src/gromacs/mdlib/ns.c

   1 /*
   2  * This file is part of the GROMACS molecular simulation package.
   3  *
   4  * Copyright (c) 1991-2000, University of Groningen, The Netherlands.
   5  * Copyright (c) 2001-2004, The GROMACS development team.
   6  * Copyright (c) 2013,2014, by the GROMACS development team, led by
   7  * Mark Abraham, David van der Spoel, Berk Hess, and Erik Lindahl,
   8  * and including many others, as listed in the AUTHORS file in the
   9  * top-level source directory and at http://www.gromacs.org.
  10  *
  11  * GROMACS is free software; you can redistribute it and/or
  12  * modify it under the terms of the GNU Lesser General Public License
  13  * as published by the Free Software Foundation; either version 2.1
  14  * of the License, or (at your option) any later version.
  15  *
  16  * GROMACS is distributed in the hope that it will be useful,
  17  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  18  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  19  * Lesser General Public License for more details.
  20  *
  21  * You should have received a copy of the GNU Lesser General Public
  22  * License along with GROMACS; if not, see
  23  * http://www.gnu.org/licenses, or write to the Free Software Foundation,
  24  * Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301  USA.
  25  *
  26  * If you want to redistribute modifications to GROMACS, please
  27  * consider that scientific software is very special. Version
  28  * control is crucial - bugs must be traceable. We will be happy to
  29  * consider code for inclusion in the official distribution, but
  30  * derived work must not be called official GROMACS. Details are found
  31  * in the README & COPYING files - if they are missing, get the
  32  * official version at http://www.gromacs.org.
  33  *
  34  * To help us fund GROMACS development, we humbly ask that you cite
  35  * the research papers on the package. Check out http://www.gromacs.org.
  36  */
  37 #ifdef HAVE_CONFIG_H
  38 #include <config.h>
  39 #endif
  40
  41 #include <math.h>
  42 #include <string.h>
  43 #include "sysstuff.h"
  44 #include "gromacs/utility/smalloc.h"
  45 #include "macros.h"
  46 #include "gromacs/math/utilities.h"
  47 #include "vec.h"
  48 #include "types/commrec.h"
  49 #include "network.h"
  50 #include "nsgrid.h"
  51 #include "force.h"
  52 #include "nonbonded.h"
  53 #include "ns.h"
  54 #include "pbc.h"
  55 #include "names.h"
  56 #include "gmx_fatal.h"
  57 #include "nrnb.h"
  58 #include "txtdump.h"
  59 #include "mtop_util.h"
  60
  61 #include "domdec.h"
  62 #include "adress.h"
  63
  64
  65 /*
  66  *    E X C L U S I O N   H A N D L I N G
  67  */
  68
  69 #ifdef DEBUG
  70 static void SETEXCL_(t_excl e[], atom_id i, atom_id j)
  71 {
  72     e[j] = e[j] | (1<<i);
  73 }
  74 static void RMEXCL_(t_excl e[], atom_id i, atom_id j)
  75 {
  76     e[j] = e[j] & ~(1<<i);
  77 }
  78 static gmx_bool ISEXCL_(t_excl e[], atom_id i, atom_id j)
  79 {
  80     return (gmx_bool)(e[j] & (1<<i));
  81 }
  82 static gmx_bool NOTEXCL_(t_excl e[], atom_id i, atom_id j)
  83 {
  84     return !(ISEXCL(e, i, j));
  85 }
  86 #else
  87 #define SETEXCL(e, i, j) (e)[((atom_id) (j))] |= (1<<((atom_id) (i)))
  88 #define RMEXCL(e, i, j)  (e)[((atom_id) (j))] &= (~(1<<((atom_id) (i))))
  89 #define ISEXCL(e, i, j)  (gmx_bool) ((e)[((atom_id) (j))] & (1<<((atom_id) (i))))
  90 #define NOTEXCL(e, i, j) !(ISEXCL(e, i, j))
  91 #endif
  92
  93 static int
  94 round_up_to_simd_width(int length, int simd_width)
  95 {
  96     int offset, newlength;
  97
  98     offset = (simd_width > 0) ? length % simd_width : 0;
  99
 100     return (offset == 0) ? length : length-offset+simd_width;
 101 }
 102 /************************************************
 103  *
 104  *  U T I L I T I E S    F O R    N S
 105  *
 106  ************************************************/
 107
 108 void reallocate_nblist(t_nblist *nl)
 109 {
 110     if (gmx_debug_at)
 111     {
 112         fprintf(debug, "reallocating neigborlist (ielec=%d, ivdw=%d, igeometry=%d, type=%d), maxnri=%d\n",
 113                 nl->ielec, nl->ivdw, nl->igeometry, nl->type, nl->maxnri);
 114     }
 115     srenew(nl->iinr,   nl->maxnri);
 116     if (nl->igeometry == GMX_NBLIST_GEOMETRY_CG_CG)
 117     {
 118         srenew(nl->iinr_end, nl->maxnri);
 119     }
 120     srenew(nl->gid,    nl->maxnri);
 121     srenew(nl->shift,  nl->maxnri);
 122     srenew(nl->jindex, nl->maxnri+1);
 123 }
 124
 125
 126 static void init_nblist(FILE *log, t_nblist *nl_sr, t_nblist *nl_lr,
 127                         int maxsr, int maxlr,
 128                         int ivdw, int ivdwmod,
 129                         int ielec, int ielecmod,
 130                         int igeometry, int type)
 131 {
 132     t_nblist *nl;
 133     int       homenr;
 134     int       i, nn;
 135
 136     for (i = 0; (i < 2); i++)
 137     {
 138         nl     = (i == 0) ? nl_sr : nl_lr;
 139         homenr = (i == 0) ? maxsr : maxlr;
 140
 141         if (nl == NULL)
 142         {
 143             continue;
 144         }
 145
 146
 147         /* Set coul/vdw in neighborlist, and for the normal loops we determine
 148          * an index of which one to call.
 149          */
 150         nl->ivdw        = ivdw;
 151         nl->ivdwmod     = ivdwmod;
 152         nl->ielec       = ielec;
 153         nl->ielecmod    = ielecmod;
 154         nl->type        = type;
 155         nl->igeometry   = igeometry;
 156
 157         if (nl->type == GMX_NBLIST_INTERACTION_FREE_ENERGY)
 158         {
 159             nl->igeometry  = GMX_NBLIST_GEOMETRY_PARTICLE_PARTICLE;
 160         }
 161
 162         /* This will also set the simd_padding_width field */
 163         gmx_nonbonded_set_kernel_pointers( (i == 0) ? log : NULL, nl);
 164
 165         /* maxnri is influenced by the number of shifts (maximum is 8)
 166          * and the number of energy groups.
 167          * If it is not enough, nl memory will be reallocated during the run.
 168          * 4 seems to be a reasonable factor, which only causes reallocation
 169          * during runs with tiny and many energygroups.
 170          */
 171         nl->maxnri      = homenr*4;
 172         nl->maxnrj      = 0;
 173         nl->nri         = -1;
 174         nl->nrj         = 0;
 175         nl->iinr        = NULL;
 176         nl->gid         = NULL;
 177         nl->shift       = NULL;
 178         nl->jindex      = NULL;
 179         nl->jjnr        = NULL;
 180         nl->excl_fep    = NULL;
 181         reallocate_nblist(nl);
 182         nl->jindex[0] = 0;
 183
 184         if (debug)
 185         {
 186             fprintf(debug, "Initiating neighbourlist (ielec=%d, ivdw=%d, type=%d) for %s interactions,\nwith %d SR, %d LR atoms.\n",
 187                     nl->ielec, nl->ivdw, nl->type, gmx_nblist_geometry_names[nl->igeometry], maxsr, maxlr);
 188         }
 189     }
 190 }
 191
 192 void init_neighbor_list(FILE *log, t_forcerec *fr, int homenr)
 193 {
 194     /* Make maxlr tunable! (does not seem to be a big difference though)
 195      * This parameter determines the number of i particles in a long range
 196      * neighbourlist. Too few means many function calls, too many means
 197      * cache trashing.
 198      */
 199     int        maxsr, maxsr_wat, maxlr, maxlr_wat;
 200     int        ielec, ielecf, ivdw, ielecmod, ielecmodf, ivdwmod, type;
 201     int        solvent;
 202     int        igeometry_def, igeometry_w, igeometry_ww;
 203     int        i;
 204     t_nblists *nbl;
 205
 206     /* maxsr     = homenr-fr->nWatMol*3; */
 207     maxsr     = homenr;
 208
 209     if (maxsr < 0)
 210     {
 211         gmx_fatal(FARGS, "%s, %d: Negative number of short range atoms.\n"
 212                   "Call your Gromacs dealer for assistance.", __FILE__, __LINE__);
 213     }
 214     /* This is just for initial allocation, so we do not reallocate
 215      * all the nlist arrays many times in a row.
 216      * The numbers seem very accurate, but they are uncritical.
 217      */
 218     maxsr_wat = min(fr->nWatMol, (homenr+2)/3);
 219     if (fr->bTwinRange)
 220     {
 221         maxlr     = 50;
 222         maxlr_wat = min(maxsr_wat, maxlr);
 223     }
 224     else
 225     {
 226         maxlr = maxlr_wat = 0;
 227     }
 228
 229     /* Determine the values for ielec/ivdw. */
 230     ielec    = fr->nbkernel_elec_interaction;
 231     ivdw     = fr->nbkernel_vdw_interaction;
 232     ielecmod = fr->nbkernel_elec_modifier;
 233     ivdwmod  = fr->nbkernel_vdw_modifier;
 234     type     = GMX_NBLIST_INTERACTION_STANDARD;
 235
 236     fr->ns.bCGlist = (getenv("GMX_NBLISTCG") != 0);
 237     if (!fr->ns.bCGlist)
 238     {
 239         igeometry_def = GMX_NBLIST_GEOMETRY_PARTICLE_PARTICLE;
 240     }
 241     else
 242     {
 243         igeometry_def = GMX_NBLIST_GEOMETRY_CG_CG;
 244         if (log != NULL)
 245         {
 246             fprintf(log, "\nUsing charge-group - charge-group neighbor lists and kernels\n\n");
 247         }
 248     }
 249
 250     if (fr->solvent_opt == esolTIP4P)
 251     {
 252         igeometry_w  = GMX_NBLIST_GEOMETRY_WATER4_PARTICLE;
 253         igeometry_ww = GMX_NBLIST_GEOMETRY_WATER4_WATER4;
 254     }
 255     else
 256     {
 257         igeometry_w  = GMX_NBLIST_GEOMETRY_WATER3_PARTICLE;
 258         igeometry_ww = GMX_NBLIST_GEOMETRY_WATER3_WATER3;
 259     }
 260
 261     for (i = 0; i < fr->nnblists; i++)
 262     {
 263         nbl = &(fr->nblists[i]);
 264
 265         if ((fr->adress_type != eAdressOff) && (i >= fr->nnblists/2))
 266         {
 267             type = GMX_NBLIST_INTERACTION_ADRESS;
 268         }
 269         init_nblist(log, &nbl->nlist_sr[eNL_VDWQQ], &nbl->nlist_lr[eNL_VDWQQ],
 270                     maxsr, maxlr, ivdw, ivdwmod, ielec, ielecmod, igeometry_def, type);
 271         init_nblist(log, &nbl->nlist_sr[eNL_VDW], &nbl->nlist_lr[eNL_VDW],
 272                     maxsr, maxlr, ivdw, ivdwmod, GMX_NBKERNEL_ELEC_NONE, eintmodNONE, igeometry_def, type);
 273         init_nblist(log, &nbl->nlist_sr[eNL_QQ], &nbl->nlist_lr[eNL_QQ],
 274                     maxsr, maxlr, GMX_NBKERNEL_VDW_NONE, eintmodNONE, ielec, ielecmod, igeometry_def, type);
 275         init_nblist(log, &nbl->nlist_sr[eNL_VDWQQ_WATER], &nbl->nlist_lr[eNL_VDWQQ_WATER],
 276                     maxsr_wat, maxlr_wat, ivdw, ivdwmod, ielec, ielecmod, igeometry_w, type);
 277         init_nblist(log, &nbl->nlist_sr[eNL_QQ_WATER], &nbl->nlist_lr[eNL_QQ_WATER],
 278                     maxsr_wat, maxlr_wat, GMX_NBKERNEL_VDW_NONE, eintmodNONE, ielec, ielecmod, igeometry_w, type);
 279         init_nblist(log, &nbl->nlist_sr[eNL_VDWQQ_WATERWATER], &nbl->nlist_lr[eNL_VDWQQ_WATERWATER],
 280                     maxsr_wat, maxlr_wat, ivdw, ivdwmod, ielec, ielecmod, igeometry_ww, type);
 281         init_nblist(log, &nbl->nlist_sr[eNL_QQ_WATERWATER], &nbl->nlist_lr[eNL_QQ_WATERWATER],
 282                     maxsr_wat, maxlr_wat, GMX_NBKERNEL_VDW_NONE, eintmodNONE, ielec, ielecmod, igeometry_ww, type);
 283
 284         /* Did we get the solvent loops so we can use optimized water kernels? */
 285         if (nbl->nlist_sr[eNL_VDWQQ_WATER].kernelptr_vf == NULL
 286             || nbl->nlist_sr[eNL_QQ_WATER].kernelptr_vf == NULL
 287 #ifndef DISABLE_WATERWATER_NLIST
 288             || nbl->nlist_sr[eNL_VDWQQ_WATERWATER].kernelptr_vf == NULL
 289             || nbl->nlist_sr[eNL_QQ_WATERWATER].kernelptr_vf == NULL
 290 #endif
 291             )
 292         {
 293             fr->solvent_opt = esolNO;
 294             fprintf(log, "Note: The available nonbonded kernels do not support water optimization - disabling.\n");
 295         }
 296
 297         if (fr->efep != efepNO)
 298         {
 299             if ((fr->bEwald) && (fr->sc_alphacoul > 0)) /* need to handle long range differently if using softcore */
 300             {
 301                 ielecf    = GMX_NBKERNEL_ELEC_EWALD;
 302                 ielecmodf = eintmodNONE;
 303             }
 304             else
 305             {
 306                 ielecf    = ielec;
 307                 ielecmodf = ielecmod;
 308             }
 309
 310             init_nblist(log, &nbl->nlist_sr[eNL_VDWQQ_FREE], &nbl->nlist_lr[eNL_VDWQQ_FREE],
 311                         maxsr, maxlr, ivdw, ivdwmod, ielecf, ielecmod, GMX_NBLIST_GEOMETRY_PARTICLE_PARTICLE, GMX_NBLIST_INTERACTION_FREE_ENERGY);
 312             init_nblist(log, &nbl->nlist_sr[eNL_VDW_FREE], &nbl->nlist_lr[eNL_VDW_FREE],
 313                         maxsr, maxlr, ivdw, ivdwmod, GMX_NBKERNEL_ELEC_NONE, eintmodNONE, GMX_NBLIST_GEOMETRY_PARTICLE_PARTICLE, GMX_NBLIST_INTERACTION_FREE_ENERGY);
 314             init_nblist(log, &nbl->nlist_sr[eNL_QQ_FREE], &nbl->nlist_lr[eNL_QQ_FREE],
 315                         maxsr, maxlr, GMX_NBKERNEL_VDW_NONE, eintmodNONE, ielecf, ielecmod, GMX_NBLIST_GEOMETRY_PARTICLE_PARTICLE, GMX_NBLIST_INTERACTION_FREE_ENERGY);
 316         }
 317     }
 318     /* QMMM MM list */
 319     if (fr->bQMMM && fr->qr->QMMMscheme != eQMMMschemeoniom)
 320     {
 321         init_nblist(log, &fr->QMMMlist, NULL,
 322                     maxsr, maxlr, 0, 0, ielec, ielecmod, GMX_NBLIST_GEOMETRY_PARTICLE_PARTICLE, GMX_NBLIST_INTERACTION_STANDARD);
 323     }
 324
 325     if (log != NULL)
 326     {
 327         fprintf(log, "\n");
 328     }
 329
 330     fr->ns.nblist_initialized = TRUE;
 331 }
 332
 333 static void reset_nblist(t_nblist *nl)
 334 {
 335     nl->nri       = -1;
 336     nl->nrj       = 0;
 337     if (nl->jindex)
 338     {
 339         nl->jindex[0] = 0;
 340     }
 341 }
 342
 343 static void reset_neighbor_lists(t_forcerec *fr, gmx_bool bResetSR, gmx_bool bResetLR)
 344 {
 345     int n, i;
 346
 347     if (fr->bQMMM)
 348     {
 349         /* only reset the short-range nblist */
 350         reset_nblist(&(fr->QMMMlist));
 351     }
 352
 353     for (n = 0; n < fr->nnblists; n++)
 354     {
 355         for (i = 0; i < eNL_NR; i++)
 356         {
 357             if (bResetSR)
 358             {
 359                 reset_nblist( &(fr->nblists[n].nlist_sr[i]) );
 360             }
 361             if (bResetLR)
 362             {
 363                 reset_nblist( &(fr->nblists[n].nlist_lr[i]) );
 364             }
 365         }
 366     }
 367 }
 368
 369
 370
 371
 372 static gmx_inline void new_i_nblist(t_nblist *nlist, atom_id i_atom, int shift, int gid)
 373 {
 374     int    i, k, nri, nshift;
 375
 376     nri = nlist->nri;
 377
 378     /* Check whether we have to increase the i counter */
 379     if ((nri == -1) ||
 380         (nlist->iinr[nri]  != i_atom) ||
 381         (nlist->shift[nri] != shift) ||
 382         (nlist->gid[nri]   != gid))
 383     {
 384         /* This is something else. Now see if any entries have
 385          * been added in the list of the previous atom.
