src/mdlib/tpi.c

   1 /* -*- mode: c; tab-width: 4; indent-tabs-mode: nil; c-basic-offset: 4; c-file-style: "stroustrup"; -*-
   2  *
   3  *
   4  *                This source code is part of
   5  *
   6  *                 G   R   O   M   A   C   S
   7  *
   8  *          GROningen MAchine for Chemical Simulations
   9  *
  10  *                        VERSION 3.2.0
  11  * Written by David van der Spoel, Erik Lindahl, Berk Hess, and others.
  12  * Copyright (c) 1991-2000, University of Groningen, The Netherlands.
  13  * Copyright (c) 2001-2004, The GROMACS development team,
  14  * check out http://www.gromacs.org for more information.
  15
  16  * This program is free software; you can redistribute it and/or
  17  * modify it under the terms of the GNU General Public License
  18  * as published by the Free Software Foundation; either version 2
  19  * of the License, or (at your option) any later version.
  20  *
  21  * If you want to redistribute modifications, please consider that
  22  * scientific software is very special. Version control is crucial -
  23  * bugs must be traceable. We will be happy to consider code for
  24  * inclusion in the official distribution, but derived work must not
  25  * be called official GROMACS. Details are found in the README & COPYING
  26  * files - if they are missing, get the official version at www.gromacs.org.
  27  *
  28  * To help us fund GROMACS development, we humbly ask that you cite
  29  * the papers on the package - you can find them in the top README file.
  30  *
  31  * For more info, check our website at http://www.gromacs.org
  32  *
  33  * And Hey:
  34  * GROwing Monsters And Cloning Shrimps
  35  */
  36 #ifdef HAVE_CONFIG_H
  37 #include <config.h>
  38 #endif
  39
  40 #include <string.h>
  41 #include <time.h>
  42 #include <math.h>
  43 #include "sysstuff.h"
  44 #include "string2.h"
  45 #include "network.h"
  46 #include "confio.h"
  47 #include "copyrite.h"
  48 #include "smalloc.h"
  49 #include "nrnb.h"
  50 #include "main.h"
  51 #include "chargegroup.h"
  52 #include "force.h"
  53 #include "macros.h"
  54 #include "random.h"
  55 #include "names.h"
  56 #include "gmx_fatal.h"
  57 #include "txtdump.h"
  58 #include "typedefs.h"
  59 #include "update.h"
  60 #include "random.h"
  61 #include "constr.h"
  62 #include "vec.h"
  63 #include "statutil.h"
  64 #include "tgroup.h"
  65 #include "mdebin.h"
  66 #include "vsite.h"
  67 #include "force.h"
  68 #include "mdrun.h"
  69 #include "domdec.h"
  70 #include "partdec.h"
  71 #include "gmx_random.h"
  72 #include "physics.h"
  73 #include "xvgr.h"
  74 #include "mdatoms.h"
  75 #include "ns.h"
  76 #include "gmx_wallcycle.h"
  77 #include "mtop_util.h"
  78 #include "gmxfio.h"
  79 #include "pme.h"
  80 #include "gbutil.h"
  81
  82 #ifdef GMX_X86_SSE2
  83 #include "gmx_x86_sse2.h"
  84 #endif
  85
  86
  87 static void global_max(t_commrec *cr,int *n)
  88 {
  89   int *sum,i;
  90
  91   snew(sum,cr->nnodes);
  92   sum[cr->nodeid] = *n;
  93   gmx_sumi(cr->nnodes,sum,cr);
  94   for(i=0; i<cr->nnodes; i++)
  95     *n = max(*n,sum[i]);
  96
  97   sfree(sum);
  98 }
  99
 100 static void realloc_bins(double **bin,int *nbin,int nbin_new)
 101 {
 102   int i;
 103
 104   if (nbin_new != *nbin) {
 105     srenew(*bin,nbin_new);
 106     for(i=*nbin; i<nbin_new; i++)
 107       (*bin)[i] = 0;
 108     *nbin = nbin_new;
 109   }
 110 }
 111
 112 double do_tpi(FILE *fplog,t_commrec *cr,
 113               int nfile, const t_filenm fnm[],
 114               const output_env_t oenv, gmx_bool bVerbose,gmx_bool bCompact,
 115               int nstglobalcomm,
 116               gmx_vsite_t *vsite,gmx_constr_t constr,
 117               int stepout,
 118               t_inputrec *inputrec,
 119               gmx_mtop_t *top_global,t_fcdata *fcd,
 120               t_state *state,
 121               t_mdatoms *mdatoms,
 122               t_nrnb *nrnb,gmx_wallcycle_t wcycle,
