src/gromacs/linearalgebra/gmx_arpack.h

   1 /*
   2  * This file is part of the GROMACS molecular simulation package.
   3  *
   4  * Copyright (c) 1991-2004 David van der Spoel, Erik Lindahl, University of Groningen.
   5  * Copyright (c) 2012,2013,2014, by the GROMACS development team, led by
   6  * Mark Abraham, David van der Spoel, Berk Hess, and Erik Lindahl,
   7  * and including many others, as listed in the AUTHORS file in the
   8  * top-level source directory and at http://www.gromacs.org.
   9  *
  10  * GROMACS is free software; you can redistribute it and/or
  11  * modify it under the terms of the GNU Lesser General Public License
  12  * as published by the Free Software Foundation; either version 2.1
  13  * of the License, or (at your option) any later version.
  14  *
  15  * GROMACS is distributed in the hope that it will be useful,
  16  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  17  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  18  * Lesser General Public License for more details.
  19  *
  20  * You should have received a copy of the GNU Lesser General Public
  21  * License along with GROMACS; if not, see
  22  * http://www.gnu.org/licenses, or write to the Free Software Foundation,
  23  * Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301  USA.
  24  *
  25  * If you want to redistribute modifications to GROMACS, please
  26  * consider that scientific software is very special. Version
  27  * control is crucial - bugs must be traceable. We will be happy to
  28  * consider code for inclusion in the official distribution, but
  29  * derived work must not be called official GROMACS. Details are found
  30  * in the README & COPYING files - if they are missing, get the
  31  * official version at http://www.gromacs.org.
  32  *
  33  * To help us fund GROMACS development, we humbly ask that you cite
  34  * the research papers on the package. Check out http://www.gromacs.org.
  35  */
  36 /*! \internal \file
  37  * \brief
  38  * Selected routines from ARPACK
  39  *
  40  * This file contains a subset of ARPACK functions to perform
  41  * diagonalization and SVD for sparse matrices in Gromacs.
  42  *
  43  * Consult the main ARPACK site for detailed documentation:
  44  * http://www.caam.rice.edu/software/ARPACK/
  45  *
  46  * Below, we just list the options and any specific differences
  47  * from ARPACK. The code is essentially the same, but the routines
  48  * have been made thread-safe by using extra workspace arrays.
  49  */
  50 #ifndef GMX_ARPACK_H
  51 #define GMX_ARPACK_H
  52
  53 #include "config.h"
  54 #ifdef __cplusplus
  55 extern "C" {
  56 #endif
  57
  58 /*! \brief Implicitly Restarted Arnoldi Iteration, double precision.
  59  *
  60  *  Reverse communication interface for the Implicitly Restarted Arnoldi
  61  *  Iteration.  For symmetric problems this reduces to a variant of the
  62  *  Lanczos method. See the ARPACK site for details.
  63  *
  64  *  \param ido     Reverse communication flag. Set to 0 first time.
  65  *                 Upon return with ido=-1 or ido=1 you should calculate
  66  *                 Y=A*X and recall the routine. Return with ido=2 means
  67  *                 Y=B*X should be calculated. ipntr[0] is the pointer in
  68  *                 workd for X, ipntr[1] is the index for Y.
  69  *                 Return with ido=99 means it finished.
  70  *  \param bmat    'I' for standard eigenproblem, 'G' for generalized.
  71  *  \param n       Order of eigenproblem.
  72  *  \param which   Which eigenvalues to calculate. 'LA' for largest
  73  *                 algebraic, 'SA' for smallest algebraic, 'LM' for largest
  74  *                 magnitude, 'SM' for smallest magnitude, and finally
  75  *                 'BE' (both ends) to calculate half from each end of
  76  *                 the spectrum.
  77  *  \param nev     Number of eigenvalues to calculate. 0<nev<n.
  78  *  \param tol     Tolerance. Machine precision of it is 0.