 386          */
 387         if ((nri == -1) ||
 388             ((nlist->jindex[nri+1] > nlist->jindex[nri]) &&
 389              (nlist->gid[nri] != -1)))
 390         {
 391             /* If so increase the counter */
 392             nlist->nri++;
 393             nri++;
 394             if (nlist->nri >= nlist->maxnri)
 395             {
 396                 nlist->maxnri += over_alloc_large(nlist->nri);
 397                 reallocate_nblist(nlist);
 398             }
 399         }
 400         /* Set the number of neighbours and the atom number */
 401         nlist->jindex[nri+1] = nlist->jindex[nri];
 402         nlist->iinr[nri]     = i_atom;
 403         nlist->gid[nri]      = gid;
 404         nlist->shift[nri]    = shift;
 405     }
 406     else
 407     {
 408         /* Adding to previous list. First remove possible previous padding */
 409         if (nlist->simd_padding_width > 1)
 410         {
 411             while (nlist->nrj > 0 && nlist->jjnr[nlist->nrj-1] < 0)
 412             {
 413                 nlist->nrj--;
 414             }
 415         }
 416     }
 417 }
 418
 419 static gmx_inline void close_i_nblist(t_nblist *nlist)
 420 {
 421     int nri = nlist->nri;
 422     int len;
 423
 424     if (nri >= 0)
 425     {
 426         /* Add elements up to padding. Since we allocate memory in units
 427          * of the simd_padding width, we do not have to check for possible
 428          * list reallocation here.
 429          */
 430         while ((nlist->nrj % nlist->simd_padding_width) != 0)
 431         {
 432             /* Use -4 here, so we can write forces for 4 atoms before real data */
 433             nlist->jjnr[nlist->nrj++] = -4;
 434         }
 435         nlist->jindex[nri+1] = nlist->nrj;
 436
 437         len = nlist->nrj -  nlist->jindex[nri];
 438     }
 439 }
 440
 441 static gmx_inline void close_nblist(t_nblist *nlist)
 442 {
 443     /* Only close this nblist when it has been initialized.
 444      * Avoid the creation of i-lists with no j-particles.
 445      */
 446     if (nlist->nrj == 0)
 447     {
 448         /* Some assembly kernels do not support empty lists,
 449          * make sure here that we don't generate any empty lists.
 450          * With the current ns code this branch is taken in two cases:
 451          * No i-particles at all: nri=-1 here
 452          * There are i-particles, but no j-particles; nri=0 here
 453          */
 454         nlist->nri = 0;
 455     }
 456     else
 457     {
 458         /* Close list number nri by incrementing the count */
 459         nlist->nri++;
 460     }
 461 }
 462
 463 static gmx_inline void close_neighbor_lists(t_forcerec *fr, gmx_bool bMakeQMMMnblist)
 464 {
 465     int n, i;
 466
 467     if (bMakeQMMMnblist)
 468     {
 469         close_nblist(&(fr->QMMMlist));
 470     }
 471
 472     for (n = 0; n < fr->nnblists; n++)
 473     {
 474         for (i = 0; (i < eNL_NR); i++)
 475         {
 476             close_nblist(&(fr->nblists[n].nlist_sr[i]));
 477             close_nblist(&(fr->nblists[n].nlist_lr[i]));
 478         }
 479     }
 480 }
 481
 482
 483 static gmx_inline void add_j_to_nblist(t_nblist *nlist, atom_id j_atom, gmx_bool bLR)
 484 {
 485     int nrj = nlist->nrj;
 486
 487     if (nlist->nrj >= nlist->maxnrj)
 488     {
 489         nlist->maxnrj = round_up_to_simd_width(over_alloc_small(nlist->nrj + 1), nlist->simd_padding_width);
 490
 491         if (gmx_debug_at)
 492         {
 493             fprintf(debug, "Increasing %s nblist (ielec=%d,ivdw=%d,type=%d,igeometry=%d) j size to %d\n",
 494                     bLR ? "LR" : "SR", nlist->ielec, nlist->ivdw, nlist->type, nlist->igeometry, nlist->maxnrj);
 495         }
 496
 497         srenew(nlist->jjnr, nlist->maxnrj);
 498     }
 499
 500     nlist->jjnr[nrj] = j_atom;
 501     nlist->nrj++;
 502 }
 503
 504 static gmx_inline void add_j_to_nblist_cg(t_nblist *nlist,
 505                                           atom_id j_start, int j_end,
 506                                           t_excl *bexcl, gmx_bool i_is_j,
 507                                           gmx_bool bLR)
 508 {
 509     int nrj = nlist->nrj;
 510     int j;
 511
 512     if (nlist->nrj >= nlist->maxnrj)
 513     {
 514         nlist->maxnrj = over_alloc_small(nlist->nrj + 1);
 515         if (gmx_debug_at)
 516         {
 517             fprintf(debug, "Increasing %s nblist (ielec=%d,ivdw=%d,type=%d,igeometry=%d) j size to %d\n",
 518                     bLR ? "LR" : "SR", nlist->ielec, nlist->ivdw, nlist->type, nlist->igeometry, nlist->maxnrj);
 519         }
 520
 521         srenew(nlist->jjnr, nlist->maxnrj);
 522         srenew(nlist->jjnr_end, nlist->maxnrj);
 523         srenew(nlist->excl, nlist->maxnrj*MAX_CGCGSIZE);
 524     }
 525
 526     nlist->jjnr[nrj]     = j_start;
 527     nlist->jjnr_end[nrj] = j_end;
 528
 529     if (j_end - j_start > MAX_CGCGSIZE)
 530     {
 531         gmx_fatal(FARGS, "The charge-group - charge-group neighborlist do not support charge groups larger than %d, found a charge group of size %d", MAX_CGCGSIZE, j_end-j_start);
 532     }
 533
 534     /* Set the exclusions */
 535     for (j = j_start; j < j_end; j++)
 536     {
 537         nlist->excl[nrj*MAX_CGCGSIZE + j - j_start] = bexcl[j];
 538     }
 539     if (i_is_j)
 540     {
 541         /* Avoid double counting of intra-cg interactions */
 542         for (j = 1; j < j_end-j_start; j++)
 543         {
 544             nlist->excl[nrj*MAX_CGCGSIZE + j] |= (1<<j) - 1;
 545         }
 546     }
 547
 548     nlist->nrj++;
 549 }
 550
 551 typedef void
 552     put_in_list_t (gmx_bool              bHaveVdW[],
 553                    int                   ngid,
 554                    t_mdatoms     *       md,
 555                    int                   icg,
 556                    int                   jgid,
 557                    int                   nj,
 558                    atom_id               jjcg[],
 559                    atom_id               index[],
 560                    t_excl                bExcl[],
 561                    int                   shift,
 562                    t_forcerec     *      fr,
 563                    gmx_bool              bLR,
 564                    gmx_bool              bDoVdW,
 565                    gmx_bool              bDoCoul,
 566                    int                   solvent_opt);
 567
 568 static void
 569 put_in_list_at(gmx_bool              bHaveVdW[],
 570                int                   ngid,
 571                t_mdatoms     *       md,
 572                int                   icg,
 573                int                   jgid,
 574                int                   nj,
 575                atom_id               jjcg[],
 576                atom_id               index[],
 577                t_excl                bExcl[],
 578                int                   shift,
 579                t_forcerec     *      fr,
 580                gmx_bool              bLR,
 581                gmx_bool              bDoVdW,
 582                gmx_bool              bDoCoul,
 583                int                   solvent_opt)
 584 {
 585     /* The a[] index has been removed,
 586      * to put it back in i_atom should be a[i0] and jj should be a[jj].
 587      */
 588     t_nblist  *   vdwc;
 589     t_nblist  *   vdw;
 590     t_nblist  *   coul;
 591     t_nblist  *   vdwc_free  = NULL;
 592     t_nblist  *   vdw_free   = NULL;
 593     t_nblist  *   coul_free  = NULL;
 594     t_nblist  *   vdwc_ww    = NULL;
 595     t_nblist  *   coul_ww    = NULL;
 596
 597     int           i, j, jcg, igid, gid, nbl_ind, ind_ij;
 598     atom_id       jj, jj0, jj1, i_atom;
 599     int           i0, nicg, len;
 600
 601     int          *cginfo;
 602     int          *type, *typeB;
 603     real         *charge, *chargeB;
 604     real          qi, qiB, qq, rlj;
 605     gmx_bool      bFreeEnergy, bFree, bFreeJ, bNotEx, *bPert;
 606     gmx_bool      bDoVdW_i, bDoCoul_i, bDoCoul_i_sol;
 607     int           iwater, jwater;
 608     t_nblist     *nlist;
 609
 610     /* Copy some pointers */
 611     cginfo  = fr->cginfo;
 612     charge  = md->chargeA;
 613     chargeB = md->chargeB;
 614     type    = md->typeA;
 615     typeB   = md->typeB;
 616     bPert   = md->bPerturbed;
 617
 618     /* Get atom range */
 619     i0     = index[icg];
 620     nicg   = index[icg+1]-i0;
 621
 622     /* Get the i charge group info */
 623     igid   = GET_CGINFO_GID(cginfo[icg]);
 624
 625     iwater = (solvent_opt != esolNO) ? GET_CGINFO_SOLOPT(cginfo[icg]) : esolNO;
 626
 627     bFreeEnergy = FALSE;
 628     if (md->nPerturbed)
 629     {
 630         /* Check if any of the particles involved are perturbed.
 631          * If not we can do the cheaper normal put_in_list
 632          * and use more solvent optimization.
