src/mdlib/sim_util.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 #ifdef GMX_CRAY_XT3
  41 #include<catamount/dclock.h>
  42 #endif
  43
  44
  45 #include <stdio.h>
  46 #include <time.h>
  47 #ifdef HAVE_SYS_TIME_H
  48 #include <sys/time.h>
  49 #endif
  50 #include <math.h>
  51 #include "typedefs.h"
  52 #include "string2.h"
  53 #include "gmxfio.h"
  54 #include "smalloc.h"
  55 #include "names.h"
  56 #include "confio.h"
  57 #include "mvdata.h"
  58 #include "txtdump.h"
  59 #include "pbc.h"
  60 #include "chargegroup.h"
  61 #include "vec.h"
  62 #include <time.h>
  63 #include "nrnb.h"
  64 #include "mshift.h"
  65 #include "mdrun.h"
  66 #include "update.h"
  67 #include "physics.h"
  68 #include "main.h"
  69 #include "mdatoms.h"
  70 #include "force.h"
  71 #include "bondf.h"
  72 #include "pme.h"
  73 #include "disre.h"
  74 #include "orires.h"
  75 #include "network.h"
  76 #include "calcmu.h"
  77 #include "constr.h"
  78 #include "xvgr.h"
  79 #include "trnio.h"
  80 #include "xtcio.h"
  81 #include "copyrite.h"
  82 #include "pull_rotation.h"
  83 #include "gmx_random.h"
  84 #include "mpelogging.h"
  85 #include "domdec.h"
  86 #include "partdec.h"
  87 #include "gmx_wallcycle.h"
  88 #include "genborn.h"
  89
  90 #ifdef GMX_LIB_MPI
  91 #include <mpi.h>
  92 #endif
  93 #ifdef GMX_THREAD_MPI
  94 #include "tmpi.h"
  95 #endif
  96
  97 #include "adress.h"
  98 #include "qmmm.h"
  99
 100 #if 0
 101 typedef struct gmx_timeprint {
 102
 103 } t_gmx_timeprint;
 104 #endif
 105
 106 /* Portable version of ctime_r implemented in src/gmxlib/string2.c, but we do not want it declared in public installed headers */
 107 char *
 108 gmx_ctime_r(const time_t *clock,char *buf, int n);
 109
 110
 111 double
 112 gmx_gettime()
 113 {
 114 #ifdef HAVE_GETTIMEOFDAY
 115         struct timeval t;
 116         double seconds;
 117
 118         gettimeofday(&t,NULL);
 119
 120         seconds = (double) t.tv_sec + 1e-6*(double)t.tv_usec;
 121
 122         return seconds;
 123 #else
 124         double  seconds;
 125
 126         seconds = time(NULL);
 127
 128         return seconds;
 129 #endif
 130 }
 131
 132
 133 #define difftime(end,start) ((double)(end)-(double)(start))
 134
 135 void print_time(FILE *out,gmx_runtime_t *runtime,gmx_large_int_t step,
 136                 t_inputrec *ir, t_commrec *cr)
 137 {
 138     time_t finish;
 139     char   timebuf[STRLEN];
 140     double dt;
 141     char buf[48];
 142
 143 #ifndef GMX_THREAD_MPI
 144     if (!PAR(cr))
 145 #endif
 146     {
 147         fprintf(out,"\r");
 148     }
 149     fprintf(out,"step %s",gmx_step_str(step,buf));
 150     if ((step >= ir->nstlist))
 151     {
 152         if ((ir->nstlist == 0) || ((step % ir->nstlist) == 0))
 153         {
 154             /* We have done a full cycle let's update time_per_step */
 155             runtime->last = gmx_gettime();
 156             dt = difftime(runtime->last,runtime->real);
 157             runtime->time_per_step = dt/(step - ir->init_step + 1);
 158         }
 159         dt = (ir->nsteps + ir->init_step - step)*runtime->time_per_step;
 160
 161         if (ir->nsteps >= 0)
 162         {
 163             if (dt >= 300)
 164             {
 165                 finish = (time_t) (runtime->last + dt);
 166                 gmx_ctime_r(&finish,timebuf,STRLEN);
 167                 sprintf(buf,"%s",timebuf);