 123               gmx_edsam_t ed,
 124               t_forcerec *fr,
 125               int repl_ex_nst, int repl_ex_nex, int repl_ex_seed,
 126               gmx_membed_t membed,
 127               real cpt_period,real max_hours,
 128               const char *deviceOptions,
 129               unsigned long Flags,
 130               gmx_runtime_t *runtime)
 131 {
 132   const char *TPI="Test Particle Insertion";
 133   gmx_localtop_t *top;
 134   gmx_groups_t *groups;
 135   gmx_enerdata_t *enerd;
 136   rvec   *f;
 137   real   lambda,t,temp,beta,drmax,epot;
 138   double embU,sum_embU,*sum_UgembU,V,V_all,VembU_all;
 139   t_trxstatus   *status;
 140   t_trxframe rerun_fr;
 141   gmx_bool   bDispCorr,bCharge,bRFExcl,bNotLastFrame,bStateChanged,bNS,bOurStep;
 142   tensor force_vir,shake_vir,vir,pres;
 143   int    cg_tp,a_tp0,a_tp1,ngid,gid_tp,nener,e;
 144   rvec   *x_mol;
 145   rvec   mu_tot,x_init,dx,x_tp;
 146   int    nnodes,frame,nsteps,step;
 147   int    i,start,end;
 148   gmx_rng_t tpi_rand;
 149   FILE   *fp_tpi=NULL;
 150   char   *ptr,*dump_pdb,**leg,str[STRLEN],str2[STRLEN];
 151   double dbl,dump_ener;
 152   gmx_bool   bCavity;
 153   int    nat_cavity=0,d;
 154   real   *mass_cavity=NULL,mass_tot;
 155   int    nbin;
 156   double invbinw,*bin,refvolshift,logV,bUlogV;
 157   real dvdl,prescorr,enercorr,dvdlcorr;
 158   gmx_bool bEnergyOutOfBounds;
 159   const char *tpid_leg[2]={"direct","reweighted"};
 160
 161   /* Since there is no upper limit to the insertion energies,
 162    * we need to set an upper limit for the distribution output.
 163    */
 164   real bU_bin_limit      = 50;
 165   real bU_logV_bin_limit = bU_bin_limit + 10;
 166
 167   nnodes = cr->nnodes;
 168
 169   top = gmx_mtop_generate_local_top(top_global,inputrec);
 170
 171   groups = &top_global->groups;
 172
 173   bCavity = (inputrec->eI == eiTPIC);
 174   if (bCavity) {
 175     ptr = getenv("GMX_TPIC_MASSES");
 176     if (ptr == NULL) {
 177       nat_cavity = 1;
 178     } else {
 179       /* Read (multiple) masses from env var GMX_TPIC_MASSES,
 180        * The center of mass of the last atoms is then used for TPIC.
 181        */
 182       nat_cavity = 0;
 183       while (sscanf(ptr,"%lf%n",&dbl,&i) > 0) {
 184         srenew(mass_cavity,nat_cavity+1);
 185         mass_cavity[nat_cavity] = dbl;
 186         fprintf(fplog,"mass[%d] = %f\n",
 187                 nat_cavity+1,mass_cavity[nat_cavity]);
 188         nat_cavity++;
 189         ptr += i;
 190       }
 191       if (nat_cavity == 0)
 192         gmx_fatal(FARGS,"Found %d masses in GMX_TPIC_MASSES",nat_cavity);
 193     }
 194   }
 195
 196   /*
 197   init_em(fplog,TPI,inputrec,&lambda,nrnb,mu_tot,
 198   state->box,fr,mdatoms,top,cr,nfile,fnm,NULL,NULL);*/
 199   /* We never need full pbc for TPI */
 200   fr->ePBC = epbcXYZ;
 201   /* Determine the temperature for the Boltzmann weighting */
 202   temp = inputrec->opts.ref_t[0];
 203   if (fplog) {
 204     for(i=1; (i<inputrec->opts.ngtc); i++) {
 205       if (inputrec->opts.ref_t[i] != temp) {
 206         fprintf(fplog,"\nWARNING: The temperatures of the different temperature coupling groups are not identical\n\n");
 207         fprintf(stderr,"\nWARNING: The temperatures of the different temperature coupling groups are not identical\n\n");
 208       }
 209     }
 210     fprintf(fplog,
 211             "\n  The temperature for test particle insertion is %.3f K\n\n",
 212             temp);
 213   }
 214   beta = 1.0/(BOLTZ*temp);
 215
 216   /* Number of insertions per frame */
 217   nsteps = inputrec->nsteps;
 218
 219   /* Use the same neighborlist with more insertions points
 220    * in a sphere of radius drmax around the initial point
 221    */
 222   /* This should be a proper mdp parameter */
 223   drmax = inputrec->rtpi;
 224
 225   /* An environment variable can be set to dump all configurations
 226    * to pdb with an insertion energy <= this value.