 633          */
 634         for (i = 0; i < nicg; i++)
 635         {
 636             bFreeEnergy |= bPert[i0+i];
 637         }
 638         /* Loop over the j charge groups */
 639         for (j = 0; (j < nj && !bFreeEnergy); j++)
 640         {
 641             jcg = jjcg[j];
 642             jj0 = index[jcg];
 643             jj1 = index[jcg+1];
 644             /* Finally loop over the atoms in the j-charge group */
 645             for (jj = jj0; jj < jj1; jj++)
 646             {
 647                 bFreeEnergy |= bPert[jj];
 648             }
 649         }
 650     }
 651
 652     /* Unpack pointers to neighbourlist structs */
 653     if (fr->nnblists == 1)
 654     {
 655         nbl_ind = 0;
 656     }
 657     else
 658     {
 659         nbl_ind = fr->gid2nblists[GID(igid, jgid, ngid)];
 660     }
 661     if (bLR)
 662     {
 663         nlist = fr->nblists[nbl_ind].nlist_lr;
 664     }
 665     else
 666     {
 667         nlist = fr->nblists[nbl_ind].nlist_sr;
 668     }
 669
 670     if (iwater != esolNO)
 671     {
 672         vdwc = &nlist[eNL_VDWQQ_WATER];
 673         vdw  = &nlist[eNL_VDW];
 674         coul = &nlist[eNL_QQ_WATER];
 675 #ifndef DISABLE_WATERWATER_NLIST
 676         vdwc_ww = &nlist[eNL_VDWQQ_WATERWATER];
 677         coul_ww = &nlist[eNL_QQ_WATERWATER];
 678 #endif
 679     }
 680     else
 681     {
 682         vdwc = &nlist[eNL_VDWQQ];
 683         vdw  = &nlist[eNL_VDW];
 684         coul = &nlist[eNL_QQ];
 685     }
 686
 687     if (!bFreeEnergy)
 688     {
 689         if (iwater != esolNO)
 690         {
 691             /* Loop over the atoms in the i charge group */
 692             i_atom  = i0;
 693             gid     = GID(igid, jgid, ngid);
 694             /* Create new i_atom for each energy group */
 695             if (bDoCoul && bDoVdW)
 696             {
 697                 new_i_nblist(vdwc, i_atom, shift, gid);
 698 #ifndef DISABLE_WATERWATER_NLIST
 699                 new_i_nblist(vdwc_ww, i_atom, shift, gid);
 700 #endif
 701             }
 702             if (bDoVdW)
 703             {
 704                 new_i_nblist(vdw, i_atom, shift, gid);
 705             }
 706             if (bDoCoul)
 707             {
 708                 new_i_nblist(coul, i_atom, shift, gid);
 709 #ifndef DISABLE_WATERWATER_NLIST
 710                 new_i_nblist(coul_ww, i_atom, shift, gid);
 711 #endif
 712             }
 713             /* Loop over the j charge groups */
 714             for (j = 0; (j < nj); j++)
 715             {
 716                 jcg = jjcg[j];
 717
 718                 if (jcg == icg)
 719                 {
 720                     continue;
 721                 }
 722
 723                 jj0    = index[jcg];
 724                 jwater = GET_CGINFO_SOLOPT(cginfo[jcg]);
 725
 726                 if (iwater == esolSPC && jwater == esolSPC)
 727                 {
 728                     /* Interaction between two SPC molecules */
 729                     if (!bDoCoul)
 730                     {
 731                         /* VdW only - only first atoms in each water interact */
 732                         add_j_to_nblist(vdw, jj0, bLR);
 733                     }
 734                     else
 735                     {
 736 #ifdef DISABLE_WATERWATER_NLIST
 737                         /* Add entries for the three atoms - only do VdW if we need to */
 738                         if (!bDoVdW)
 739                         {
 740                             add_j_to_nblist(coul, jj0, bLR);
 741                         }
 742                         else
 743                         {
 744                             add_j_to_nblist(vdwc, jj0, bLR);
 745                         }
 746                         add_j_to_nblist(coul, jj0+1, bLR);
 747                         add_j_to_nblist(coul, jj0+2, bLR);
 748 #else
 749                         /* One entry for the entire water-water interaction */
 750                         if (!bDoVdW)
 751                         {
 752                             add_j_to_nblist(coul_ww, jj0, bLR);
 753                         }
 754                         else
 755                         {
 756                             add_j_to_nblist(vdwc_ww, jj0, bLR);
 757                         }
 758 #endif
 759                     }
 760                 }
 761                 else if (iwater == esolTIP4P && jwater == esolTIP4P)
 762                 {
 763                     /* Interaction between two TIP4p molecules */
 764                     if (!bDoCoul)
 765                     {
 766                         /* VdW only - only first atoms in each water interact */
 767                         add_j_to_nblist(vdw, jj0, bLR);
 768                     }
 769                     else
 770                     {
 771 #ifdef DISABLE_WATERWATER_NLIST
 772                         /* Add entries for the four atoms - only do VdW if we need to */
 773                         if (bDoVdW)
 774                         {
 775                             add_j_to_nblist(vdw, jj0, bLR);
 776                         }
 777                         add_j_to_nblist(coul, jj0+1, bLR);
 778                         add_j_to_nblist(coul, jj0+2, bLR);
 779                         add_j_to_nblist(coul, jj0+3, bLR);
 780 #else
 781                         /* One entry for the entire water-water interaction */
 782                         if (!bDoVdW)
 783                         {
 784                             add_j_to_nblist(coul_ww, jj0, bLR);
 785                         }
 786                         else
 787                         {
 788                             add_j_to_nblist(vdwc_ww, jj0, bLR);
 789                         }
 790 #endif
 791                     }
 792                 }
 793                 else
 794                 {
 795                     /* j charge group is not water, but i is.
 796                      * Add entries to the water-other_atom lists; the geometry of the water
 797                      * molecule doesn't matter - that is taken care of in the nonbonded kernel,
 798                      * so we don't care if it is SPC or TIP4P...
 799                      */
 800
 801                     jj1 = index[jcg+1];
 802
 803                     if (!bDoVdW)
 804                     {
 805                         for (jj = jj0; (jj < jj1); jj++)
 806                         {
 807                             if (charge[jj] != 0)
 808                             {
 809                                 add_j_to_nblist(coul, jj, bLR);
 810                             }
 811                         }
 812                     }
 813                     else if (!bDoCoul)
 814                     {
 815                         for (jj = jj0; (jj < jj1); jj++)
 816                         {
 817                             if (bHaveVdW[type[jj]])
 818                             {
 819                                 add_j_to_nblist(vdw, jj, bLR);
 820                             }
 821                         }
 822                     }
 823                     else
 824                     {
 825                         /* _charge_ _groups_ interact with both coulomb and LJ */
 826                         /* Check which atoms we should add to the lists!       */
 827                         for (jj = jj0; (jj < jj1); jj++)
 828                         {
 829                             if (bHaveVdW[type[jj]])
 830                             {
 831                                 if (charge[jj] != 0)
 832                                 {
 833                                     add_j_to_nblist(vdwc, jj, bLR);
 834                                 }
 835                                 else
 836                                 {
 837                                     add_j_to_nblist(vdw, jj, bLR);
 838                                 }
 839                             }
 840                             else if (charge[jj] != 0)
 841                             {
 842                                 add_j_to_nblist(coul, jj, bLR);
 843                             }
 844                         }
 845                     }
 846                 }
 847             }
 848             close_i_nblist(vdw);
 849             close_i_nblist(coul);
 850             close_i_nblist(vdwc);
 851 #ifndef DISABLE_WATERWATER_NLIST
 852             close_i_nblist(coul_ww);
 853             close_i_nblist(vdwc_ww);
 854 #endif
 855         }
 856         else
 857         {
 858             /* no solvent as i charge group */
 859             /* Loop over the atoms in the i charge group */
 860             for (i = 0; i < nicg; i++)
 861             {
 862                 i_atom  = i0+i;
 863                 gid     = GID(igid, jgid, ngid);
 864                 qi      = charge[i_atom];
 865
 866                 /* Create new i_atom for each energy group */
 867                 if (bDoVdW && bDoCoul)
 868                 {
 869                     new_i_nblist(vdwc, i_atom, shift, gid);
 870                 }
 871                 if (bDoVdW)
 872                 {
 873                     new_i_nblist(vdw, i_atom, shift, gid);
 874                 }
 875                 if (bDoCoul)
 876                 {
 877                     new_i_nblist(coul, i_atom, shift, gid);
 878                 }
 879                 bDoVdW_i  = (bDoVdW  && bHaveVdW[type[i_atom]]);
 880                 bDoCoul_i = (bDoCoul && qi != 0);
 881
 882                 if (bDoVdW_i || bDoCoul_i)
 883                 {
 884                     /* Loop over the j charge groups */
 885                     for (j = 0; (j < nj); j++)
 886                     {
 887                         jcg = jjcg[j];
 888
 889                         /* Check for large charge groups */
 890                         if (jcg == icg)
 891                         {
 892                             jj0 = i0 + i + 1;
 893                         }
 894                         else
 895                         {
 896                             jj0 = index[jcg];
 897                         }
 898
 899                         jj1 = index[jcg+1];
 900                         /* Finally loop over the atoms in the j-charge group */
 901                         for (jj = jj0; jj < jj1; jj++)
 902                         {
 903                             bNotEx = NOTEXCL(bExcl, i, jj);
 904
 905                             if (bNotEx)
 906                             {
 907                                 if (!bDoVdW_i)
 908                                 {
 909                                     if (charge[jj] != 0)
 910                                     {
 911                                         add_j_to_nblist(coul, jj, bLR);
 912                                     }
 913                                 }
 914                                 else if (!bDoCoul_i)
 915                                 {
 916                                     if (bHaveVdW[type[jj]])
 917                                     {
 918                                         add_j_to_nblist(vdw, jj, bLR);
 919                                     }
 920                                 }
 921                                 else
 922                                 {
 923                                     if (bHaveVdW[type[jj]])
 924                                     {
 925                                         if (charge[jj] != 0)
 926                                         {
 927                                             add_j_to_nblist(vdwc, jj, bLR);
 928                                         }
 929                                         else
 930                                         {
 931                                             add_j_to_nblist(vdw, jj, bLR);
 932                                         }
 933                                     }
 934                                     else if (charge[jj] != 0)
 935                                     {
 936                                         add_j_to_nblist(coul, jj, bLR);
 937                                     }
 938                                 }
 939                             }
 940                         }
 941                     }
 942                 }
 943                 close_i_nblist(vdw);
 944                 close_i_nblist(coul);
 945                 close_i_nblist(vdwc);
 946             }
 947         }
 948     }
 949     else
 950     {
 951         /* we are doing free energy */
 952         vdwc_free = &nlist[eNL_VDWQQ_FREE];
 953         vdw_free  = &nlist[eNL_VDW_FREE];
 954         coul_free = &nlist[eNL_QQ_FREE];
 955         /* Loop over the atoms in the i charge group */
 956         for (i = 0; i < nicg; i++)
 957         {
 958             i_atom  = i0+i;
 959             gid     = GID(igid, jgid, ngid);
 960             qi      = charge[i_atom];
 961             qiB     = chargeB[i_atom];
 962
 963             /* Create new i_atom for each energy group */
 964             if (bDoVdW && bDoCoul)
 965             {
 966                 new_i_nblist(vdwc, i_atom, shift, gid);
 967             }
 968             if (bDoVdW)
 969             {
 970                 new_i_nblist(vdw, i_atom, shift, gid);
 971             }
 972             if (bDoCoul)
 973             {
 974                 new_i_nblist(coul, i_atom, shift, gid);
 975             }
 976
 977             new_i_nblist(vdw_free, i_atom, shift, gid);
 978             new_i_nblist(coul_free, i_atom, shift, gid);
 979             new_i_nblist(vdwc_free, i_atom, shift, gid);
 980
 981             bDoVdW_i  = (bDoVdW  &&
 982                          (bHaveVdW[type[i_atom]] || bHaveVdW[typeB[i_atom]]));
 983             bDoCoul_i = (bDoCoul && (qi != 0 || qiB != 0));
 984             /* For TIP4P the first atom does not have a charge,
 985              * but the last three do. So we should still put an atom
 986              * without LJ but with charge in the water-atom neighborlist
 987              * for a TIP4p i charge group.
 988              * For SPC type water the first atom has LJ and charge,
 989              * so there is no such problem.
 990              */
 991             if (iwater == esolNO)
 992             {
 993                 bDoCoul_i_sol = bDoCoul_i;
 994             }
 995             else
 996             {
 997                 bDoCoul_i_sol = bDoCoul;
 998             }
 999
1000             if (bDoVdW_i || bDoCoul_i_sol)
1001             {
1002                 /* Loop over the j charge groups */
1003                 for (j = 0; (j < nj); j++)
1004                 {
1005                     jcg = jjcg[j];
1006
1007                     /* Check for large charge groups */
1008                     if (jcg == icg)
1009                     {
1010                         jj0 = i0 + i + 1;
1011                     }
1012                     else
1013                     {
1014                         jj0 = index[jcg];
1015                     }
1016
1017                     jj1 = index[jcg+1];
1018                     /* Finally loop over the atoms in the j-charge group */
1019                     bFree = bPert[i_atom];
1020                     for (jj = jj0; (jj < jj1); jj++)
1021                     {
1022                         bFreeJ = bFree || bPert[jj];
1023                         /* Complicated if, because the water H's should also
1024                          * see perturbed j-particles
1025                          */
1026                         if (iwater == esolNO || i == 0 || bFreeJ)
1027                         {
1028                             bNotEx = NOTEXCL(bExcl, i, jj);
1029
1030                             if (bNotEx)
1031                             {
1032                                 if (bFreeJ)
1033                                 {
1034                                     if (!bDoVdW_i)
1035                                     {
1036                                         if (charge[jj] != 0 || chargeB[jj] != 0)
1037                                         {
1038                                             add_j_to_nblist(coul_free, jj, bLR);
1039                                         }
1040                                     }
1041                                     else if (!bDoCoul_i)
1042                                     {
1043                                         if (bHaveVdW[type[jj]] || bHaveVdW[typeB[jj]])
1044                                         {
1045                                             add_j_to_nblist(vdw_free, jj, bLR);
1046                                         }
1047                                     }
1048                                     else
1049                                     {
1050                                         if (bHaveVdW[type[jj]] || bHaveVdW[typeB[jj]])
1051                                         {
1052                                             if (charge[jj] != 0 || chargeB[jj] != 0)
1053                                             {
1054                                                 add_j_to_nblist(vdwc_free, jj, bLR);
1055                                             }
1056                                             else
1057                                             {
1058                                                 add_j_to_nblist(vdw_free, jj, bLR);
1059                                             }
1060                                         }
1061                                         else if (charge[jj] != 0 || chargeB[jj] != 0)
1062                                         {
1063                                             add_j_to_nblist(coul_free, jj, bLR);
1064                                         }
1065                                     }
1066                                 }
1067                                 else if (!bDoVdW_i)
1068                                 {
1069                                     /* This is done whether or not bWater is set */
1070                                     if (charge[jj] != 0)
1071                                     {
1072                                         add_j_to_nblist(coul, jj, bLR);
1073                                     }
1074                                 }
1075                                 else if (!bDoCoul_i_sol)
1076                                 {
1077                                     if (bHaveVdW[type[jj]])
1078                                     {
1079                                         add_j_to_nblist(vdw, jj, bLR);
1080                                     }
1081                                 }
1082                                 else
1083                                 {
1084                                     if (bHaveVdW[type[jj]])
1085                                     {
1086                                         if (charge[jj] != 0)
1087                                         {
1088                                             add_j_to_nblist(vdwc, jj, bLR);
1089                                         }
1090                                         else
1091                                         {
1092                                             add_j_to_nblist(vdw, jj, bLR);
1093                                         }
1094                                     }
1095                                     else if (charge[jj] != 0)
1096                                     {
1097                                         add_j_to_nblist(coul, jj, bLR);
1098                                     }
1099                                 }
1100                             }
1101                         }
1102                     }
1103                 }
1104             }
1105             close_i_nblist(vdw);
1106             close_i_nblist(coul);
1107             close_i_nblist(vdwc);
1108             close_i_nblist(vdw_free);
1109             close_i_nblist(coul_free);
1110             close_i_nblist(vdwc_free);
1111         }
1112     }
1113 }
1114
1115 static void
1116 put_in_list_adress(gmx_bool              bHaveVdW[],
1117                    int                   ngid,
1118                    t_mdatoms     *       md,
1119                    int                   icg,
1120                    int                   jgid,
1121                    int                   nj,
1122                    atom_id               jjcg[],
1123                    atom_id               index[],
1124                    t_excl                bExcl[],
1125                    int                   shift,
1126                    t_forcerec     *      fr,
1127                    gmx_bool              bLR,
1128                    gmx_bool              bDoVdW,
1129                    gmx_bool              bDoCoul,
1130                    int                   solvent_opt)
1131 {
1132     /* The a[] index has been removed,
1133      * to put it back in i_atom should be a[i0] and jj should be a[jj].
