include/sparsematrix.h

   1 /*
   2  *
   3  *                This source code is part of
   4  *
   5  *                 G   R   O   M   A   C   S
   6  *
   7  *          GROningen MAchine for Chemical Simulations
   8  *
   9  *                        VERSION 3.2.0
  10  * Written by David van der Spoel, Erik Lindahl, Berk Hess, and others.
  11  * Copyright (c) 1991-2000, University of Groningen, The Netherlands.
  12  * Copyright (c) 2001-2004, The GROMACS development team,
  13  * check out http://www.gromacs.org for more information.
  14
  15  * This program is free software; you can redistribute it and/or
  16  * modify it under the terms of the GNU General Public License
  17  * as published by the Free Software Foundation; either version 2
  18  * of the License, or (at your option) any later version.
  19  *
  20  * If you want to redistribute modifications, please consider that
  21  * scientific software is very special. Version control is crucial -
  22  * bugs must be traceable. We will be happy to consider code for
  23  * inclusion in the official distribution, but derived work must not
  24  * be called official GROMACS. Details are found in the README & COPYING
  25  * files - if they are missing, get the official version at www.gromacs.org.
  26  *
  27  * To help us fund GROMACS development, we humbly ask that you cite
  28  * the papers on the package - you can find them in the top README file.
  29  *
  30  * For more info, check our website at http://www.gromacs.org
  31  *
  32  * And Hey:
  33  * Green Red Orange Magenta Azure Cyan Skyblue
  34  */
  35 #ifdef HAVE_CONFIG_H
  36 #include <config.h>
  37 #endif
  38
  39 #ifndef _SPARSEMATRIX_H_
  40 #define _SPARSEMATRIX_H_
  41
  42
  43 #include "types/simple.h"
  44
  45 #ifdef __cplusplus
  46 extern "C" {
  47 #endif
  48
  49 /*! \brief Sparse matrix storage format
  50  *
  51  *  This structure specifies a storage format for a sparse matrix.
  52  *  The memory requirements are only proportional to the number
  53  *  of nonzero elements, and it provides a reasonably fast way to
  54  *  perform matrix-vector multiplications.
  55  *
  56  *  The data format is very similar to a neighborlist. It is optimized
  57  *  for fast access, but it is difficult to add entries. If you are
  58  *  constructing a matrix you should either do it in exactly the order
  59  *  specified here, or use some other more flexible intermediate structure.
  60  *
  61  *  The index array is of size nrow+1. All non-zero matrix elements
  62  *  on row i are stored in positions index[i] through index[i+1]-1 in
  63  *  the arrays column and value. The column array contains the column
  64  *  index for each entry, in ascending order, and the corresponding
  65  *  position in the value array contains the floating point matrix element.
  66  *
  67  *  index[nrow] should be equal to the total number of elements stored.
  68  *
  69  *  Thus, to find the value of matrix element [5,4] you should loop
  70  *  over positions index[5] to index[6]-1 in column until you either find
  71  *  the value 4, or a higher value (meaning the element was zero).
  72  *
  73  *  It is fairly easy to construct the matrix on-the-fly if you can do
  74  *  it row-by-row.
  75  *
  76  *  IMPORTANT:
  77  *  If compressed_symmetric is set to TRUE, you should only store EITHER the upper OR
  78  *  lower triangle (and the diagonal), and the other half is assumed to be
  79  *  symmetric. Otherwise, if compressed_symmetric==FALSE, no symmetry is implied and all
  80  *  elements should be stored.
  81  *
  82  *  The symmetry compression saves us a factor 2 both in storage and
  83  *  matrix multiplication CPU-time, which can be very useful for huge eigenproblems.
  84  *
  85  *  If you are unsure, just set compressed_symmetric to FALSE and list all elements. If
  86  *  you enable it but still list all elements (both upper and lower triangle) you will be sorry...
  87  *
  88  *  Internally, the sparse data is stored as a separate list for each row, where the list
  89  *  element is a structure with a column and (floating-point) data value. This makes it
  90  *  possible, although not completely transparent, to update values in random access order.