 168                 buf[strlen(buf)-1]='\0';
 169                 fprintf(out,", will finish %s",buf);
 170             }
 171             else
 172                 fprintf(out,", remaining runtime: %5d s          ",(int)dt);
 173         }
 174         else
 175         {
 176             fprintf(out," performance: %.1f ns/day    ",
 177                     ir->delta_t/1000*24*60*60/runtime->time_per_step);
 178         }
 179     }
 180 #ifndef GMX_THREAD_MPI
 181     if (PAR(cr))
 182     {
 183         fprintf(out,"\n");
 184     }
 185 #endif
 186
 187     fflush(out);
 188 }
 189
 190 #ifdef NO_CLOCK
 191 #define clock() -1
 192 #endif
 193
 194 static double set_proctime(gmx_runtime_t *runtime)
 195 {
 196     double diff;
 197 #ifdef GMX_CRAY_XT3
 198     double prev;
 199
 200     prev = runtime->proc;
 201     runtime->proc = dclock();
 202
 203     diff = runtime->proc - prev;
 204 #else
 205     clock_t prev;
 206
 207     prev = runtime->proc;
 208     runtime->proc = clock();
 209
 210     diff = (double)(runtime->proc - prev)/(double)CLOCKS_PER_SEC;
 211 #endif
 212     if (diff < 0)
 213     {
 214         /* The counter has probably looped, ignore this data */
 215         diff = 0;
 216     }
 217
 218     return diff;
 219 }
 220
 221 void runtime_start(gmx_runtime_t *runtime)
 222 {
 223     runtime->real = gmx_gettime();
 224     runtime->proc          = 0;
 225     set_proctime(runtime);
 226     runtime->realtime      = 0;
 227     runtime->proctime      = 0;
 228     runtime->last          = 0;
 229     runtime->time_per_step = 0;
 230 }
 231
 232 void runtime_end(gmx_runtime_t *runtime)
 233 {
 234     double now;
 235
 236     now = gmx_gettime();
 237
 238     runtime->proctime += set_proctime(runtime);
 239     runtime->realtime  = now - runtime->real;
 240     runtime->real      = now;
 241 }
 242
 243 void runtime_upd_proc(gmx_runtime_t *runtime)
 244 {
 245     runtime->proctime += set_proctime(runtime);
 246 }
 247
 248 void print_date_and_time(FILE *fplog,int nodeid,const char *title,
 249                          const gmx_runtime_t *runtime)
 250 {
 251     int i;
 252     char timebuf[STRLEN];
 253     char time_string[STRLEN];
 254     time_t tmptime;
 255
 256     if (fplog)
 257     {
 258         if (runtime != NULL)
 259         {
 260             tmptime = (time_t) runtime->real;
 261             gmx_ctime_r(&tmptime,timebuf,STRLEN);
 262         }
 263         else
 264         {
 265             tmptime = (time_t) gmx_gettime();
 266             gmx_ctime_r(&tmptime,timebuf,STRLEN);
 267         }
 268         for(i=0; timebuf[i]>=' '; i++)
 269         {
 270             time_string[i]=timebuf[i];
 271         }
 272         time_string[i]='\0';
 273
 274         fprintf(fplog,"%s on node %d %s\n",title,nodeid,time_string);
 275     }
 276 }
 277
 278 static void sum_forces(int start,int end,rvec f[],rvec flr[])
 279 {
 280   int i;
 281
 282   if (gmx_debug_at) {
 283     pr_rvecs(debug,0,"fsr",f+start,end-start);
 284     pr_rvecs(debug,0,"flr",flr+start,end-start);
 285   }
 286   for(i=start; (i<end); i++)
 287     rvec_inc(f[i],flr[i]);
 288 }
 289
 290 /*
 291  * calc_f_el calculates forces due to an electric field.
 292  *
 293  * force is kJ mol^-1 nm^-1 = e * kJ mol^-1 nm^-1 / e
 294  *
 295  * Et[] contains the parameters for the time dependent
 296  * part of the field (not yet used).