1134      */
1135     t_nblist  *   vdwc;
1136     t_nblist  *   vdw;
1137     t_nblist  *   coul;
1138     t_nblist  *   vdwc_adress  = NULL;
1139     t_nblist  *   vdw_adress   = NULL;
1140     t_nblist  *   coul_adress  = NULL;
1141     t_nblist  *   vdwc_ww      = NULL;
1142     t_nblist  *   coul_ww      = NULL;
1143
1144     int           i, j, jcg, igid, gid, nbl_ind, nbl_ind_adress;
1145     atom_id       jj, jj0, jj1, i_atom;
1146     int           i0, nicg, len;
1147
1148     int          *cginfo;
1149     int          *type, *typeB;
1150     real         *charge, *chargeB;
1151     real         *wf;
1152     real          qi, qiB, qq, rlj;
1153     gmx_bool      bFreeEnergy, bFree, bFreeJ, bNotEx, *bPert;
1154     gmx_bool      bDoVdW_i, bDoCoul_i, bDoCoul_i_sol;
1155     gmx_bool      b_hybrid;
1156     gmx_bool      j_all_atom;
1157     int           iwater, jwater;
1158     t_nblist     *nlist, *nlist_adress;
1159     gmx_bool      bEnergyGroupCG;
1160
1161     /* Copy some pointers */
1162     cginfo  = fr->cginfo;
1163     charge  = md->chargeA;
1164     chargeB = md->chargeB;
1165     type    = md->typeA;
1166     typeB   = md->typeB;
1167     bPert   = md->bPerturbed;
1168     wf      = md->wf;
1169
1170     /* Get atom range */
1171     i0     = index[icg];
1172     nicg   = index[icg+1]-i0;
1173
1174     /* Get the i charge group info */
1175     igid   = GET_CGINFO_GID(cginfo[icg]);
1176
1177     iwater = (solvent_opt != esolNO) ? GET_CGINFO_SOLOPT(cginfo[icg]) : esolNO;
1178
1179     if (md->nPerturbed)
1180     {
1181         gmx_fatal(FARGS, "AdResS does not support free energy pertubation\n");
1182     }
1183
1184     /* Unpack pointers to neighbourlist structs */
1185     if (fr->nnblists == 2)
1186     {
1187         nbl_ind        = 0;
1188         nbl_ind_adress = 1;
1189     }
1190     else
1191     {
1192         nbl_ind        = fr->gid2nblists[GID(igid, jgid, ngid)];
1193         nbl_ind_adress = nbl_ind+fr->nnblists/2;
1194     }
1195     if (bLR)
1196     {
1197         nlist        = fr->nblists[nbl_ind].nlist_lr;
1198         nlist_adress = fr->nblists[nbl_ind_adress].nlist_lr;
1199     }
1200     else
1201     {
1202         nlist        = fr->nblists[nbl_ind].nlist_sr;
1203         nlist_adress = fr->nblists[nbl_ind_adress].nlist_sr;
1204     }
1205
1206
1207     vdwc = &nlist[eNL_VDWQQ];
1208     vdw  = &nlist[eNL_VDW];
1209     coul = &nlist[eNL_QQ];
1210
1211     vdwc_adress = &nlist_adress[eNL_VDWQQ];
1212     vdw_adress  = &nlist_adress[eNL_VDW];
1213     coul_adress = &nlist_adress[eNL_QQ];
1214
1215     /* We do not support solvent optimization with AdResS for now.
1216        For this we would need hybrid solvent-other kernels */
1217
1218     /* no solvent as i charge group */
1219     /* Loop over the atoms in the i charge group */
1220     for (i = 0; i < nicg; i++)
1221     {
1222         i_atom  = i0+i;
1223         gid     = GID(igid, jgid, ngid);
1224         qi      = charge[i_atom];
1225
1226         /* Create new i_atom for each energy group */
1227         if (bDoVdW && bDoCoul)
1228         {
1229             new_i_nblist(vdwc, i_atom, shift, gid);
1230             new_i_nblist(vdwc_adress, i_atom, shift, gid);
1231
1232         }
1233         if (bDoVdW)
1234         {
1235             new_i_nblist(vdw, i_atom, shift, gid);
1236             new_i_nblist(vdw_adress, i_atom, shift, gid);
1237
1238         }
1239         if (bDoCoul)
1240         {
1241             new_i_nblist(coul, i_atom, shift, gid);
1242             new_i_nblist(coul_adress, i_atom, shift, gid);
1243         }
1244         bDoVdW_i  = (bDoVdW  && bHaveVdW[type[i_atom]]);
1245         bDoCoul_i = (bDoCoul && qi != 0);
1246
1247         /* Here we find out whether the energy groups interaction belong to a
1248          * coarse-grained (vsite) or atomistic interaction. Note that, beacuse
1249          * interactions between coarse-grained and other (atomistic) energygroups
1250          * are excluded automatically by grompp, it is sufficient to check for
1251          * the group id of atom i (igid) */
1252         bEnergyGroupCG = !egp_explicit(fr, igid);
1253
1254         if (bDoVdW_i || bDoCoul_i)
1255         {
1256             /* Loop over the j charge groups */
1257             for (j = 0; (j < nj); j++)
1258             {
1259                 jcg = jjcg[j];
1260
1261                 /* Check for large charge groups */
1262                 if (jcg == icg)
1263                 {
1264                     jj0 = i0 + i + 1;
1265                 }
1266                 else
1267                 {
1268                     jj0 = index[jcg];
1269                 }
1270
1271                 jj1 = index[jcg+1];
1272                 /* Finally loop over the atoms in the j-charge group */
1273                 for (jj = jj0; jj < jj1; jj++)
1274                 {
1275                     bNotEx = NOTEXCL(bExcl, i, jj);
1276
1277                     /* Now we have to exclude interactions which will be zero
1278                      * anyway due to the AdResS weights (in previous implementations
1279                      * this was done in the force kernel). This is necessary as
1280                      * pure interactions (those with b_hybrid=false, i.e. w_i*w_j==1 or 0)
1281                      * are put into neighbour lists which will be passed to the
1282                      * standard (optimized) kernels for speed. The interactions with
1283                      * b_hybrid=true are placed into the _adress neighbour lists and
1284                      * processed by the generic AdResS kernel.
1285                      */
1286                     if ( (bEnergyGroupCG &&
1287                           wf[i_atom] >= 1-GMX_REAL_EPS && wf[jj] >= 1-GMX_REAL_EPS ) ||
1288                          ( !bEnergyGroupCG && wf[jj] <= GMX_REAL_EPS ) )
1289                     {
1290                         continue;
1291                     }
1292
1293                     b_hybrid = !((wf[i_atom] >= 1-GMX_REAL_EPS && wf[jj] >= 1-GMX_REAL_EPS) ||
1294                                  (wf[i_atom] <= GMX_REAL_EPS && wf[jj] <= GMX_REAL_EPS));
1295
1296                     if (bNotEx)
1297                     {
1298                         if (!bDoVdW_i)
1299                         {
1300                             if (charge[jj] != 0)
1301                             {
1302                                 if (!b_hybrid)
1303                                 {
1304                                     add_j_to_nblist(coul, jj, bLR);
1305                                 }
1306                                 else
1307                                 {
1308                                     add_j_to_nblist(coul_adress, jj, bLR);
1309                                 }
1310                             }
1311                         }
1312                         else if (!bDoCoul_i)
1313                         {
1314                             if (bHaveVdW[type[jj]])
1315                             {
1316                                 if (!b_hybrid)
1317                                 {
1318                                     add_j_to_nblist(vdw, jj, bLR);
1319                                 }
1320                                 else
1321                                 {
1322                                     add_j_to_nblist(vdw_adress, jj, bLR);
1323                                 }
1324                             }
1325                         }
1326                         else
1327                         {
1328                             if (bHaveVdW[type[jj]])
1329                             {
1330                                 if (charge[jj] != 0)
1331                                 {
1332                                     if (!b_hybrid)
1333                                     {
1334                                         add_j_to_nblist(vdwc, jj, bLR);
1335                                     }
1336                                     else
1337                                     {
1338                                         add_j_to_nblist(vdwc_adress, jj, bLR);
1339                                     }
1340                                 }
1341                                 else
1342                                 {
1343                                     if (!b_hybrid)
1344                                     {
1345                                         add_j_to_nblist(vdw, jj, bLR);
1346                                     }
1347                                     else
1348                                     {
1349                                         add_j_to_nblist(vdw_adress, jj, bLR);
1350                                     }
1351
1352                                 }
1353                             }
1354                             else if (charge[jj] != 0)
1355                             {
1356                                 if (!b_hybrid)
1357                                 {
1358                                     add_j_to_nblist(coul, jj, bLR);
1359                                 }
1360                                 else
1361                                 {
1362                                     add_j_to_nblist(coul_adress, jj, bLR);
1363                                 }
1364
1365                             }
1366                         }
1367                     }
1368                 }
1369             }
1370
1371             close_i_nblist(vdw);
1372             close_i_nblist(coul);
1373             close_i_nblist(vdwc);
1374             close_i_nblist(vdw_adress);
1375             close_i_nblist(coul_adress);
1376             close_i_nblist(vdwc_adress);
1377         }
1378     }
1379 }
1380
1381 static void
1382 put_in_list_qmmm(gmx_bool gmx_unused              bHaveVdW[],
1383                  int                              ngid,
1384                  t_mdatoms gmx_unused     *       md,
1385                  int                              icg,
1386                  int                              jgid,
1387                  int                              nj,
1388                  atom_id                          jjcg[],
1389                  atom_id                          index[],
1390                  t_excl                           bExcl[],
1391                  int                              shift,
1392                  t_forcerec                *      fr,
1393                  gmx_bool                         bLR,
1394                  gmx_bool  gmx_unused             bDoVdW,
1395                  gmx_bool  gmx_unused             bDoCoul,
1396                  int       gmx_unused             solvent_opt)
1397 {
1398     t_nblist  *   coul;
1399     int           i, j, jcg, igid, gid;
1400     atom_id       jj, jj0, jj1, i_atom;
1401     int           i0, nicg;
1402     gmx_bool      bNotEx;
1403
1404     /* Get atom range */
1405     i0     = index[icg];
1406     nicg   = index[icg+1]-i0;
1407
1408     /* Get the i charge group info */
1409     igid   = GET_CGINFO_GID(fr->cginfo[icg]);
1410
1411     coul = &fr->QMMMlist;
1412
1413     /* Loop over atoms in the ith charge group */
1414     for (i = 0; i < nicg; i++)
1415     {
1416         i_atom = i0+i;
1417         gid    = GID(igid, jgid, ngid);
1418         /* Create new i_atom for each energy group */
1419         new_i_nblist(coul, i_atom, shift, gid);
1420
1421         /* Loop over the j charge groups */
1422         for (j = 0; j < nj; j++)
1423         {
1424             jcg = jjcg[j];
1425
1426             /* Charge groups cannot have QM and MM atoms simultaneously */
1427             if (jcg != icg)
1428             {
1429                 jj0 = index[jcg];
1430                 jj1 = index[jcg+1];
1431                 /* Finally loop over the atoms in the j-charge group */
1432                 for (jj = jj0; jj < jj1; jj++)
1433                 {
1434                     bNotEx = NOTEXCL(bExcl, i, jj);
1435                     if (bNotEx)
1436                     {
1437                         add_j_to_nblist(coul, jj, bLR);
1438                     }
1439                 }
1440             }
1441         }
1442         close_i_nblist(coul);
1443     }
1444 }
1445
1446 static void
1447 put_in_list_cg(gmx_bool  gmx_unused             bHaveVdW[],
1448                int                              ngid,
1449                t_mdatoms  gmx_unused    *       md,
1450                int                              icg,
1451                int                              jgid,
1452                int                              nj,
1453                atom_id                          jjcg[],
1454                atom_id                          index[],
1455                t_excl                           bExcl[],
1456                int                              shift,
1457                t_forcerec                *      fr,
1458                gmx_bool                         bLR,
1459                gmx_bool   gmx_unused            bDoVdW,
1460                gmx_bool   gmx_unused            bDoCoul,
1461                int        gmx_unused            solvent_opt)
1462 {
1463     int          cginfo;
1464     int          igid, gid, nbl_ind;
1465     t_nblist *   vdwc;
1466     int          j, jcg;
1467
1468     cginfo = fr->cginfo[icg];
1469
1470     igid = GET_CGINFO_GID(cginfo);
1471     gid  = GID(igid, jgid, ngid);
1472
1473     /* Unpack pointers to neighbourlist structs */
1474     if (fr->nnblists == 1)
1475     {
1476         nbl_ind = 0;
1477     }
1478     else
1479     {
1480         nbl_ind = fr->gid2nblists[gid];
1481     }
1482     if (bLR)
1483     {
1484         vdwc = &fr->nblists[nbl_ind].nlist_lr[eNL_VDWQQ];
1485     }
1486     else
1487     {
1488         vdwc = &fr->nblists[nbl_ind].nlist_sr[eNL_VDWQQ];
1489     }
1490
1491     /* Make a new neighbor list for charge group icg.