  79  *  \param resid   Optional starting residual vector at input if info=1,
  80  *                 otherwise a random one is used. Final residual vector on
  81  *                 return.
  82  *  \param ncv     Number of columns in matrix v.
  83  *  \param v       N*NCV matrix. V contain the Lanczos basis vectors.
  84  *  \param ldv     Leading dimension of v.
  85  *  \param iparam  Integer array, size 11. Same contents as arpack.
  86  *  \param ipntr   Integer array, size 11. Points to starting locations
  87  *                 in the workd/workl arrays. Same contents as arpack.
  88  *  \param workd   Double precision work array, length 3*n+4.
  89  *                 Provide the same array for all calls, and don't touch it.
  90  *                 IMPORTANT: This is 4 units larger than standard ARPACK!
  91  *  \param iwork   Integer work array, size 80.
  92  *                 Provide the same array for all calls, and don't touch it.
  93  *                 IMPORTANT: New argument compared to standard ARPACK!
  94  *  \param workl   Double precision work array, length lwork.
  95  *  \param lworkl  Length of the work array workl. Must be at least ncv*(ncv+8)
  96  *  \param info    Set info to 0 to use random initial residual vector,
  97  *                 or to 1 if you provide a one. On output, info=0 means
  98  *                 normal exit, 1 that max number of iterations was reached,
  99  *                 and 3 that no shifts could be applied. Negative numbers
 100  *                 correspond to errors in the arguments provided.
 101  */
 102 void
 103     F77_FUNC(dsaupd, DSAUPD) (int *     ido,
 104                               const char *    bmat,
 105                               int *     n,
 106                               const char *      which,
 107                               int *     nev,
 108                               double *  tol,
 109                               double *  resid,
 110                               int *     ncv,
 111                               double *  v,
 112                               int *     ldv,
 113                               int *     iparam,
 114                               int *     ipntr,
 115                               double *  workd,
 116                               int *     iwork,
 117                               double *  workl,
 118                               int *     lworkl,
 119                               int *     info);
 120
 121
 122
 123 /*! \brief Get eigenvalues/vectors after Arnoldi iteration, double prec.
 124  *
 125  *  See the ARPACK site for details. You must have finished the interative
 126  *  part with dsaupd() before calling this function.
 127  *
 128  *  \param rvec    1 if you want eigenvectors, 0 if not.
 129  *  \param howmny  'A' if you want all nvec vectors, 'S' if you
 130  *                 provide a subset selection in select[].
 131  *  \param select  Integer array, dimension nev. Indices of the
 132  *                 eigenvectors to calculate. Fortran code means we
 133  *                 start counting on 1. This array must be given even in
 134  *                 howmny is 'A'. (Arpack documentation is wrong on this).
 135  *  \param d       Double precision array, length nev. Eigenvalues.
 136  *  \param z       Double precision array, n*nev. Eigenvectors.
 137  *  \param ldz     Leading dimension of z. Normally n.
 138  *  \param sigma   Shift if iparam[6] is 3,4, or 5. Ignored otherwise.
 139  *  \param bmat    Provide the same argument as you did to dsaupd()
 140  *  \param n       Provide the same argument as you did to dsaupd()
 141  *  \param which   Provide the same argument as you did to dsaupd()
 142  *  \param nev     Provide the same argument as you did to dsaupd()
 143  *  \param tol     Provide the same argument as you did to dsaupd()
 144  *  \param resid   Provide the same argument as you did to dsaupd()
 145  *                 The array must not be touched between the two function calls!
 146  *  \param ncv     Provide the same argument as you did to dsaupd()
 147  *  \param v       Provide the same argument as you did to dsaupd()
 148  *                 The array must not be touched between the two function calls!
 149  *  \param ldv     Provide the same argument as you did to dsaupd()
 150  *  \param iparam  Provide the same argument as you did to dsaupd()
 151  *                 The array must not be touched between the two function calls!