  91  *  The drawback is that the structure will allocate nrow memory regions.
  92  *  The matrix data could be stored in a single contiguous array with indices for each row,
  93  *  but then we could only insert elements at the end without copying the entire matrix.
  94  *
  95  *  After you have
  96  *
  97  *  In other words: Not perfect, but it works.
  98  */
  99
 100 typedef struct
 101 gmx_sparsematrix_entry
 102 {
 103     int      col;
 104     real     value;
 105 } gmx_sparsematrix_entry_t;
 106
 107 typedef struct
 108 gmx_sparsematrix
 109 {
 110     bool                         compressed_symmetric; /*!< Store half elements and assume symmetry. */
 111     int                          nrow;                 /*!< Number of rows in matrix                 */
 112     int *                        ndata;                /*!< Number of entries on each row (list)     */
 113     int *                        nalloc;               /*!< Allocated entry list length for each row */
 114     gmx_sparsematrix_entry_t **  data;                 /*!< data[i] is a list with entries on row i  */
 115 }
 116 gmx_sparsematrix_t;
 117
 118
 119 /*! \Allocate a new sparse matrix structure
 120  *
 121  *  The number of rows is used to allocate the index array entry. Obviously you
 122  *  can reallocate these later yourself if necessary - this is a
 123  *  convenience routine.
 124  *
 125  *  By default, the compressed_symmetric flag in the structure will
 126  *  be FALSE. Set it to TRUE manually if you are only storing either the
 127  *  upper or lower half of the matrix.
 128  */
 129 gmx_sparsematrix_t *
 130 gmx_sparsematrix_init            (int                    nrow);
 131
 132
 133 /*! \brief Release all resources used by a sparse matrix structure
 134  *
 135  *  All arrays in the structure will be freed, and the structure itself.
 136  */
 137 void
 138 gmx_sparsematrix_destroy         (gmx_sparsematrix_t *   A);
 139
 140
 141 /*! \brief Print sparse matrix to a stream.
 142  *
 143  *  Mainly used for debugging. Be warned that the real sparse matrices used
 144  *  in Gromacs runs can be HUGE (think 100,000 rows).
 145  */
 146 void
 147 gmx_sparsematrix_print           (FILE *                 stream,
 148                                   gmx_sparsematrix_t *   A);
 149
 150 /* Adds value at row,col. If the value did not exist
 151  * previously it is added, otherwise it is incremented with difference.
 152  *
 153  * The column sort order might change, so you need to run fix_sparsematrix
 154  * once you are done changing the matrix.
 155  */
 156 real
 157 gmx_sparsematrix_value          (gmx_sparsematrix_t *    A,
 158                                  int                     row,
 159                                  int                     col);
 160
 161
 162 /* Adds value at row,col. If the value did not exist
 163 * previously it is added, otherwise it is incremented with difference.
 164 *
 165 * The column sort order might change, so you need to run fix_sparsematrix
 166 * once you are done changing the matrix.
 167 */
 168 void
 169 gmx_sparsematrix_increment_value(gmx_sparsematrix_t *    A,
 170                                  int                     row,
 171                                  int                     col,
 172                                  real                    difference);
 173
 174
 175
 176 /*! \brief Sort elements in each column and remove zeros.
 177  *
 178  *  Sparse matrix access is faster when the elements are stored in
 179  *  increasing column order in each row. In some cases previously non-zero
 180  *  elements will be zero after adding more data, and this routine also removes
 181  *  those entries to reduce the storage requirements.
 182  *
 183  *  It never hurts to run this routine if you have been updating the matrix...
 184  */
 185 void
 186 gmx_sparsematrix_compress       (gmx_sparsematrix_t *    A);
 187
 188
 189
 190 /*! \brief Sparse matrix vector multiplication
 191  *
 192  * Calculate y = A * x for a sparse matrix A.
 193  */
 194 void
 195 gmx_sparsematrix_vector_multiply(gmx_sparsematrix_t *    A,
 196                                  real *                  x,
 197                                  real *                  y);
 198
 199 #ifdef __cplusplus
 200 }
 201 #endif
 202
 203
 204 #endif
 205
 206