 227    */
 228   dump_pdb = getenv("GMX_TPI_DUMP");
 229   dump_ener = 0;
 230   if (dump_pdb)
 231     sscanf(dump_pdb,"%lf",&dump_ener);
 232
 233   atoms2md(top_global,inputrec,0,NULL,0,top_global->natoms,mdatoms);
 234   update_mdatoms(mdatoms,inputrec->fepvals->init_lambda);
 235
 236   snew(enerd,1);
 237   init_enerdata(groups->grps[egcENER].nr,inputrec->fepvals->n_lambda,enerd);
 238   snew(f,top_global->natoms);
 239
 240   /* Print to log file  */
 241   runtime_start(runtime);
 242   print_date_and_time(fplog,cr->nodeid,
 243                       "Started Test Particle Insertion",runtime);
 244   wallcycle_start(wcycle,ewcRUN);
 245
 246   /* The last charge group is the group to be inserted */
 247   cg_tp = top->cgs.nr - 1;
 248   a_tp0 = top->cgs.index[cg_tp];
 249   a_tp1 = top->cgs.index[cg_tp+1];
 250   if (debug)
 251     fprintf(debug,"TPI cg %d, atoms %d-%d\n",cg_tp,a_tp0,a_tp1);
 252   if (a_tp1 - a_tp0 > 1 &&
 253       (inputrec->rlist < inputrec->rcoulomb ||
 254        inputrec->rlist < inputrec->rvdw))
 255     gmx_fatal(FARGS,"Can not do TPI for multi-atom molecule with a twin-range cut-off");
 256   snew(x_mol,a_tp1-a_tp0);
 257
 258   bDispCorr = (inputrec->eDispCorr != edispcNO);
 259   bCharge = FALSE;
 260   for(i=a_tp0; i<a_tp1; i++) {
 261     /* Copy the coordinates of the molecule to be insterted */
 262     copy_rvec(state->x[i],x_mol[i-a_tp0]);
 263     /* Check if we need to print electrostatic energies */
 264     bCharge |= (mdatoms->chargeA[i] != 0 ||
 265                 (mdatoms->chargeB && mdatoms->chargeB[i] != 0));
 266   }
 267   bRFExcl = (bCharge && EEL_RF(fr->eeltype) && fr->eeltype!=eelRF_NEC);
 268
 269   calc_cgcm(fplog,cg_tp,cg_tp+1,&(top->cgs),state->x,fr->cg_cm);
 270   if (bCavity) {
 271     if (norm(fr->cg_cm[cg_tp]) > 0.5*inputrec->rlist && fplog) {
 272       fprintf(fplog, "WARNING: Your TPI molecule is not centered at 0,0,0\n");
 273       fprintf(stderr,"WARNING: Your TPI molecule is not centered at 0,0,0\n");
 274     }
 275   } else {
 276     /* Center the molecule to be inserted at zero */
 277      for(i=0; i<a_tp1-a_tp0; i++)
 278       rvec_dec(x_mol[i],fr->cg_cm[cg_tp]);
 279   }
 280
 281   if (fplog) {
 282     fprintf(fplog,"\nWill insert %d atoms %s partial charges\n",
 283             a_tp1-a_tp0,bCharge ? "with" : "without");
 284
 285     fprintf(fplog,"\nWill insert %d times in each frame of %s\n",
 286             nsteps,opt2fn("-rerun",nfile,fnm));
 287   }
 288
 289     if (!bCavity)
 290     {
 291         if (inputrec->nstlist > 1)
 292         {
 293             if (drmax==0 && a_tp1-a_tp0==1)
 294             {
 295                 gmx_fatal(FARGS,"Re-using the neighborlist %d times for insertions of a single atom in a sphere of radius %f does not make sense",inputrec->nstlist,drmax);
 296             }
 297             if (fplog)
 298             {