 297  * Ex[] contains the parameters for
 298  * the spatial dependent part of the field. You can have cool periodic
 299  * fields in principle, but only a constant field is supported
 300  * now.
 301  * The function should return the energy due to the electric field
 302  * (if any) but for now returns 0.
 303  *
 304  * WARNING:
 305  * There can be problems with the virial.
 306  * Since the field is not self-consistent this is unavoidable.
 307  * For neutral molecules the virial is correct within this approximation.
 308  * For neutral systems with many charged molecules the error is small.
 309  * But for systems with a net charge or a few charged molecules
 310  * the error can be significant when the field is high.
 311  * Solution: implement a self-consitent electric field into PME.
 312  */
 313 static void calc_f_el(FILE *fp,int  start,int homenr,
 314                       real charge[],rvec x[],rvec f[],
 315                       t_cosines Ex[],t_cosines Et[],double t)
 316 {
 317     rvec Ext;
 318     real t0;
 319     int  i,m;
 320
 321     for(m=0; (m<DIM); m++)
 322     {
 323         if (Et[m].n > 0)
 324         {
 325             if (Et[m].n == 3)
 326             {
 327                 t0 = Et[m].a[1];
 328                 Ext[m] = cos(Et[m].a[0]*(t-t0))*exp(-sqr(t-t0)/(2.0*sqr(Et[m].a[2])));
 329             }
 330             else
 331             {
 332                 Ext[m] = cos(Et[m].a[0]*t);
 333             }
 334         }
 335         else
 336         {
 337             Ext[m] = 1.0;
 338         }
 339         if (Ex[m].n > 0)
 340         {
 341             /* Convert the field strength from V/nm to MD-units */
 342             Ext[m] *= Ex[m].a[0]*FIELDFAC;
 343             for(i=start; (i<start+homenr); i++)
 344                 f[i][m] += charge[i]*Ext[m];
 345         }
 346         else
 347         {
 348             Ext[m] = 0;
 349         }
 350     }
 351     if (fp != NULL)
 352     {
 353         fprintf(fp,"%10g  %10g  %10g  %10g #FIELD\n",t,
 354                 Ext[XX]/FIELDFAC,Ext[YY]/FIELDFAC,Ext[ZZ]/FIELDFAC);
 355     }
 356 }
 357
 358 static void calc_virial(FILE *fplog,int start,int homenr,rvec x[],rvec f[],
 359                         tensor vir_part,t_graph *graph,matrix box,
 360                         t_nrnb *nrnb,const t_forcerec *fr,int ePBC)
 361 {
 362   int i,j;
 363   tensor virtest;
 364
 365   /* The short-range virial from surrounding boxes */
 366   clear_mat(vir_part);
 367   calc_vir(fplog,SHIFTS,fr->shift_vec,fr->fshift,vir_part,ePBC==epbcSCREW,box);
 368   inc_nrnb(nrnb,eNR_VIRIAL,SHIFTS);
 369
 370   /* Calculate partial virial, for local atoms only, based on short range.