1492      * Currently simply one neighbor list is made with LJ and Coulomb.
1493      * If required, zero interactions could be removed here
1494      * or in the force loop.
1495      */
1496     new_i_nblist(vdwc, index[icg], shift, gid);
1497     vdwc->iinr_end[vdwc->nri] = index[icg+1];
1498
1499     for (j = 0; (j < nj); j++)
1500     {
1501         jcg = jjcg[j];
1502         /* Skip the icg-icg pairs if all self interactions are excluded */
1503         if (!(jcg == icg && GET_CGINFO_EXCL_INTRA(cginfo)))
1504         {
1505             /* Here we add the j charge group jcg to the list,
1506              * exclusions are also added to the list.
1507              */
1508             add_j_to_nblist_cg(vdwc, index[jcg], index[jcg+1], bExcl, icg == jcg, bLR);
1509         }
1510     }
1511
1512     close_i_nblist(vdwc);
1513 }
1514
1515 static void setexcl(atom_id start, atom_id end, t_blocka *excl, gmx_bool b,
1516                     t_excl bexcl[])
1517 {
1518     atom_id i, k;
1519
1520     if (b)
1521     {
1522         for (i = start; i < end; i++)
1523         {
1524             for (k = excl->index[i]; k < excl->index[i+1]; k++)
1525             {
1526                 SETEXCL(bexcl, i-start, excl->a[k]);
1527             }
1528         }
1529     }
1530     else
1531     {
1532         for (i = start; i < end; i++)
1533         {
1534             for (k = excl->index[i]; k < excl->index[i+1]; k++)
1535             {
1536                 RMEXCL(bexcl, i-start, excl->a[k]);
1537             }
1538         }
1539     }
1540 }
1541
1542 int calc_naaj(int icg, int cgtot)
1543 {
1544     int naaj;
1545
1546     if ((cgtot % 2) == 1)
1547     {
1548         /* Odd number of charge groups, easy */
1549         naaj = 1 + (cgtot/2);
1550     }
1551     else if ((cgtot % 4) == 0)
1552     {
1553         /* Multiple of four is hard */
1554         if (icg < cgtot/2)
1555         {
1556             if ((icg % 2) == 0)
1557             {
1558                 naaj = 1+(cgtot/2);
1559             }
1560             else
1561             {
1562                 naaj = cgtot/2;
1563             }
1564         }
1565         else
1566         {
1567             if ((icg % 2) == 1)
1568             {
1569                 naaj = 1+(cgtot/2);
1570             }
1571             else
1572             {
1573                 naaj = cgtot/2;
1574             }
1575         }
1576     }
1577     else
1578     {
1579         /* cgtot/2 = odd */
1580         if ((icg % 2) == 0)
1581         {
1582             naaj = 1+(cgtot/2);
1583         }
1584         else
1585         {
1586             naaj = cgtot/2;
1587         }
1588     }
1589 #ifdef DEBUG
1590     fprintf(log, "naaj=%d\n", naaj);
1591 #endif
1592
1593     return naaj;
1594 }
1595
1596 /************************************************
1597  *
1598  *  S I M P L E      C O R E     S T U F F
1599  *
1600  ************************************************/
1601
1602 static real calc_image_tric(rvec xi, rvec xj, matrix box,
1603                             rvec b_inv, int *shift)
1604 {
1605     /* This code assumes that the cut-off is smaller than
1606      * a half times the smallest diagonal element of the box.
1607      */
1608     const real h25 = 2.5;
1609     real       dx, dy, dz;
1610     real       r2;
1611     int        tx, ty, tz;
1612
1613     /* Compute diff vector */
1614     dz = xj[ZZ] - xi[ZZ];
1615     dy = xj[YY] - xi[YY];
1616     dx = xj[XX] - xi[XX];
1617
1618     /* Perform NINT operation, using trunc operation, therefore
1619      * we first add 2.5 then subtract 2 again
1620      */
1621     tz  = dz*b_inv[ZZ] + h25;
1622     tz -= 2;
1623     dz -= tz*box[ZZ][ZZ];
1624     dy -= tz*box[ZZ][YY];
1625     dx -= tz*box[ZZ][XX];
1626
1627     ty  = dy*b_inv[YY] + h25;
1628     ty -= 2;
1629     dy -= ty*box[YY][YY];
1630     dx -= ty*box[YY][XX];
1631
1632     tx  = dx*b_inv[XX]+h25;
1633     tx -= 2;
1634     dx -= tx*box[XX][XX];
1635
1636     /* Distance squared */
1637     r2 = (dx*dx) + (dy*dy) + (dz*dz);
1638
1639     *shift = XYZ2IS(tx, ty, tz);
1640
1641     return r2;
1642 }
1643
1644 static real calc_image_rect(rvec xi, rvec xj, rvec box_size,
1645                             rvec b_inv, int *shift)
1646 {
1647     const real h15 = 1.5;
1648     real       ddx, ddy, ddz;
1649     real       dx, dy, dz;
1650     real       r2;
1651     int        tx, ty, tz;
1652
1653     /* Compute diff vector */
1654     dx = xj[XX] - xi[XX];
1655     dy = xj[YY] - xi[YY];
1656     dz = xj[ZZ] - xi[ZZ];
1657
1658     /* Perform NINT operation, using trunc operation, therefore
1659      * we first add 1.5 then subtract 1 again
1660      */
1661     tx = dx*b_inv[XX] + h15;
1662     ty = dy*b_inv[YY] + h15;
1663     tz = dz*b_inv[ZZ] + h15;
1664     tx--;
1665     ty--;
1666     tz--;
1667
1668     /* Correct diff vector for translation */
1669     ddx = tx*box_size[XX] - dx;
1670     ddy = ty*box_size[YY] - dy;
1671     ddz = tz*box_size[ZZ] - dz;
1672
1673     /* Distance squared */
1674     r2 = (ddx*ddx) + (ddy*ddy) + (ddz*ddz);
1675
1676     *shift = XYZ2IS(tx, ty, tz);
1677
1678     return r2;
1679 }
1680
1681 static void add_simple(t_ns_buf *nsbuf, int nrj, atom_id cg_j,
1682                        gmx_bool bHaveVdW[], int ngid, t_mdatoms *md,
1683                        int icg, int jgid, t_block *cgs, t_excl bexcl[],
1684                        int shift, t_forcerec *fr, put_in_list_t *put_in_list)
1685 {
1686     if (nsbuf->nj + nrj > MAX_CG)
1687     {
1688         put_in_list(bHaveVdW, ngid, md, icg, jgid, nsbuf->ncg, nsbuf->jcg,
1689                     cgs->index, bexcl, shift, fr, FALSE, TRUE, TRUE, fr->solvent_opt);
1690         /* Reset buffer contents */
1691         nsbuf->ncg = nsbuf->nj = 0;
1692     }
1693     nsbuf->jcg[nsbuf->ncg++] = cg_j;
1694     nsbuf->nj               += nrj;
1695 }
1696
1697 static void ns_inner_tric(rvec x[], int icg, int *i_egp_flags,
1698                           int njcg, atom_id jcg[],
1699                           matrix box, rvec b_inv, real rcut2,
1700                           t_block *cgs, t_ns_buf **ns_buf,
1701                           gmx_bool bHaveVdW[], int ngid, t_mdatoms *md,
1702                           t_excl bexcl[], t_forcerec *fr,
1703                           put_in_list_t *put_in_list)
1704 {
1705     int       shift;
1706     int       j, nrj, jgid;
1707     int      *cginfo = fr->cginfo;
1708     atom_id   cg_j, *cgindex;
1709     t_ns_buf *nsbuf;
1710
1711     cgindex = cgs->index;
1712     shift   = CENTRAL;
1713     for (j = 0; (j < njcg); j++)
1714     {
1715         cg_j   = jcg[j];
1716         nrj    = cgindex[cg_j+1]-cgindex[cg_j];
1717         if (calc_image_tric(x[icg], x[cg_j], box, b_inv, &shift) < rcut2)
1718         {
1719             jgid  = GET_CGINFO_GID(cginfo[cg_j]);
1720             if (!(i_egp_flags[jgid] & EGP_EXCL))
1721             {
1722                 add_simple(&ns_buf[jgid][shift], nrj, cg_j,
1723                            bHaveVdW, ngid, md, icg, jgid, cgs, bexcl, shift, fr,
1724                            put_in_list);
1725             }
1726         }
1727     }
1728 }
1729
1730 static void ns_inner_rect(rvec x[], int icg, int *i_egp_flags,
1731                           int njcg, atom_id jcg[],
1732                           gmx_bool bBox, rvec box_size, rvec b_inv, real rcut2,
1733                           t_block *cgs, t_ns_buf **ns_buf,
1734                           gmx_bool bHaveVdW[], int ngid, t_mdatoms *md,
1735                           t_excl bexcl[], t_forcerec *fr,
1736                           put_in_list_t *put_in_list)
1737 {
1738     int       shift;
1739     int       j, nrj, jgid;
1740     int      *cginfo = fr->cginfo;
1741     atom_id   cg_j, *cgindex;
1742     t_ns_buf *nsbuf;
1743
1744     cgindex = cgs->index;
1745     if (bBox)
1746     {
1747         shift = CENTRAL;
1748         for (j = 0; (j < njcg); j++)
1749         {
1750             cg_j   = jcg[j];
1751             nrj    = cgindex[cg_j+1]-cgindex[cg_j];
1752             if (calc_image_rect(x[icg], x[cg_j], box_size, b_inv, &shift) < rcut2)
1753             {
1754                 jgid  = GET_CGINFO_GID(cginfo[cg_j]);
1755                 if (!(i_egp_flags[jgid] & EGP_EXCL))
1756                 {
1757                     add_simple(&ns_buf[jgid][shift], nrj, cg_j,
1758                                bHaveVdW, ngid, md, icg, jgid, cgs, bexcl, shift, fr,
1759                                put_in_list);
1760                 }
1761             }
1762         }
1763     }
1764     else
1765     {
1766         for (j = 0; (j < njcg); j++)
1767         {
1768             cg_j   = jcg[j];
1769             nrj    = cgindex[cg_j+1]-cgindex[cg_j];
1770             if ((rcut2 == 0) || (distance2(x[icg], x[cg_j]) < rcut2))
1771             {
1772                 jgid  = GET_CGINFO_GID(cginfo[cg_j]);
1773                 if (!(i_egp_flags[jgid] & EGP_EXCL))
1774                 {
1775                     add_simple(&ns_buf[jgid][CENTRAL], nrj, cg_j,
1776                                bHaveVdW, ngid, md, icg, jgid, cgs, bexcl, CENTRAL, fr,
1777                                put_in_list);
1778                 }
1779             }
1780         }
1781     }
1782 }
1783
1784 /* ns_simple_core needs to be adapted for QMMM still 2005 */
1785
1786 static int ns_simple_core(t_forcerec *fr,
1787                           gmx_localtop_t *top,
1788                           t_mdatoms *md,
1789                           matrix box, rvec box_size,
1790                           t_excl bexcl[], atom_id *aaj,
1791                           int ngid, t_ns_buf **ns_buf,
1792                           put_in_list_t *put_in_list, gmx_bool bHaveVdW[])
1793 {
1794     int          naaj, k;
1795     real         rlist2;
1796     int          nsearch, icg, jcg, igid, i0, nri, nn;
1797     int         *cginfo;
1798     t_ns_buf    *nsbuf;
1799     /* atom_id  *i_atoms; */
1800     t_block     *cgs  = &(top->cgs);
1801     t_blocka    *excl = &(top->excls);
1802     rvec         b_inv;
1803     int          m;
1804     gmx_bool     bBox, bTriclinic;
1805     int         *i_egp_flags;
1806
1807     rlist2 = sqr(fr->rlist);
1808
1809     bBox = (fr->ePBC != epbcNONE);
1810     if (bBox)
1811     {
1812         for (m = 0; (m < DIM); m++)
1813         {
1814             b_inv[m] = divide_err(1.0, box_size[m]);
1815         }
1816         bTriclinic = TRICLINIC(box);
1817     }
1818     else
1819     {
1820         bTriclinic = FALSE;
1821     }
1822
1823     cginfo = fr->cginfo;
1824
1825     nsearch = 0;
1826     for (icg = fr->cg0; (icg < fr->hcg); icg++)
1827     {
1828         /*
1829            i0        = cgs->index[icg];
1830            nri       = cgs->index[icg+1]-i0;
1831            i_atoms   = &(cgs->a[i0]);
1832            i_eg_excl = fr->eg_excl + ngid*md->cENER[*i_atoms];
1833            setexcl(nri,i_atoms,excl,TRUE,bexcl);
1834          */
1835         igid        = GET_CGINFO_GID(cginfo[icg]);
1836         i_egp_flags = fr->egp_flags + ngid*igid;
1837         setexcl(cgs->index[icg], cgs->index[icg+1], excl, TRUE, bexcl);
1838
1839         naaj = calc_naaj(icg, cgs->nr);
1840         if (bTriclinic)
1841         {
1842             ns_inner_tric(fr->cg_cm, icg, i_egp_flags, naaj, &(aaj[icg]),
1843                           box, b_inv, rlist2, cgs, ns_buf,
1844                           bHaveVdW, ngid, md, bexcl, fr, put_in_list);
1845         }
1846         else
1847         {
1848             ns_inner_rect(fr->cg_cm, icg, i_egp_flags, naaj, &(aaj[icg]),
1849                           bBox, box_size, b_inv, rlist2, cgs, ns_buf,
1850                           bHaveVdW, ngid, md, bexcl, fr, put_in_list);
1851         }
1852         nsearch += naaj;
1853
1854         for (nn = 0; (nn < ngid); nn++)
1855         {
1856             for (k = 0; (k < SHIFTS); k++)
1857             {
1858                 nsbuf = &(ns_buf[nn][k]);
1859                 if (nsbuf->ncg > 0)
1860                 {