 299                 fprintf(fplog,"Will use the same neighborlist for %d insertions in a sphere of radius %f\n",inputrec->nstlist,drmax);
 300             }
 301         }
 302     }
 303     else
 304     {
 305         if (fplog)
 306         {
 307             fprintf(fplog,"Will insert randomly in a sphere of radius %f around the center of the cavity\n",drmax);
 308         }
 309     }
 310
 311   ngid = groups->grps[egcENER].nr;
 312   gid_tp = GET_CGINFO_GID(fr->cginfo[cg_tp]);
 313   nener = 1 + ngid;
 314   if (bDispCorr)
 315     nener += 1;
 316   if (bCharge) {
 317     nener += ngid;
 318     if (bRFExcl)
 319       nener += 1;
 320     if (EEL_FULL(fr->eeltype))
 321       nener += 1;
 322   }
 323   snew(sum_UgembU,nener);
 324
 325   /* Initialize random generator */
 326   tpi_rand = gmx_rng_init(inputrec->ld_seed);
 327
 328   if (MASTER(cr)) {
 329     fp_tpi = xvgropen(opt2fn("-tpi",nfile,fnm),
 330                       "TPI energies","Time (ps)",
 331                       "(kJ mol\\S-1\\N) / (nm\\S3\\N)",oenv);
 332     xvgr_subtitle(fp_tpi,"f. are averages over one frame",oenv);
 333     snew(leg,4+nener);
 334     e = 0;
 335     sprintf(str,"-kT log(<Ve\\S-\\betaU\\N>/<V>)");
 336     leg[e++] = strdup(str);
 337     sprintf(str,"f. -kT log<e\\S-\\betaU\\N>");
 338     leg[e++] = strdup(str);
 339     sprintf(str,"f. <e\\S-\\betaU\\N>");
 340     leg[e++] = strdup(str);
 341     sprintf(str,"f. V");
 342     leg[e++] = strdup(str);
 343     sprintf(str,"f. <Ue\\S-\\betaU\\N>");
 344     leg[e++] = strdup(str);
 345     for(i=0; i<ngid; i++) {
 346       sprintf(str,"f. <U\\sVdW %s\\Ne\\S-\\betaU\\N>",
 347               *(groups->grpname[groups->grps[egcENER].nm_ind[i]]));
 348       leg[e++] = strdup(str);
 349     }
 350     if (bDispCorr) {
 351       sprintf(str,"f. <U\\sdisp c\\Ne\\S-\\betaU\\N>");
 352       leg[e++] = strdup(str);
 353     }
 354     if (bCharge) {
 355       for(i=0; i<ngid; i++) {
 356         sprintf(str,"f. <U\\sCoul %s\\Ne\\S-\\betaU\\N>",
 357                 *(groups->grpname[groups->grps[egcENER].nm_ind[i]]));
 358         leg[e++] = strdup(str);
 359       }
 360       if (bRFExcl) {
 361         sprintf(str,"f. <U\\sRF excl\\Ne\\S-\\betaU\\N>");
 362         leg[e++] = strdup(str);
 363       }
 364       if (EEL_FULL(fr->eeltype)) {
 365         sprintf(str,"f. <U\\sCoul recip\\Ne\\S-\\betaU\\N>");
 366         leg[e++] = strdup(str);
 367       }
 368     }
 369     xvgr_legend(fp_tpi,4+nener,(const char**)leg,oenv);
 370     for(i=0; i<4+nener; i++)
 371       sfree(leg[i]);
 372     sfree(leg);
 373   }
 374   clear_rvec(x_init);
 375   V_all = 0;
 376   VembU_all = 0;
 377
 378   invbinw = 10;
 379   nbin = 10;
 380   snew(bin,nbin);
 381
 382   bNotLastFrame = read_first_frame(oenv,&status,opt2fn("-rerun",nfile,fnm),
 383                                    &rerun_fr,TRX_NEED_X);
 384   frame = 0;
 385
 386   if (rerun_fr.natoms - (bCavity ? nat_cavity : 0) !=
 387       mdatoms->nr - (a_tp1 - a_tp0))