 371    * Total virial is computed in global_stat, called from do_md
 372    */
 373   f_calc_vir(fplog,start,start+homenr,x,f,vir_part,graph,box);
 374   inc_nrnb(nrnb,eNR_VIRIAL,homenr);
 375
 376   /* Add position restraint contribution */
 377   for(i=0; i<DIM; i++) {
 378     vir_part[i][i] += fr->vir_diag_posres[i];
 379   }
 380
 381   /* Add wall contribution */
 382   for(i=0; i<DIM; i++) {
 383     vir_part[i][ZZ] += fr->vir_wall_z[i];
 384   }
 385
 386   if (debug)
 387     pr_rvecs(debug,0,"vir_part",vir_part,DIM);
 388 }
 389
 390 static void print_large_forces(FILE *fp,t_mdatoms *md,t_commrec *cr,
 391                                gmx_large_int_t step,real pforce,rvec *x,rvec *f)
 392 {
 393   int  i;
 394   real pf2,fn2;
 395   char buf[STEPSTRSIZE];
 396
 397   pf2 = sqr(pforce);
 398   for(i=md->start; i<md->start+md->homenr; i++) {
 399     fn2 = norm2(f[i]);
 400     /* We also catch NAN, if the compiler does not optimize this away. */
 401     if (fn2 >= pf2 || fn2 != fn2) {
 402       fprintf(fp,"step %s  atom %6d  x %8.3f %8.3f %8.3f  force %12.5e\n",
 403               gmx_step_str(step,buf),
 404               ddglatnr(cr->dd,i),x[i][XX],x[i][YY],x[i][ZZ],sqrt(fn2));
 405     }
 406   }
 407 }
 408
 409 void do_force(FILE *fplog,t_commrec *cr,
 410               t_inputrec *inputrec,
 411               gmx_large_int_t step,t_nrnb *nrnb,gmx_wallcycle_t wcycle,
 412               gmx_localtop_t *top,
 413               gmx_mtop_t *mtop,
 414               gmx_groups_t *groups,
 415               matrix box,rvec x[],history_t *hist,
 416               rvec f[],
 417               tensor vir_force,
 418               t_mdatoms *mdatoms,
 419               gmx_enerdata_t *enerd,t_fcdata *fcd,
 420               real *lambda,t_graph *graph,
 421               t_forcerec *fr,gmx_vsite_t *vsite,rvec mu_tot,
 422               double t,FILE *field,gmx_edsam_t ed,
 423               gmx_bool bBornRadii,
 424               int flags)
 425 {
 426     int    cg0,cg1,i,j;
 427     int    start,homenr;
 428     double mu[2*DIM];
 429     gmx_bool   bSepDVDL,bStateChanged,bNS,bFillGrid,bCalcCGCM,bBS;
 430     gmx_bool   bDoLongRange,bDoForces,bSepLRF;
 431     gmx_bool   bDoAdressWF;
 432     matrix boxs;
 433     real   e,v,dvdlambda[efptNR];
 434     real   dvdl_dum,lambda_dum;
 435     t_pbc  pbc;
 436     float  cycles_ppdpme,cycles_pme,cycles_seppme,cycles_force;
 437
 438     start  = mdatoms->start;
 439     homenr = mdatoms->homenr;
 440
 441     bSepDVDL = (fr->bSepDVDL && do_per_step(step,inputrec->nstlog));
 442
 443     clear_mat(vir_force);
 444
 445     if (PARTDECOMP(cr))
 446     {
 447         pd_cg_range(cr,&cg0,&cg1);
 448     }
 449     else
 450     {
 451         cg0 = 0;
 452         if (DOMAINDECOMP(cr))
 453         {
 454             cg1 = cr->dd->ncg_tot;
 455         }
 456         else
 457         {
 458             cg1 = top->cgs.nr;
 459         }
 460         if (fr->n_tpi > 0)
 461         {
 462             cg1--;
 463         }
 464     }
 465
 466     bStateChanged = (flags & GMX_FORCE_STATECHANGED);
 467     bNS           = (flags & GMX_FORCE_NS) && (fr->bAllvsAll==FALSE);
 468     bFillGrid     = (bNS && bStateChanged);
 469     bCalcCGCM     = (bFillGrid && !DOMAINDECOMP(cr));
 470     bDoLongRange  = (fr->bTwinRange && bNS && (flags & GMX_FORCE_DOLR));
 471     bDoForces     = (flags & GMX_FORCE_FORCES);
 472     bSepLRF       = (bDoLongRange && bDoForces && (flags & GMX_FORCE_SEPLRF));
 473     /* should probably move this to the forcerec since it doesn't change */
 474     bDoAdressWF   = ((fr->adress_type!=eAdressOff));
 475
 476     if (bStateChanged)
 477     {
 478         update_forcerec(fplog,fr,box);
 479
 480         /* Calculate total (local) dipole moment in a temporary common array.
 481          * This makes it possible to sum them over nodes faster.
 482          */
 483         calc_mu(start,homenr,
 484                 x,mdatoms->chargeA,mdatoms->chargeB,mdatoms->nChargePerturbed,
 485                 mu,mu+DIM);
 486     }
 487
 488   if (fr->ePBC != epbcNONE) {
 489     /* Compute shift vectors every step,
 490      * because of pressure coupling or box deformation!