1861                     put_in_list(bHaveVdW, ngid, md, icg, nn, nsbuf->ncg, nsbuf->jcg,
1862                                 cgs->index, bexcl, k, fr, FALSE, TRUE, TRUE, fr->solvent_opt);
1863                     nsbuf->ncg = nsbuf->nj = 0;
1864                 }
1865             }
1866         }
1867         /* setexcl(nri,i_atoms,excl,FALSE,bexcl); */
1868         setexcl(cgs->index[icg], cgs->index[icg+1], excl, FALSE, bexcl);
1869     }
1870     close_neighbor_lists(fr, FALSE);
1871
1872     return nsearch;
1873 }
1874
1875 /************************************************
1876  *
1877  *    N S 5     G R I D     S T U F F
1878  *
1879  ************************************************/
1880
1881 static gmx_inline void get_dx(int Nx, real gridx, real rc2, int xgi, real x,
1882                               int *dx0, int *dx1, real *dcx2)
1883 {
1884     real dcx, tmp;
1885     int  xgi0, xgi1, i;
1886
1887     if (xgi < 0)
1888     {
1889         *dx0 = 0;
1890         xgi0 = -1;
1891         *dx1 = -1;
1892         xgi1 = 0;
1893     }
1894     else if (xgi >= Nx)
1895     {
1896         *dx0 = Nx;
1897         xgi0 = Nx-1;
1898         *dx1 = Nx-1;
1899         xgi1 = Nx;
1900     }
1901     else
1902     {
1903         dcx2[xgi] = 0;
1904         *dx0      = xgi;
1905         xgi0      = xgi-1;
1906         *dx1      = xgi;
1907         xgi1      = xgi+1;
1908     }
1909
1910     for (i = xgi0; i >= 0; i--)
1911     {
1912         dcx = (i+1)*gridx-x;
1913         tmp = dcx*dcx;
1914         if (tmp >= rc2)
1915         {
1916             break;
1917         }
1918         *dx0    = i;
1919         dcx2[i] = tmp;
1920     }
1921     for (i = xgi1; i < Nx; i++)
1922     {
1923         dcx = i*gridx-x;
1924         tmp = dcx*dcx;
1925         if (tmp >= rc2)
1926         {
1927             break;
1928         }
1929         *dx1    = i;
1930         dcx2[i] = tmp;
1931     }
1932 }
1933
1934 static gmx_inline void get_dx_dd(int Nx, real gridx, real rc2, int xgi, real x,
1935                                  int ncpddc, int shift_min, int shift_max,
1936                                  int *g0, int *g1, real *dcx2)
1937 {
1938     real dcx, tmp;
1939     int  g_min, g_max, shift_home;
1940
1941     if (xgi < 0)
1942     {
1943         g_min = 0;
1944         g_max = Nx - 1;
1945         *g0   = 0;
1946         *g1   = -1;
1947     }
1948     else if (xgi >= Nx)
1949     {
1950         g_min = 0;
1951         g_max = Nx - 1;
1952         *g0   = Nx;
1953         *g1   = Nx - 1;
1954     }
1955     else
1956     {
1957         if (ncpddc == 0)
1958         {
1959             g_min = 0;
1960             g_max = Nx - 1;
1961         }
1962         else
1963         {
1964             if (xgi < ncpddc)
1965             {
1966                 shift_home = 0;
1967             }
1968             else
1969             {
1970                 shift_home = -1;
1971             }
1972             g_min = (shift_min == shift_home ? 0          : ncpddc);
1973             g_max = (shift_max == shift_home ? ncpddc - 1 : Nx - 1);
1974         }
1975         if (shift_min > 0)
1976         {
1977             *g0 = g_min;
1978             *g1 = g_min - 1;
1979         }
1980         else if (shift_max < 0)
1981         {
1982             *g0 = g_max + 1;
1983             *g1 = g_max;
1984         }
1985         else
1986         {
1987             *g0       = xgi;
1988             *g1       = xgi;
1989             dcx2[xgi] = 0;
1990         }
1991     }
1992
1993     while (*g0 > g_min)
1994     {
1995         /* Check one grid cell down */
1996         dcx = ((*g0 - 1) + 1)*gridx - x;
1997         tmp = dcx*dcx;
1998         if (tmp >= rc2)
1999         {
2000             break;
2001         }
2002         (*g0)--;
2003         dcx2[*g0] = tmp;
2004     }
2005
2006     while (*g1 < g_max)
2007     {
2008         /* Check one grid cell up */
2009         dcx = (*g1 + 1)*gridx - x;
2010         tmp = dcx*dcx;
2011         if (tmp >= rc2)
2012         {
2013             break;
2014         }
2015         (*g1)++;
2016         dcx2[*g1] = tmp;
2017     }
2018 }
2019
2020
2021 #define sqr(x) ((x)*(x))
2022 #define calc_dx2(XI, YI, ZI, y) (sqr(XI-y[XX]) + sqr(YI-y[YY]) + sqr(ZI-y[ZZ]))
2023 #define calc_cyl_dx2(XI, YI, y) (sqr(XI-y[XX]) + sqr(YI-y[YY]))
2024 /****************************************************
2025  *
2026  *    F A S T   N E I G H B O R  S E A R C H I N G
2027  *
2028  *    Optimized neighboursearching routine using grid
2029  *    at least 1x1x1, see GROMACS manual
2030  *
2031  ****************************************************/
2032
2033
2034 static void get_cutoff2(t_forcerec *fr, gmx_bool bDoLongRange,
2035                         real *rvdw2, real *rcoul2,
2036                         real *rs2, real *rm2, real *rl2)
2037 {
2038     *rs2 = sqr(fr->rlist);
2039
2040     if (bDoLongRange && fr->bTwinRange)
2041     {
2042         /* With plain cut-off or RF we need to make the list exactly
2043          * up to the cut-off and the cut-off's can be different,
2044          * so we can not simply set them to rlistlong.
2045          * To keep this code compatible with (exotic) old cases,
2046          * we also create lists up to rvdw/rcoulomb for PME and Ewald.
2047          * The interaction check should correspond to:
2048          * !ir_vdw/coulomb_might_be_zero_at_cutoff from inputrec.c.
2049          */
2050         if (((fr->vdwtype == evdwCUT || fr->vdwtype == evdwPME) &&
2051              fr->vdw_modifier == eintmodNONE) ||
2052             fr->rvdw <= fr->rlist)
2053         {
2054             *rvdw2  = sqr(fr->rvdw);
2055         }
2056         else
2057         {
2058             *rvdw2  = sqr(fr->rlistlong);
2059         }
2060         if (((fr->eeltype == eelCUT ||
2061               (EEL_RF(fr->eeltype) && fr->eeltype != eelRF_ZERO) ||
2062               fr->eeltype == eelPME ||
2063               fr->eeltype == eelEWALD) &&
2064              fr->coulomb_modifier == eintmodNONE) ||
2065             fr->rcoulomb <= fr->rlist)
2066         {
2067             *rcoul2 = sqr(fr->rcoulomb);
2068         }
2069         else
2070         {
2071             *rcoul2 = sqr(fr->rlistlong);
2072         }
2073     }
2074     else
2075     {
2076         /* Workaround for a gcc -O3 or -ffast-math problem */
2077         *rvdw2  = *rs2;
2078         *rcoul2 = *rs2;
2079     }
2080     *rm2 = min(*rvdw2, *rcoul2);
2081     *rl2 = max(*rvdw2, *rcoul2);
2082 }
2083
2084 static void init_nsgrid_lists(t_forcerec *fr, int ngid, gmx_ns_t *ns)
2085 {
2086     real rvdw2, rcoul2, rs2, rm2, rl2;
2087     int  j;
2088
2089     get_cutoff2(fr, TRUE, &rvdw2, &rcoul2, &rs2, &rm2, &rl2);
2090
2091     /* Short range buffers */
2092     snew(ns->nl_sr, ngid);
2093     /* Counters */
2094     snew(ns->nsr, ngid);
2095     snew(ns->nlr_ljc, ngid);
2096     snew(ns->nlr_one, ngid);
2097
2098     /* Always allocate both list types, since rcoulomb might now change with PME load balancing */
2099     /* Long range VdW and Coul buffers */
2100     snew(ns->nl_lr_ljc, ngid);
2101     /* Long range VdW or Coul only buffers */
2102     snew(ns->nl_lr_one, ngid);
2103
2104     for (j = 0; (j < ngid); j++)
2105     {
2106         snew(ns->nl_sr[j], MAX_CG);
2107         snew(ns->nl_lr_ljc[j], MAX_CG);
2108         snew(ns->nl_lr_one[j], MAX_CG);
2109     }
2110     if (debug)
2111     {
2112         fprintf(debug,
2113                 "ns5_core: rs2 = %g, rm2 = %g, rl2 = %g (nm^2)\n",
2114                 rs2, rm2, rl2);
2115     }
2116 }
2117
2118 static int nsgrid_core(t_commrec *cr, t_forcerec *fr,
2119                        matrix box, int ngid,
2120                        gmx_localtop_t *top,
2121                        t_grid *grid,
2122                        t_excl bexcl[], gmx_bool *bExcludeAlleg,
2123                        t_mdatoms *md,
2124                        put_in_list_t *put_in_list,
2125                        gmx_bool bHaveVdW[],
2126                        gmx_bool bDoLongRange, gmx_bool bMakeQMMMnblist)
2127 {
2128     gmx_ns_t     *ns;
2129     atom_id     **nl_lr_ljc, **nl_lr_one, **nl_sr;
2130     int          *nlr_ljc, *nlr_one, *nsr;
2131     gmx_domdec_t *dd     = NULL;
2132     t_block      *cgs    = &(top->cgs);
2133     int          *cginfo = fr->cginfo;
2134     /* atom_id *i_atoms,*cgsindex=cgs->index; */
2135     ivec          sh0, sh1, shp;
2136     int           cell_x, cell_y, cell_z;
2137     int           d, tx, ty, tz, dx, dy, dz, cj;
2138 #ifdef ALLOW_OFFDIAG_LT_HALFDIAG
2139     int           zsh_ty, zsh_tx, ysh_tx;
2140 #endif
2141     int           dx0, dx1, dy0, dy1, dz0, dz1;
2142     int           Nx, Ny, Nz, shift = -1, j, nrj, nns, nn = -1;
2143     real          gridx, gridy, gridz, grid_x, grid_y, grid_z;
2144     real         *dcx2, *dcy2, *dcz2;
2145     int           zgi, ygi, xgi;
2146     int           cg0, cg1, icg = -1, cgsnr, i0, igid, nri, naaj, max_jcg;
2147     int           jcg0, jcg1, jjcg, cgj0, jgid;
2148     int          *grida, *gridnra, *gridind;
2149     gmx_bool      rvdw_lt_rcoul, rcoul_lt_rvdw;
2150     rvec          xi, *cgcm, grid_offset;
2151     real          r2, rs2, rvdw2, rcoul2, rm2, rl2, XI, YI, ZI, dcx, dcy, dcz, tmp1, tmp2;
2152     int          *i_egp_flags;
2153     gmx_bool      bDomDec, bTriclinicX, bTriclinicY;
2154     ivec          ncpddc;
2155
2156     ns = &fr->ns;
2157
2158     bDomDec = DOMAINDECOMP(cr);
2159     if (bDomDec)
2160     {
2161         dd = cr->dd;
2162     }
2163
2164     bTriclinicX = ((YY < grid->npbcdim &&
2165                     (!bDomDec || dd->nc[YY] == 1) && box[YY][XX] != 0) ||
2166                    (ZZ < grid->npbcdim &&
2167                     (!bDomDec || dd->nc[ZZ] == 1) && box[ZZ][XX] != 0));
2168     bTriclinicY =  (ZZ < grid->npbcdim &&
2169                     (!bDomDec || dd->nc[ZZ] == 1) && box[ZZ][YY] != 0);
2170
2171     cgsnr    = cgs->nr;
2172
2173     get_cutoff2(fr, bDoLongRange, &rvdw2, &rcoul2, &rs2, &rm2, &rl2);
2174
2175     rvdw_lt_rcoul = (rvdw2 >= rcoul2);
2176     rcoul_lt_rvdw = (rcoul2 >= rvdw2);
2177
2178     if (bMakeQMMMnblist)
2179     {
2180         rm2 = rl2;
2181         rs2 = rl2;
2182     }
2183
2184     nl_sr     = ns->nl_sr;
2185     nsr       = ns->nsr;
2186     nl_lr_ljc = ns->nl_lr_ljc;
2187     nl_lr_one = ns->nl_lr_one;
2188     nlr_ljc   = ns->nlr_ljc;
2189     nlr_one   = ns->nlr_one;
2190
2191     /* Unpack arrays */
2192     cgcm    = fr->cg_cm;
2193     Nx      = grid->n[XX];
2194     Ny      = grid->n[YY];
2195     Nz      = grid->n[ZZ];
2196     grida   = grid->a;
2197     gridind = grid->index;
2198     gridnra = grid->nra;
2199     nns     = 0;
2200
2201     gridx      = grid->cell_size[XX];
2202     gridy      = grid->cell_size[YY];
2203     gridz      = grid->cell_size[ZZ];
2204     grid_x     = 1/gridx;
2205     grid_y     = 1/gridy;
2206     grid_z     = 1/gridz;
2207     copy_rvec(grid->cell_offset, grid_offset);
2208     copy_ivec(grid->ncpddc, ncpddc);
2209     dcx2       = grid->dcx2;
2210     dcy2       = grid->dcy2;
2211     dcz2       = grid->dcz2;
2212
2213 #ifdef ALLOW_OFFDIAG_LT_HALFDIAG
2214     zsh_ty = floor(-box[ZZ][YY]/box[YY][YY]+0.5);
2215     zsh_tx = floor(-box[ZZ][XX]/box[XX][XX]+0.5);
2216     ysh_tx = floor(-box[YY][XX]/box[XX][XX]+0.5);
2217     if (zsh_tx != 0 && ysh_tx != 0)
2218     {
2219         /* This could happen due to rounding, when both ratios are 0.5 */
2220         ysh_tx = 0;
2221     }
2222 #endif
2223
2224     debug_gmx();
2225
2226     if (fr->n_tpi)
2227     {
2228         /* We only want a list for the test particle */
2229         cg0 = cgsnr - 1;
2230     }
2231     else
2232     {
2233         cg0 = grid->icg0;
2234     }
2235     cg1 = grid->icg1;
2236
2237     /* Set the shift range */
2238     for (d = 0; d < DIM; d++)
2239     {
2240         sh0[d] = -1;
2241         sh1[d] = 1;
2242         /* Check if we need periodicity shifts.
2243          * Without PBC or with domain decomposition we don't need them.
