139 
140     if (*maxgrp == *mingrp) PetscFunctionReturn(PETSC_SUCCESS);
141   }
142 
143   /*     Color the intersection graph of the columns of A */
144   /*     with the largest-first (LF) ordering. */
145 
146   i__1 = *n - 1;
147   PetscCall(MINPACKnumsrt(n, &i__1, &iwa[*n * 5 + 1], &c_n1, &iwa[(*n << 2) + 1], &iwa[(*n << 1) + 1], &iwa[*n + 1]));
148   PetscCall(MINPACKseq(n, &indrow[1], &jpntr[1], &indcol[1], &ipntr[1], &iwa[(*n << 2) + 1], &iwa[1], &numgrp, &iwa[*n + 1]));
149 
150   /*     Retain the best of the three orderings and exit. */
151 
152   if (numgrp < *maxgrp) {
153     *maxgrp = numgrp;
154     i__1    = *n;
155     for (j = 1; j <= i__1; ++j) ngrp[j] = iwa[j];
156   }
157   PetscFunctionReturn(PETSC_SUCCESS);
158 }
159