 491      */
 492     if ((flags & GMX_FORCE_DYNAMICBOX) && bStateChanged)
 493       calc_shifts(box,fr->shift_vec);
 494
 495     if (bCalcCGCM) {
 496       put_charge_groups_in_box(fplog,cg0,cg1,fr->ePBC,box,
 497                                &(top->cgs),x,fr->cg_cm);
 498       inc_nrnb(nrnb,eNR_CGCM,homenr);
 499       inc_nrnb(nrnb,eNR_RESETX,cg1-cg0);
 500     }
 501     else if (EI_ENERGY_MINIMIZATION(inputrec->eI) && graph) {
 502       unshift_self(graph,box,x);
 503     }
 504   }
 505   else if (bCalcCGCM) {
 506     calc_cgcm(fplog,cg0,cg1,&(top->cgs),x,fr->cg_cm);
 507     inc_nrnb(nrnb,eNR_CGCM,homenr);
 508   }
 509
 510   if (bCalcCGCM) {
 511     if (PAR(cr)) {
 512       move_cgcm(fplog,cr,fr->cg_cm);
 513     }
 514     if (gmx_debug_at)
 515       pr_rvecs(debug,0,"cgcm",fr->cg_cm,top->cgs.nr);
 516   }
 517
 518 #ifdef GMX_MPI
 519   if (!(cr->duty & DUTY_PME)) {
 520     /* Send particle coordinates to the pme nodes.
 521      * Since this is only implemented for domain decomposition
 522      * and domain decomposition does not use the graph,
 523      * we do not need to worry about shifting.
 524      */
 525
 526     wallcycle_start(wcycle,ewcPP_PMESENDX);
 527     GMX_MPE_LOG(ev_send_coordinates_start);
 528
 529     bBS = (inputrec->nwall == 2);
 530     if (bBS) {
 531       copy_mat(box,boxs);
 532       svmul(inputrec->wall_ewald_zfac,boxs[ZZ],boxs[ZZ]);
 533     }
 534
 535     gmx_pme_send_x(cr,bBS ? boxs : box,x,
 536                    mdatoms->nChargePerturbed,lambda[efptCOUL],
 537                    ( flags & GMX_FORCE_VIRIAL),step);
 538
 539     GMX_MPE_LOG(ev_send_coordinates_finish);
 540     wallcycle_stop(wcycle,ewcPP_PMESENDX);
 541   }
 542 #endif /* GMX_MPI */
 543
 544     /* Communicate coordinates and sum dipole if necessary */
 545     if (PAR(cr))
 546     {
 547         wallcycle_start(wcycle,ewcMOVEX);
 548         if (DOMAINDECOMP(cr))
 549         {
 550             dd_move_x(cr->dd,box,x);
 551         }
 552         else
 553         {
 554             move_x(fplog,cr,GMX_LEFT,GMX_RIGHT,x,nrnb);
 555         }
 556         /* When we don't need the total dipole we sum it in global_stat */
 557         if (bStateChanged && NEED_MUTOT(*inputrec))
 558         {
 559             gmx_sumd(2*DIM,mu,cr);
 560         }
 561         wallcycle_stop(wcycle,ewcMOVEX);
 562     }
 563     if (bStateChanged)
 564     {
 565
 566         /* update adress weight beforehand */
 567         if(bDoAdressWF)
 568         {
 569             /* need pbc for adress weight calculation with pbc_dx */
 570             set_pbc(&pbc,inputrec->ePBC,box);
 571             if(fr->adress_site == eAdressSITEcog)
 572             {
 573                 update_adress_weights_cog(top->idef.iparams,top->idef.il,x,fr,mdatoms,
 574                                           inputrec->ePBC==epbcNONE ? NULL : &pbc);
 575             }
 576             else if (fr->adress_site == eAdressSITEcom)
 577             {
 578                 update_adress_weights_com(fplog,cg0,cg1,&(top->cgs),x,fr,mdatoms,
 579                                           inputrec->ePBC==epbcNONE ? NULL : &pbc);
 580             }
 581             else if (fr->adress_site == eAdressSITEatomatom){
 582                 update_adress_weights_atom_per_atom(cg0,cg1,&(top->cgs),x,fr,mdatoms,