 388     gmx_fatal(FARGS,"Number of atoms in trajectory (%d)%s "
 389               "is not equal the number in the run input file (%d) "
 390               "minus the number of atoms to insert (%d)\n",
 391               rerun_fr.natoms,bCavity ? " minus one" : "",
 392               mdatoms->nr,a_tp1-a_tp0);
 393
 394   refvolshift = log(det(rerun_fr.box));
 395
 396 #ifdef GMX_X86_SSE2
 397     /* Make sure we don't detect SSE overflow generated before this point */
 398     gmx_mm_check_and_reset_overflow();
 399 #endif
 400
 401     while (bNotLastFrame)
 402     {
 403         lambda = rerun_fr.lambda;
 404         t = rerun_fr.time;
 405
 406         sum_embU = 0;
 407         for(e=0; e<nener; e++)
 408         {
 409             sum_UgembU[e] = 0;
 410         }
 411
 412         /* Copy the coordinates from the input trajectory */
 413         for(i=0; i<rerun_fr.natoms; i++)
 414         {
 415             copy_rvec(rerun_fr.x[i],state->x[i]);
 416         }
 417         copy_mat(rerun_fr.box,state->box);
 418
 419         V = det(state->box);
 420         logV = log(V);
 421
 422         bStateChanged = TRUE;
 423         bNS = TRUE;
 424         for(step=0; step<nsteps; step++)
 425         {
 426             /* In parallel all nodes generate all random configurations.
 427              * In that way the result is identical to a single cpu tpi run.
 428              */
 429             if (!bCavity)
 430             {
 431                 /* Random insertion in the whole volume */
 432                 bNS = (step % inputrec->nstlist == 0);
 433                 if (bNS)
 434                 {
 435                     /* Generate a random position in the box */
 436                     x_init[XX] = gmx_rng_uniform_real(tpi_rand)*state->box[XX][XX];
 437                     x_init[YY] = gmx_rng_uniform_real(tpi_rand)*state->box[YY][YY];
 438                     x_init[ZZ] = gmx_rng_uniform_real(tpi_rand)*state->box[ZZ][ZZ];
 439                 }
 440                 if (inputrec->nstlist == 1)
 441                 {
 442                     copy_rvec(x_init,x_tp);
 443                 }
 444                 else
 445                 {
 446                     /* Generate coordinates within |dx|=drmax of x_init */
 447                     do
 448                     {
 449                         dx[XX] = (2*gmx_rng_uniform_real(tpi_rand) - 1)*drmax;
 450                         dx[YY] = (2*gmx_rng_uniform_real(tpi_rand) - 1)*drmax;
 451                         dx[ZZ] = (2*gmx_rng_uniform_real(tpi_rand) - 1)*drmax;
 452                     }
 453                     while (norm2(dx) > drmax*drmax);
 454                     rvec_add(x_init,dx,x_tp);
 455                 }
 456             }
 457             else
 458             {
 459                 /* Random insertion around a cavity location
 460                  * given by the last coordinate of the trajectory.