2244          */
2245         if (d >= ePBC2npbcdim(fr->ePBC) || (bDomDec && dd->nc[d] > 1))
2246         {
2247             shp[d] = 0;
2248         }
2249         else
2250         {
2251             if (d == XX &&
2252                 box[XX][XX] - fabs(box[YY][XX]) - fabs(box[ZZ][XX]) < sqrt(rl2))
2253             {
2254                 shp[d] = 2;
2255             }
2256             else
2257             {
2258                 shp[d] = 1;
2259             }
2260         }
2261     }
2262
2263     /* Loop over charge groups */
2264     for (icg = cg0; (icg < cg1); icg++)
2265     {
2266         igid = GET_CGINFO_GID(cginfo[icg]);
2267         /* Skip this charge group if all energy groups are excluded! */
2268         if (bExcludeAlleg[igid])
2269         {
2270             continue;
2271         }
2272
2273         i0   = cgs->index[icg];
2274
2275         if (bMakeQMMMnblist)
2276         {
2277             /* Skip this charge group if it is not a QM atom while making a
2278              * QM/MM neighbourlist
2279              */
2280             if (md->bQM[i0] == FALSE)
2281             {
2282                 continue; /* MM particle, go to next particle */
2283             }
2284
2285             /* Compute the number of charge groups that fall within the control
2286              * of this one (icg)
2287              */
2288             naaj    = calc_naaj(icg, cgsnr);
2289             jcg0    = icg;
2290             jcg1    = icg + naaj;
2291             max_jcg = cgsnr;
2292         }
2293         else
2294         {
2295             /* make a normal neighbourlist */
2296
2297             if (bDomDec)
2298             {
2299                 /* Get the j charge-group and dd cell shift ranges */
2300                 dd_get_ns_ranges(cr->dd, icg, &jcg0, &jcg1, sh0, sh1);
2301                 max_jcg = 0;
2302             }
2303             else
2304             {
2305                 /* Compute the number of charge groups that fall within the control
2306                  * of this one (icg)
2307                  */
2308                 naaj = calc_naaj(icg, cgsnr);
2309                 jcg0 = icg;
2310                 jcg1 = icg + naaj;
2311
2312                 if (fr->n_tpi)
2313                 {
2314                     /* The i-particle is awlways the test particle,
2315                      * so we want all j-particles
2316                      */
2317                     max_jcg = cgsnr - 1;
2318                 }
2319                 else
2320                 {
2321                     max_jcg  = jcg1 - cgsnr;
2322                 }
2323             }
2324         }
2325
2326         i_egp_flags = fr->egp_flags + igid*ngid;
2327
2328         /* Set the exclusions for the atoms in charge group icg using a bitmask */
2329         setexcl(i0, cgs->index[icg+1], &top->excls, TRUE, bexcl);
2330
2331         ci2xyz(grid, icg, &cell_x, &cell_y, &cell_z);
2332
2333         /* Changed iicg to icg, DvdS 990115
2334          * (but see consistency check above, DvdS 990330)
2335          */
2336 #ifdef NS5DB
2337         fprintf(log, "icg=%5d, naaj=%5d, cell %d %d %d\n",
2338                 icg, naaj, cell_x, cell_y, cell_z);
2339 #endif
2340         /* Loop over shift vectors in three dimensions */
2341         for (tz = -shp[ZZ]; tz <= shp[ZZ]; tz++)
2342         {
2343             ZI = cgcm[icg][ZZ]+tz*box[ZZ][ZZ];
2344             /* Calculate range of cells in Z direction that have the shift tz */
2345             zgi = cell_z + tz*Nz;
2346 #define FAST_DD_NS
2347 #ifndef FAST_DD_NS
2348             get_dx(Nz, gridz, rl2, zgi, ZI, &dz0, &dz1, dcz2);
2349 #else
2350             get_dx_dd(Nz, gridz, rl2, zgi, ZI-grid_offset[ZZ],
2351                       ncpddc[ZZ], sh0[ZZ], sh1[ZZ], &dz0, &dz1, dcz2);
2352 #endif
2353             if (dz0 > dz1)
2354             {
2355                 continue;
2356             }
2357             for (ty = -shp[YY]; ty <= shp[YY]; ty++)
2358             {
2359                 YI = cgcm[icg][YY]+ty*box[YY][YY]+tz*box[ZZ][YY];
2360                 /* Calculate range of cells in Y direction that have the shift ty */
2361                 if (bTriclinicY)
2362                 {
2363                     ygi = (int)(Ny + (YI - grid_offset[YY])*grid_y) - Ny;
2364                 }
2365                 else
2366                 {
2367                     ygi = cell_y + ty*Ny;
2368                 }
2369 #ifndef FAST_DD_NS
2370                 get_dx(Ny, gridy, rl2, ygi, YI, &dy0, &dy1, dcy2);
2371 #else
2372                 get_dx_dd(Ny, gridy, rl2, ygi, YI-grid_offset[YY],
2373                           ncpddc[YY], sh0[YY], sh1[YY], &dy0, &dy1, dcy2);
2374 #endif
2375                 if (dy0 > dy1)
2376                 {
2377                     continue;
2378                 }
2379                 for (tx = -shp[XX]; tx <= shp[XX]; tx++)
2380                 {
2381                     XI = cgcm[icg][XX]+tx*box[XX][XX]+ty*box[YY][XX]+tz*box[ZZ][XX];
2382                     /* Calculate range of cells in X direction that have the shift tx */
2383                     if (bTriclinicX)
2384                     {
2385                         xgi = (int)(Nx + (XI - grid_offset[XX])*grid_x) - Nx;
2386                     }
2387                     else
2388                     {
2389                         xgi = cell_x + tx*Nx;
2390                     }
2391 #ifndef FAST_DD_NS
2392                     get_dx(Nx, gridx, rl2, xgi*Nx, XI, &dx0, &dx1, dcx2);
2393 #else
2394                     get_dx_dd(Nx, gridx, rl2, xgi, XI-grid_offset[XX],
2395                               ncpddc[XX], sh0[XX], sh1[XX], &dx0, &dx1, dcx2);
2396 #endif
2397                     if (dx0 > dx1)
2398                     {
2399                         continue;
2400                     }
2401                     /* Adress: an explicit cg that has a weigthing function of 0 is excluded
2402                      *  from the neigbour list as it will not interact  */
2403                     if (fr->adress_type != eAdressOff)
2404                     {
2405                         if (md->wf[cgs->index[icg]] <= GMX_REAL_EPS && egp_explicit(fr, igid))
2406                         {
2407                             continue;
2408                         }
2409                     }
2410                     /* Get shift vector */
2411                     shift = XYZ2IS(tx, ty, tz);
2412 #ifdef NS5DB
2413                     range_check(shift, 0, SHIFTS);
2414 #endif
2415                     for (nn = 0; (nn < ngid); nn++)
2416                     {
2417                         nsr[nn]      = 0;
2418                         nlr_ljc[nn]  = 0;
2419                         nlr_one[nn]  = 0;
2420                     }
2421 #ifdef NS5DB
2422                     fprintf(log, "shift: %2d, dx0,1: %2d,%2d, dy0,1: %2d,%2d, dz0,1: %2d,%2d\n",
2423                             shift, dx0, dx1, dy0, dy1, dz0, dz1);
2424                     fprintf(log, "cgcm: %8.3f  %8.3f  %8.3f\n", cgcm[icg][XX],
2425                             cgcm[icg][YY], cgcm[icg][ZZ]);
2426                     fprintf(log, "xi:   %8.3f  %8.3f  %8.3f\n", XI, YI, ZI);
2427 #endif
2428                     for (dx = dx0; (dx <= dx1); dx++)
2429                     {
2430                         tmp1 = rl2 - dcx2[dx];
2431                         for (dy = dy0; (dy <= dy1); dy++)
2432                         {
2433                             tmp2 = tmp1 - dcy2[dy];
2434                             if (tmp2 > 0)
2435                             {
2436                                 for (dz = dz0; (dz <= dz1); dz++)
2437                                 {
2438                                     if (tmp2 > dcz2[dz])
2439                                     {
2440                                         /* Find grid-cell cj in which possible neighbours are */
2441                                         cj   = xyz2ci(Ny, Nz, dx, dy, dz);
2442
2443                                         /* Check out how many cgs (nrj) there in this cell */
2444                                         nrj  = gridnra[cj];
2445
2446                                         /* Find the offset in the cg list */
2447                                         cgj0 = gridind[cj];
2448
2449                                         /* Check if all j's are out of range so we
2450                                          * can skip the whole cell.
2451                                          * Should save some time, especially with DD.
2452                                          */
2453                                         if (nrj == 0 ||
2454                                             (grida[cgj0] >= max_jcg &&
2455                                              (grida[cgj0] >= jcg1 || grida[cgj0+nrj-1] < jcg0)))
2456                                         {
2457                                             continue;
2458                                         }
2459
2460                                         /* Loop over cgs */
2461                                         for (j = 0; (j < nrj); j++)
2462                                         {
2463                                             jjcg = grida[cgj0+j];
2464
2465                                             /* check whether this guy is in range! */
2466                                             if ((jjcg >= jcg0 && jjcg < jcg1) ||
2467                                                 (jjcg < max_jcg))
2468                                             {
2469                                                 r2 = calc_dx2(XI, YI, ZI, cgcm[jjcg]);
2470                                                 if (r2 < rl2)
2471                                                 {
2472                                                     /* jgid = gid[cgsatoms[cgsindex[jjcg]]]; */
2473                                                     jgid = GET_CGINFO_GID(cginfo[jjcg]);
2474                                                     /* check energy group exclusions */
2475                                                     if (!(i_egp_flags[jgid] & EGP_EXCL))
2476                                                     {
2477                                                         if (r2 < rs2)
2478                                                         {
2479                                                             if (nsr[jgid] >= MAX_CG)
2480                                                             {
2481                                                                 /* Add to short-range list */
2482                                                                 put_in_list(bHaveVdW, ngid, md, icg, jgid,
2483                                                                             nsr[jgid], nl_sr[jgid],
2484                                                                             cgs->index, /* cgsatoms, */ bexcl,
2485                                                                             shift, fr, FALSE, TRUE, TRUE, fr->solvent_opt);
2486                                                                 nsr[jgid] = 0;
2487                                                             }
2488                                                             nl_sr[jgid][nsr[jgid]++] = jjcg;
2489                                                         }
2490                                                         else if (r2 < rm2)
2491                                                         {
2492                                                             if (nlr_ljc[jgid] >= MAX_CG)
2493                                                             {
2494                                                                 /* Add to LJ+coulomb long-range list */
2495                                                                 put_in_list(bHaveVdW, ngid, md, icg, jgid,
2496                                                                             nlr_ljc[jgid], nl_lr_ljc[jgid], top->cgs.index,
2497                                                                             bexcl, shift, fr, TRUE, TRUE, TRUE, fr->solvent_opt);
2498                                                                 nlr_ljc[jgid] = 0;
2499                                                             }
2500                                                             nl_lr_ljc[jgid][nlr_ljc[jgid]++] = jjcg;
2501                                                         }
2502                                                         else
2503                                                         {
2504                                                             if (nlr_one[jgid] >= MAX_CG)
2505                                                             {
2506                                                                 /* Add to long-range list with only coul, or only LJ */
2507                                                                 put_in_list(bHaveVdW, ngid, md, icg, jgid,
2508                                                                             nlr_one[jgid], nl_lr_one[jgid], top->cgs.index,
2509                                                                             bexcl, shift, fr, TRUE, rvdw_lt_rcoul, rcoul_lt_rvdw, fr->solvent_opt);
2510                                                                 nlr_one[jgid] = 0;
2511                                                             }
2512                                                             nl_lr_one[jgid][nlr_one[jgid]++] = jjcg;
2513                                                         }
2514                                                     }
2515                                                 }
2516                                                 nns++;
2517                                             }
2518                                         }
2519                                     }
2520                                 }
2521                             }
2522                         }
2523                     }
2524                     /* CHECK whether there is anything left in the buffers */
2525                     for (nn = 0; (nn < ngid); nn++)
2526                     {
2527                         if (nsr[nn] > 0)
2528                         {
2529                             put_in_list(bHaveVdW, ngid, md, icg, nn, nsr[nn], nl_sr[nn],
2530                                         cgs->index, /* cgsatoms, */ bexcl,
2531                                         shift, fr, FALSE, TRUE, TRUE, fr->solvent_opt);
2532                         }
2533
2534                         if (nlr_ljc[nn] > 0)
2535                         {
2536                             put_in_list(bHaveVdW, ngid, md, icg, nn, nlr_ljc[nn],
2537                                         nl_lr_ljc[nn], top->cgs.index,
2538                                         bexcl, shift, fr, TRUE, TRUE, TRUE, fr->solvent_opt);
2539                         }
2540
2541                         if (nlr_one[nn] > 0)
2542                         {
2543                             put_in_list(bHaveVdW, ngid, md, icg, nn, nlr_one[nn],
2544                                         nl_lr_one[nn], top->cgs.index,
2545                                         bexcl, shift, fr, TRUE, rvdw_lt_rcoul, rcoul_lt_rvdw, fr->solvent_opt);
2546                         }
2547                     }
2548                 }
2549             }
2550         }
2551         /* setexcl(nri,i_atoms,&top->atoms.excl,FALSE,bexcl); */
2552         setexcl(cgs->index[icg], cgs->index[icg+1], &top->excls, FALSE, bexcl);
2553     }
2554     /* No need to perform any left-over force calculations anymore (as we used to do here)
2555      * since we now save the proper long-range lists for later evaluation.
2556      */
2557
2558     debug_gmx();
2559
2560     /* Close neighbourlists */
2561     close_neighbor_lists(fr, bMakeQMMMnblist);
2562
2563     return nns;
2564 }
2565
2566 void ns_realloc_natoms(gmx_ns_t *ns, int natoms)
2567 {
2568     int i;
2569
2570     if (natoms > ns->nra_alloc)
2571     {
2572         ns->nra_alloc = over_alloc_dd(natoms);
2573         srenew(ns->bexcl, ns->nra_alloc);
2574         for (i = 0; i < ns->nra_alloc; i++)
2575         {
2576             ns->bexcl[i] = 0;
2577         }
2578     }
2579 }
2580
2581 void init_ns(FILE *fplog, const t_commrec *cr,
2582              gmx_ns_t *ns, t_forcerec *fr,
2583              const gmx_mtop_t *mtop)
2584 {
2585     int  mt, icg, nr_in_cg, maxcg, i, j, jcg, ngid, ncg;
2586     t_block *cgs;
2587     char *ptr;
2588
2589     /* Compute largest charge groups size (# atoms) */
2590     nr_in_cg = 1;
2591     for (mt = 0; mt < mtop->nmoltype; mt++)
2592     {
2593         cgs = &mtop->moltype[mt].cgs;
2594         for (icg = 0; (icg < cgs->nr); icg++)
2595         {
2596             nr_in_cg = max(nr_in_cg, (int)(cgs->index[icg+1]-cgs->index[icg]));
2597         }
2598     }
2599
2600     /* Verify whether largest charge group is <= max cg.