 152  *  \param ipntr   Provide the same argument as you did to dsaupd()
 153  *                 The array must not be touched between the two function calls!
 154  *  \param workd   Provide the same argument as you did to dsaupd()
 155  *                 The array must not be touched between the two function calls!
 156  *  \param workl   Double precision work array, length lwork.
 157  *                 The array must not be touched between the two function calls!
 158  *  \param lworkl  Provide the same argument as you did to dsaupd()
 159  *  \param info    Provide the same argument as you did to dsaupd()
 160  */
 161 void
 162     F77_FUNC(dseupd, DSEUPD) (int *     rvec,
 163                               const char *    howmny,
 164                               int *     select,
 165                               double *  d,
 166                               double *  z,
 167                               int *     ldz,
 168                               double *  sigma,
 169                               const char *    bmat,
 170                               int *     n,
 171                               const char *    which,
 172                               int *     nev,
 173                               double *  tol,
 174                               double *  resid,
 175                               int *     ncv,
 176                               double *  v,
 177                               int *     ldv,
 178                               int *     iparam,
 179                               int *     ipntr,
 180                               double *  workd,
 181                               double *  workl,
 182                               int *     lworkl,
 183                               int *     info);
 184
 185
 186
 187
 188
 189 /*! \brief Implicitly Restarted Arnoldi Iteration, single precision.
 190  *
 191  *  Reverse communication interface for the Implicitly Restarted Arnoldi
 192  *  Iteration.  For symmetric problems this reduces to a variant of the
 193  *  Lanczos method. See the ARPACK site for details.
 194  *
 195  *  \param ido     Reverse communication flag. Set to 0 first time.
 196  *                 Upon return with ido=-1 or ido=1 you should calculate
 197  *                 Y=A*X and recall the routine. Return with ido=2 means
 198  *                 Y=B*X should be calculated. ipntr[0] is the pointer in
 199  *                 workd for X, ipntr[1] is the index for Y.
 200  *                 Return with ido=99 means it finished.
 201  *  \param bmat    'I' for standard eigenproblem, 'G' for generalized.
 202  *  \param n       Order of eigenproblem.
 203  *  \param which   Which eigenvalues to calculate. 'LA' for largest
 204  *                 algebraic, 'SA' for smallest algebraic, 'LM' for largest
 205  *                 magnitude, 'SM' for smallest magnitude, and finally
 206  *                 'BE' (both ends) to calculate half from each end of
 207  *                 the spectrum.
 208  *  \param nev     Number of eigenvalues to calculate. 0<nev<n.
 209  *  \param tol     Tolerance. Machine precision of it is 0.
 210  *  \param resid   Optional starting residual vector at input if info=1,
 211  *                 otherwise a random one is used. Final residual vector on
 212  *                 return.
 213  *  \param ncv     Number of columns in matrix v.
 214  *  \param v       N*NCV matrix. V contain the Lanczos basis vectors.
 215  *  \param ldv     Leading dimension of v.
 216  *  \param iparam  Integer array, size 11. Same contents as arpack.
 217  *  \param ipntr   Integer array, size 11. Points to starting locations
 218  *                 in the workd/workl arrays. Same contents as arpack.
 219  *  \param workd   Single precision work array, length 3*n+4.
 220  *                 Provide the same array for all calls, and don't touch it.
 221  *                 IMPORTANT: This is 4 units larger than standard ARPACK!
 222  *  \param iwork   Integer work array, size 80.
 223  *                 Provide the same array for all calls, and don't touch it.
 224  *                 IMPORTANT: New argument compared to standard ARPACK!
 225  *  \param workl   Single precision work array, length lwork.
 226  *  \param lworkl  Length of the work array workl. Must be at least ncv*(ncv+8)
 227  *  \param info    Set info to 0 to use random initial residual vector,
 228  *                 or to 1 if you provide a one. On output, info=0 means
 229  *                 normal exit, 1 that max number of iterations was reached,
 230  *                 and 3 that no shifts could be applied. Negative numbers
 231  *                 correspond to errors in the arguments provided.