 461                  */
 462                 if (step == 0)
 463                 {
 464                     if (nat_cavity == 1)
 465                     {
 466                         /* Copy the location of the cavity */
 467                         copy_rvec(rerun_fr.x[rerun_fr.natoms-1],x_init);
 468                     }
 469                     else
 470                     {
 471                         /* Determine the center of mass of the last molecule */
 472                         clear_rvec(x_init);
 473                         mass_tot = 0;
 474                         for(i=0; i<nat_cavity; i++)
 475                         {
 476                             for(d=0; d<DIM; d++)
 477                             {
 478                                 x_init[d] +=
 479                                     mass_cavity[i]*rerun_fr.x[rerun_fr.natoms-nat_cavity+i][d];
 480                             }
 481                             mass_tot += mass_cavity[i];
 482                         }
 483                         for(d=0; d<DIM; d++)
 484                         {
 485                             x_init[d] /= mass_tot;
 486                         }
 487                     }
 488                 }
 489                 /* Generate coordinates within |dx|=drmax of x_init */
 490                 do
 491                 {
 492                     dx[XX] = (2*gmx_rng_uniform_real(tpi_rand) - 1)*drmax;
 493                     dx[YY] = (2*gmx_rng_uniform_real(tpi_rand) - 1)*drmax;
 494                     dx[ZZ] = (2*gmx_rng_uniform_real(tpi_rand) - 1)*drmax;
 495                 }
 496                 while (norm2(dx) > drmax*drmax);
 497                 rvec_add(x_init,dx,x_tp);
 498             }
 499
 500             if (a_tp1 - a_tp0 == 1)
 501             {
 502                 /* Insert a single atom, just copy the insertion location */
 503         copy_rvec(x_tp,state->x[a_tp0]);
 504             }
 505             else
 506             {
 507                 /* Copy the coordinates from the top file */
 508                 for(i=a_tp0; i<a_tp1; i++)
 509                 {
 510                     copy_rvec(x_mol[i-a_tp0],state->x[i]);
 511                 }
 512                 /* Rotate the molecule randomly */
 513                 rotate_conf(a_tp1-a_tp0,state->x+a_tp0,NULL,
 514                             2*M_PI*gmx_rng_uniform_real(tpi_rand),
 515                             2*M_PI*gmx_rng_uniform_real(tpi_rand),
 516                             2*M_PI*gmx_rng_uniform_real(tpi_rand));
 517                 /* Shift to the insertion location */
 518                 for(i=a_tp0; i<a_tp1; i++)
 519                 {
 520                     rvec_inc(state->x[i],x_tp);
 521                 }
 522             }
 523
 524             /* Check if this insertion belongs to this node */
 525             bOurStep = TRUE;
 526             if (PAR(cr))
 527             {
 528                 switch (inputrec->eI)
 529                 {
 530                 case eiTPI:
 531                     bOurStep = ((step / inputrec->nstlist) % nnodes == cr->nodeid);
 532                     break;
 533                 case eiTPIC:
 534                     bOurStep = (step % nnodes == cr->nodeid);
 535                     break;
 536                 default:
 537                     gmx_fatal(FARGS,"Unknown integrator %s",ei_names[inputrec->eI]);
 538                 }
 539             }
 540             if (bOurStep)
 541             {
 542                 /* Clear some matrix variables  */
 543                 clear_mat(force_vir);
 544                 clear_mat(shake_vir);
 545                 clear_mat(vir);
 546                 clear_mat(pres);
 547
 548                 /* Set the charge group center of mass of the test particle */
 549                 copy_rvec(x_init,fr->cg_cm[top->cgs.nr-1]);
 550
 551                 /* Calc energy (no forces) on new positions.
 552                  * Since we only need the intermolecular energy
 553                  * and the RF exclusion terms of the inserted molecule occur
 554                  * within a single charge group we can pass NULL for the graph.
 555                  * This also avoids shifts that would move charge groups
 556                  * out of the box.
 557                  *
 558                  * Some checks above ensure than we can not have
 559                  * twin-range interactions together with nstlist > 1,
 560                  * therefore we do not need to remember the LR energies.
 561                  */
 562                 /* Make do_force do a single node force calculation */
 563                 cr->nnodes = 1;
 564                 do_force(fplog,cr,inputrec,
 565                          step,nrnb,wcycle,top,top_global,&top_global->groups,
 566                          state->box,state->x,&state->hist,
 567                          f,force_vir,mdatoms,enerd,fcd,
 568                          state->lambda,
 569                          NULL,fr,NULL,mu_tot,t,NULL,NULL,FALSE,
 570                          GMX_FORCE_NONBONDED | GMX_FORCE_ENERGY |
 571                          (bNS ? GMX_FORCE_DYNAMICBOX | GMX_FORCE_NS | GMX_FORCE_DOLR : 0) |
 572                          (bStateChanged ? GMX_FORCE_STATECHANGED : 0));
 573                 cr->nnodes = nnodes;
 574                 bStateChanged = FALSE;
 575                 bNS = FALSE;
 576
 577                 /* Calculate long range corrections to pressure and energy */
 578                 calc_dispcorr(fplog,inputrec,fr,step,top_global->natoms,state->box,
 579                               lambda,pres,vir,&prescorr,&enercorr,&dvdlcorr);
 580                 /* figure out how to rearrange the next 4 lines MRS 8/4/2009 */
 581                 enerd->term[F_DISPCORR] = enercorr;
 582                 enerd->term[F_EPOT] += enercorr;
 583                 enerd->term[F_PRES] += prescorr;
 584                 enerd->term[F_DVDL_VDW] += dvdlcorr;
 585
 586                 epot = enerd->term[F_EPOT];
 587                 bEnergyOutOfBounds = FALSE;
 588 #ifdef GMX_X86_SSE2
 589                 /* With SSE the energy can overflow, check for this */
 590                 if (gmx_mm_check_and_reset_overflow())
 591                 {
 592                     if (debug)
 593                     {
 594                         fprintf(debug,"Found an SSE overflow, assuming the energy is out of bounds\n");
 595                     }
 596                     bEnergyOutOfBounds = TRUE;
 597                 }
 598 #endif
 599                 /* If the compiler doesn't optimize this check away
 600                  * we catch the NAN energies.