2601      * This is determined by the type of the local exclusion type
2602      * Exclusions are stored in bits. (If the type is not large
2603      * enough, enlarge it, unsigned char -> unsigned short -> unsigned long)
2604      */
2605     maxcg = sizeof(t_excl)*8;
2606     if (nr_in_cg > maxcg)
2607     {
2608         gmx_fatal(FARGS, "Max #atoms in a charge group: %d > %d\n",
2609                   nr_in_cg, maxcg);
2610     }
2611
2612     ngid = mtop->groups.grps[egcENER].nr;
2613     snew(ns->bExcludeAlleg, ngid);
2614     for (i = 0; i < ngid; i++)
2615     {
2616         ns->bExcludeAlleg[i] = TRUE;
2617         for (j = 0; j < ngid; j++)
2618         {
2619             if (!(fr->egp_flags[i*ngid+j] & EGP_EXCL))
2620             {
2621                 ns->bExcludeAlleg[i] = FALSE;
2622             }
2623         }
2624     }
2625
2626     if (fr->bGrid)
2627     {
2628         /* Grid search */
2629         ns->grid = init_grid(fplog, fr);
2630         init_nsgrid_lists(fr, ngid, ns);
2631     }
2632     else
2633     {
2634         /* Simple search */
2635         snew(ns->ns_buf, ngid);
2636         for (i = 0; (i < ngid); i++)
2637         {
2638             snew(ns->ns_buf[i], SHIFTS);
2639         }
2640         ncg = ncg_mtop(mtop);
2641         snew(ns->simple_aaj, 2*ncg);
2642         for (jcg = 0; (jcg < ncg); jcg++)
2643         {
2644             ns->simple_aaj[jcg]     = jcg;
2645             ns->simple_aaj[jcg+ncg] = jcg;
2646         }
2647     }
2648
2649     /* Create array that determines whether or not atoms have VdW */
2650     snew(ns->bHaveVdW, fr->ntype);
2651     for (i = 0; (i < fr->ntype); i++)
2652     {
2653         for (j = 0; (j < fr->ntype); j++)
2654         {
2655             ns->bHaveVdW[i] = (ns->bHaveVdW[i] ||
2656                                (fr->bBHAM ?
2657                                 ((BHAMA(fr->nbfp, fr->ntype, i, j) != 0) ||
2658                                  (BHAMB(fr->nbfp, fr->ntype, i, j) != 0) ||
2659                                  (BHAMC(fr->nbfp, fr->ntype, i, j) != 0)) :
2660                                 ((C6(fr->nbfp, fr->ntype, i, j) != 0) ||
2661                                  (C12(fr->nbfp, fr->ntype, i, j) != 0))));
2662         }
2663     }
2664     if (debug)
2665     {
2666         pr_bvec(debug, 0, "bHaveVdW", ns->bHaveVdW, fr->ntype, TRUE);
2667     }
2668
2669     ns->nra_alloc = 0;
2670     ns->bexcl     = NULL;
2671     if (!DOMAINDECOMP(cr))
2672     {
2673         ns_realloc_natoms(ns, mtop->natoms);
2674     }
2675
2676     ns->nblist_initialized = FALSE;
2677
2678     /* nbr list debug dump */
2679     {
2680         char *ptr = getenv("GMX_DUMP_NL");
2681         if (ptr)
2682         {
2683             ns->dump_nl = strtol(ptr, NULL, 10);
2684             if (fplog)
2685             {
2686                 fprintf(fplog, "GMX_DUMP_NL = %d", ns->dump_nl);
2687             }
2688         }
2689         else
2690         {
2691             ns->dump_nl = 0;
2692         }
2693     }
2694 }
2695
2696
2697 int search_neighbours(FILE *log, t_forcerec *fr,
2698                       matrix box,
2699                       gmx_localtop_t *top,
2700                       gmx_groups_t *groups,
2701                       t_commrec *cr,
2702                       t_nrnb *nrnb, t_mdatoms *md,
2703                       gmx_bool bFillGrid,
2704                       gmx_bool bDoLongRangeNS)
2705 {
2706     t_block  *cgs = &(top->cgs);
2707     rvec     box_size, grid_x0, grid_x1;
2708     int      i, j, m, ngid;
2709     real     min_size, grid_dens;
2710     int      nsearch;
2711     gmx_bool     bGrid;
2712     char     *ptr;
2713     gmx_bool     *i_egp_flags;
2714     int      cg_start, cg_end, start, end;
2715     gmx_ns_t *ns;
2716     t_grid   *grid;
2717     gmx_domdec_zones_t *dd_zones;
2718     put_in_list_t *put_in_list;
2719
2720     ns = &fr->ns;
2721
2722     /* Set some local variables */
2723     bGrid = fr->bGrid;
2724     ngid  = groups->grps[egcENER].nr;
2725
2726     for (m = 0; (m < DIM); m++)
2727     {
2728         box_size[m] = box[m][m];
2729     }
2730
2731     if (fr->ePBC != epbcNONE)
2732     {
2733         if (sqr(fr->rlistlong) >= max_cutoff2(fr->ePBC, box))
2734         {
2735             gmx_fatal(FARGS, "One of the box vectors has become shorter than twice the cut-off length or box_yy-|box_zy| or box_zz has become smaller than the cut-off.");
2736         }
2737         if (!bGrid)
2738         {
2739             min_size = min(box_size[XX], min(box_size[YY], box_size[ZZ]));
2740             if (2*fr->rlistlong >= min_size)
2741             {
2742                 gmx_fatal(FARGS, "One of the box diagonal elements has become smaller than twice the cut-off length.");
2743             }
2744         }
2745     }
2746
2747     if (DOMAINDECOMP(cr))
2748     {
2749         ns_realloc_natoms(ns, cgs->index[cgs->nr]);
2750     }
2751     debug_gmx();
2752
2753     /* Reset the neighbourlists */
2754     reset_neighbor_lists(fr, TRUE, TRUE);
2755
2756     if (bGrid && bFillGrid)
2757     {
2758
2759         grid = ns->grid;
2760         if (DOMAINDECOMP(cr))
2761         {
2762             dd_zones = domdec_zones(cr->dd);
2763         }
2764         else
2765         {
2766             dd_zones = NULL;
2767
2768             get_nsgrid_boundaries(grid->nboundeddim, box, NULL, NULL, NULL, NULL,
2769                                   cgs->nr, fr->cg_cm, grid_x0, grid_x1, &grid_dens);
2770
2771             grid_first(log, grid, NULL, NULL, box, grid_x0, grid_x1,
2772                        fr->rlistlong, grid_dens);
2773         }
2774         debug_gmx();
2775
2776         start = 0;
2777         end   = cgs->nr;
2778
2779         if (DOMAINDECOMP(cr))
2780         {
2781             end = cgs->nr;
2782             fill_grid(dd_zones, grid, end, -1, end, fr->cg_cm);
2783             grid->icg0 = 0;
2784             grid->icg1 = dd_zones->izone[dd_zones->nizone-1].cg1;
2785         }
2786         else
2787         {
2788             fill_grid(NULL, grid, cgs->nr, fr->cg0, fr->hcg, fr->cg_cm);
2789             grid->icg0 = fr->cg0;
2790             grid->icg1 = fr->hcg;
2791             debug_gmx();
2792         }
2793
2794         calc_elemnr(grid, start, end, cgs->nr);
2795         calc_ptrs(grid);
2796         grid_last(grid, start, end, cgs->nr);
2797
2798         if (gmx_debug_at)
2799         {
2800             check_grid(grid);
2801             print_grid(debug, grid);
2802         }
2803     }
2804     else if (fr->n_tpi)
2805     {
2806         /* Set the grid cell index for the test particle only.
2807          * The cell to cg index is not corrected, but that does not matter.
2808          */
2809         fill_grid(NULL, ns->grid, fr->hcg, fr->hcg-1, fr->hcg, fr->cg_cm);
2810     }
2811     debug_gmx();
2812
2813     if (fr->adress_type == eAdressOff)
2814     {
2815         if (!fr->ns.bCGlist)
2816         {
2817             put_in_list = put_in_list_at;
2818         }
2819         else
2820         {
2821             put_in_list = put_in_list_cg;
2822         }
2823     }
2824     else
2825     {
2826         put_in_list = put_in_list_adress;
2827     }
2828
2829     /* Do the core! */
2830     if (bGrid)
2831     {
2832         grid    = ns->grid;
2833         nsearch = nsgrid_core(cr, fr, box, ngid, top,
2834                               grid, ns->bexcl, ns->bExcludeAlleg,
2835                               md, put_in_list, ns->bHaveVdW,
2836                               bDoLongRangeNS, FALSE);
2837
2838         /* neighbour searching withouth QMMM! QM atoms have zero charge in
2839          * the classical calculation. The charge-charge interaction
2840          * between QM and MM atoms is handled in the QMMM core calculation
2841          * (see QMMM.c). The VDW however, we'd like to compute classically
2842          * and the QM MM atom pairs have just been put in the
2843          * corresponding neighbourlists. in case of QMMM we still need to
2844          * fill a special QMMM neighbourlist that contains all neighbours
2845          * of the QM atoms. If bQMMM is true, this list will now be made:
2846          */
2847         if (fr->bQMMM && fr->qr->QMMMscheme != eQMMMschemeoniom)
2848         {
2849             nsearch += nsgrid_core(cr, fr, box, ngid, top,
2850                                    grid, ns->bexcl, ns->bExcludeAlleg,
2851                                    md, put_in_list_qmmm, ns->bHaveVdW,
2852                                    bDoLongRangeNS, TRUE);
2853         }
2854     }
2855     else
2856     {
2857         nsearch = ns_simple_core(fr, top, md, box, box_size,
2858                                  ns->bexcl, ns->simple_aaj,
2859                                  ngid, ns->ns_buf, put_in_list, ns->bHaveVdW);
2860     }
2861     debug_gmx();
2862
2863 #ifdef DEBUG
2864     pr_nsblock(log);
2865 #endif
2866
2867     inc_nrnb(nrnb, eNR_NS, nsearch);
2868     /* inc_nrnb(nrnb,eNR_LR,fr->nlr); */
2869
2870     return nsearch;
2871 }
2872
2873 int natoms_beyond_ns_buffer(t_inputrec *ir, t_forcerec *fr, t_block *cgs,
2874                             matrix scale_tot, rvec *x)
2875 {
2876     int  cg0, cg1, cg, a0, a1, a, i, j;
2877     real rint, hbuf2, scale;
2878     rvec *cg_cm, cgsc;
2879     gmx_bool bIsotropic;
2880     int  nBeyond;
2881
2882     nBeyond = 0;
2883
2884     rint = max(ir->rcoulomb, ir->rvdw);
2885     if (ir->rlist < rint)
2886     {
2887         gmx_fatal(FARGS, "The neighbor search buffer has negative size: %f nm",
2888                   ir->rlist - rint);
2889     }
2890     cg_cm = fr->cg_cm;
2891
2892     cg0 = fr->cg0;
2893     cg1 = fr->hcg;
2894
2895     if (!EI_DYNAMICS(ir->eI) || !DYNAMIC_BOX(*ir))
2896     {
2897         hbuf2 = sqr(0.5*(ir->rlist - rint));
2898         for (cg = cg0; cg < cg1; cg++)
2899         {
2900             a0 = cgs->index[cg];
2901             a1 = cgs->index[cg+1];
2902             for (a = a0; a < a1; a++)
2903             {
2904                 if (distance2(cg_cm[cg], x[a]) > hbuf2)
2905                 {
2906                     nBeyond++;
2907                 }
2908             }
2909         }
2910     }
2911     else
2912     {
2913         bIsotropic = TRUE;
2914         scale      = scale_tot[0][0];
2915         for (i = 1; i < DIM; i++)
2916         {
2917             /* With anisotropic scaling, the original spherical ns volumes become
2918              * ellipsoids. To avoid costly transformations we use the minimum
2919              * eigenvalue of the scaling matrix for determining the buffer size.
2920              * Since the lower half is 0, the eigenvalues are the diagonal elements.
2921              */
2922             scale = min(scale, scale_tot[i][i]);
2923             if (scale_tot[i][i] != scale_tot[i-1][i-1])
2924             {
2925                 bIsotropic = FALSE;
2926             }
2927             for (j = 0; j < i; j++)
2928             {
2929                 if (scale_tot[i][j] != 0)
2930                 {
2931                     bIsotropic = FALSE;
2932                 }
2933             }
2934         }
2935         hbuf2 = sqr(0.5*(scale*ir->rlist - rint));
2936         if (bIsotropic)
2937         {
2938             for (cg = cg0; cg < cg1; cg++)
2939             {
2940                 svmul(scale, cg_cm[cg], cgsc);
2941                 a0 = cgs->index[cg];
2942                 a1 = cgs->index[cg+1];
2943                 for (a = a0; a < a1; a++)
2944                 {
2945                     if (distance2(cgsc, x[a]) > hbuf2)
2946                     {
2947                         nBeyond++;
2948                     }
2949                 }
2950             }
2951         }
2952         else
2953         {
2954             /* Anistropic scaling */
2955             for (cg = cg0; cg < cg1; cg++)
2956             {
2957                 /* Since scale_tot contains the transpose of the scaling matrix,
2958                  * we need to multiply with the transpose.
2959                  */
2960                 tmvmul_ur0(scale_tot, cg_cm[cg], cgsc);
2961                 a0 = cgs->index[cg];
2962                 a1 = cgs->index[cg+1];
2963                 for (a = a0; a < a1; a++)
2964                 {
2965                     if (distance2(cgsc, x[a]) > hbuf2)
2966                     {
2967                         nBeyond++;
2968                     }
2969                 }
2970             }
2971         }
2972     }
2973
2974     return nBeyond;
2975 }