 232  */
 233 void
 234     F77_FUNC(ssaupd, SSAUPD) (int *     ido,
 235                               const char *    bmat,
 236                               int *     n,
 237                               const char *    which,
 238                               int *     nev,
 239                               float *   tol,
 240                               float *   resid,
 241                               int *     ncv,
 242                               float *   v,
 243                               int *     ldv,
 244                               int *     iparam,
 245                               int *     ipntr,
 246                               float *   workd,
 247                               int *     iwork,
 248                               float *   workl,
 249                               int *     lworkl,
 250                               int *     info);
 251
 252
 253
 254
 255
 256 /*! \brief Get eigenvalues/vectors after Arnoldi iteration, single prec.
 257  *
 258  *  See the ARPACK site for details. You must have finished the interative
 259  *  part with ssaupd() before calling this function.
 260  *
 261  *  \param rvec    1 if you want eigenvectors, 0 if not.
 262  *  \param howmny  'A' if you want all nvec vectors, 'S' if you
 263  *                 provide a subset selection in select[].
 264  *  \param select  Integer array, dimension nev. Indices of the
 265  *                 eigenvectors to calculate. Fortran code means we
 266  *                 start counting on 1. This array must be given even in
 267  *                 howmny is 'A'. (Arpack documentation is wrong on this).
 268  *  \param d       Single precision array, length nev. Eigenvalues.
 269  *  \param z       Single precision array, n*nev. Eigenvectors.
 270  *  \param ldz     Leading dimension of z. Normally n.
 271  *  \param sigma   Shift if iparam[6] is 3,4, or 5. Ignored otherwise.
 272  *  \param bmat    Provide the same argument as you did to ssaupd()
 273  *  \param n       Provide the same argument as you did to ssaupd()
 274  *  \param which   Provide the same argument as you did to ssaupd()
 275  *  \param nev     Provide the same argument as you did to ssaupd()
 276  *  \param tol     Provide the same argument as you did to ssaupd()
 277  *  \param resid   Provide the same argument as you did to ssaupd()
 278  *                 The array must not be touched between the two function calls!
 279  *  \param ncv     Provide the same argument as you did to ssaupd()
 280  *  \param v       Provide the same argument as you did to ssaupd()
 281  *                 The array must not be touched between the two function calls!
 282  *  \param ldv     Provide the same argument as you did to ssaupd()
 283  *  \param iparam  Provide the same argument as you did to ssaupd()
 284  *                 The array must not be touched between the two function calls!
 285  *  \param ipntr   Provide the same argument as you did to ssaupd()
 286  *                 The array must not be touched between the two function calls!
 287  *  \param workd   Provide the same argument as you did to ssaupd()
 288  *                 The array must not be touched between the two function calls!
 289  *  \param workl   Single precision work array, length lwork.
 290  *                 The array must not be touched between the two function calls!
 291  *  \param lworkl  Provide the same argument as you did to ssaupd()
 292  *  \param info    Provide the same argument as you did to ssaupd()
 293  */
 294 void
 295     F77_FUNC(sseupd, SSEUPD) (int *     rvec,
 296                               const char *    howmny,
 297                               int *     select,
 298                               float *   d,
 299                               float *   z,
 300                               int *     ldz,
 301                               float *   sigma,
 302                               const char *    bmat,
 303                               int *     n,
 304                               const char *    which,
 305                               int *     nev,
 306                               float *   tol,
 307                               float *   resid,
 308                               int *     ncv,
 309                               float *   v,
 310                               int *     ldv,
 311                               int *     iparam,
 312                               int *     ipntr,
 313                               float *   workd,
 314                               float *   workl,
 315                               int *     lworkl,
 316                               int *     info);
 317
 318 #ifdef __cplusplus
 319 }
 320 #endif
 321
 322 #endif