 601                  * The epot>GMX_REAL_MAX check catches inf values,
 602                  * which should nicely result in embU=0 through the exp below,
 603                  * but it does not hurt to check anyhow.
 604                  */
 605                 /* Non-bonded Interaction usually diverge at r=0.
 606                  * With tabulated interaction functions the first few entries
 607                  * should be capped in a consistent fashion between
 608                  * repulsion, dispersion and Coulomb to avoid accidental
 609                  * negative values in the total energy.
 610                  * The table generation code in tables.c does this.
 611                  * With user tbales the user should take care of this.
 612                  */
 613                 if (epot != epot || epot > GMX_REAL_MAX)
 614                 {
 615                     bEnergyOutOfBounds = TRUE;
 616                 }
 617                 if (bEnergyOutOfBounds)
 618                 {
 619                     if (debug)
 620                     {
 621                         fprintf(debug,"\n  time %.3f, step %d: non-finite energy %f, using exp(-bU)=0\n",t,step,epot);
 622                     }
 623                     embU = 0;
 624                 }
 625                 else
 626                 {
 627                     embU = exp(-beta*epot);
 628                     sum_embU += embU;
 629                     /* Determine the weighted energy contributions of each energy group */
 630                     e = 0;
 631                     sum_UgembU[e++] += epot*embU;
 632                     if (fr->bBHAM)
 633                     {
 634                         for(i=0; i<ngid; i++)
 635                         {
 636                             sum_UgembU[e++] +=
 637                                 (enerd->grpp.ener[egBHAMSR][GID(i,gid_tp,ngid)] +
 638                                  enerd->grpp.ener[egBHAMLR][GID(i,gid_tp,ngid)])*embU;
 639                         }
 640                     }
 641                     else
 642                     {
 643                         for(i=0; i<ngid; i++)
 644                         {
 645                             sum_UgembU[e++] +=
 646                                 (enerd->grpp.ener[egLJSR][GID(i,gid_tp,ngid)] +
 647                                  enerd->grpp.ener[egLJLR][GID(i,gid_tp,ngid)])*embU;
 648                         }
 649                     }
 650                     if (bDispCorr)
 651                     {
 652                         sum_UgembU[e++] += enerd->term[F_DISPCORR]*embU;
 653                     }
 654                     if (bCharge)
 655                     {
 656                         for(i=0; i<ngid; i++)
 657                         {
 658                             sum_UgembU[e++] +=
 659                                 (enerd->grpp.ener[egCOULSR][GID(i,gid_tp,ngid)] +
 660                                  enerd->grpp.ener[egCOULLR][GID(i,gid_tp,ngid)])*embU;
 661                         }
 662                         if (bRFExcl)
 663                         {
 664                             sum_UgembU[e++] += enerd->term[F_RF_EXCL]*embU;
 665                         }
 666                         if (EEL_FULL(fr->eeltype))
 667                         {
 668                             sum_UgembU[e++] += enerd->term[F_COUL_RECIP]*embU;
 669                         }
 670                     }
 671                 }
 672
 673                 if (embU == 0 || beta*epot > bU_bin_limit)
 674                 {
 675                     bin[0]++;
 676                 }
 677                 else
 678                 {
 679                     i = (int)((bU_logV_bin_limit
 680                                - (beta*epot - logV + refvolshift))*invbinw
 681                               + 0.5);
 682                     if (i < 0)
 683                     {
 684                         i = 0;
 685                     }
 686                     if (i >= nbin)
 687                     {
 688                         realloc_bins(&bin,&nbin,i+10);
 689                     }
 690                     bin[i]++;
 691                 }
 692
 693                 if (debug)
 694                 {
 695                     fprintf(debug,"TPI %7d %12.5e %12.5f %12.5f %12.5f\n",
 696                             step,epot,x_tp[XX],x_tp[YY],x_tp[ZZ]);
 697                 }
 698
 699                 if (dump_pdb && epot <= dump_ener)
 700                 {
 701                     sprintf(str,"t%g_step%d.pdb",t,step);
 702                     sprintf(str2,"t: %f step %d ener: %f",t,step,epot);
 703                     write_sto_conf_mtop(str,str2,top_global,state->x,state->v,
 704                                         inputrec->ePBC,state->box);
 705                 }
 706             }
 707         }
 708
 709         if (PAR(cr))
 710         {
 711             /* When running in parallel sum the energies over the processes */
 712             gmx_sumd(1,    &sum_embU, cr);
 713             gmx_sumd(nener,sum_UgembU,cr);
 714         }
 715
 716         frame++;
 717         V_all += V;
 718         VembU_all += V*sum_embU/nsteps;
 719
 720         if (fp_tpi)
 721         {
 722             if (bVerbose || frame%10==0 || frame<10)
 723             {
 724                 fprintf(stderr,"mu %10.3e <mu> %10.3e\n",
 725                         -log(sum_embU/nsteps)/beta,-log(VembU_all/V_all)/beta);
 726             }
 727
 728             fprintf(fp_tpi,"%10.3f %12.5e %12.5e %12.5e %12.5e",
 729                     t,
 730                     VembU_all==0 ? 20/beta : -log(VembU_all/V_all)/beta,
 731                     sum_embU==0  ? 20/beta : -log(sum_embU/nsteps)/beta,
 732                     sum_embU/nsteps,V);
 733             for(e=0; e<nener; e++)
 734             {
 735                 fprintf(fp_tpi," %12.5e",sum_UgembU[e]/nsteps);
 736             }
 737             fprintf(fp_tpi,"\n");
 738             fflush(fp_tpi);
 739         }
 740
 741         bNotLastFrame = read_next_frame(oenv, status,&rerun_fr);
 742     } /* End of the loop  */
 743     runtime_end(runtime);
 744
 745     close_trj(status);
 746
 747     if (fp_tpi != NULL)
 748     {
 749         gmx_fio_fclose(fp_tpi);
 750     }
 751
 752     if (fplog != NULL)
 753     {
 754         fprintf(fplog,"\n");
 755         fprintf(fplog,"  <V>  = %12.5e nm^3\n",V_all/frame);
 756         fprintf(fplog,"  <mu> = %12.5e kJ/mol\n",-log(VembU_all/V_all)/beta);
 757     }
 758
 759     /* Write the Boltzmann factor histogram */
 760     if (PAR(cr))
 761     {
 762         /* When running in parallel sum the bins over the processes */
 763         i = nbin;
 764         global_max(cr,&i);
 765         realloc_bins(&bin,&nbin,i);
 766         gmx_sumd(nbin,bin,cr);
 767     }
 768     if (MASTER(cr))
 769     {
 770         fp_tpi = xvgropen(opt2fn("-tpid",nfile,fnm),
 771                           "TPI energy distribution",
 772                           "\\betaU - log(V/<V>)","count",oenv);
 773         sprintf(str,"number \\betaU > %g: %9.3e",bU_bin_limit,bin[0]);
 774         xvgr_subtitle(fp_tpi,str,oenv);
 775         xvgr_legend(fp_tpi,2,(const char **)tpid_leg,oenv);
 776         for(i=nbin-1; i>0; i--)
 777         {
 778             bUlogV = -i/invbinw + bU_logV_bin_limit - refvolshift + log(V_all/frame);
 779             fprintf(fp_tpi,"%6.2f %10d %12.5e\n",
 780                     bUlogV,
 781                     (int)(bin[i]+0.5),
 782                     bin[i]*exp(-bUlogV)*V_all/VembU_all);
 783         }
 784         gmx_fio_fclose(fp_tpi);
 785     }
 786     sfree(bin);
 787
 788     sfree(sum_UgembU);
 789
 790     runtime->nsteps_done = frame*inputrec->nsteps;
 791
 792     return 0;
 793 }