 583                                           inputrec->ePBC==epbcNONE ? NULL : &pbc);
 584             }
 585             else
 586             {
 587                 update_adress_weights_atom(cg0,cg1,&(top->cgs),x,fr,mdatoms,
 588                                            inputrec->ePBC==epbcNONE ? NULL : &pbc);
 589             }
 590         }
 591
 592         for(i=0; i<2; i++)
 593         {
 594             for(j=0;j<DIM;j++)
 595             {
 596                 fr->mu_tot[i][j] = mu[i*DIM + j];
 597             }
 598         }
 599     }
 600     if (fr->efep == efepNO)
 601     {
 602         copy_rvec(fr->mu_tot[0],mu_tot);
 603     }
 604     else
 605     {
 606         for(j=0; j<DIM; j++)
 607         {
 608             mu_tot[j] =
 609                 (1.0 - lambda[efptCOUL])*fr->mu_tot[0][j] + lambda[efptCOUL]*fr->mu_tot[1][j];
 610         }
 611     }
 612
 613     /* Reset energies */
 614     reset_enerdata(&(inputrec->opts),fr,bNS,enerd,MASTER(cr));
 615     clear_rvecs(SHIFTS,fr->fshift);
 616
 617     if (bNS)
 618     {
 619         wallcycle_start(wcycle,ewcNS);
 620
 621         if (graph && bStateChanged)
 622         {
 623             /* Calculate intramolecular shift vectors to make molecules whole */
 624             mk_mshift(fplog,graph,fr->ePBC,box,x);
 625         }
 626
 627         /* Reset long range forces if necessary */
 628         if (fr->bTwinRange)
 629         {
 630             /* Reset the (long-range) forces if necessary */
 631             clear_rvecs(fr->natoms_force_constr,bSepLRF ? fr->f_twin : f);
 632         }
 633
 634         /* Do the actual neighbour searching and if twin range electrostatics
 635          * also do the calculation of long range forces and energies.
 636          */
 637         for (i=0;i<efptNR;i++) {dvdlambda[i] = 0;}
 638         ns(fplog,fr,x,box,
 639            groups,&(inputrec->opts),top,mdatoms,
 640            cr,nrnb,lambda,dvdlambda,&enerd->grpp,bFillGrid,
 641            bDoLongRange,bDoForces,bSepLRF ? fr->f_twin : f);
 642         if (bSepDVDL)
 643         {
 644             fprintf(fplog,sepdvdlformat,"LR non-bonded",0.0,dvdlambda);
 645         }
 646         enerd->dvdl_lin[efptVDW] += dvdlambda[efptVDW];
 647         enerd->dvdl_lin[efptCOUL] += dvdlambda[efptCOUL];
 648
 649         wallcycle_stop(wcycle,ewcNS);
 650     }
 651
 652     if (inputrec->implicit_solvent && bNS)
 653     {
 654         make_gb_nblist(cr,inputrec->gb_algorithm,inputrec->rlist,
 655                        x,box,fr,&top->idef,graph,fr->born);
 656     }
 657
 658     if (DOMAINDECOMP(cr))
 659     {
 660         if (!(cr->duty & DUTY_PME))
 661         {
 662             wallcycle_start(wcycle,ewcPPDURINGPME);
 663             dd_force_flop_start(cr->dd,nrnb);
 664         }
 665     }
 666
 667     if (inputrec->bRot)
 668     {
 669         /* Enforced rotation has its own cycle counter that starts after the collective
 670          * coordinates have been communicated. It is added to ddCyclF to allow
 671          * for proper load-balancing */
 672         wallcycle_start(wcycle,ewcROT);
 673         do_rotation(cr,inputrec,box,x,t,step,wcycle,bNS);
 674         wallcycle_stop(wcycle,ewcROT);
 675     }
 676
 677     /* Start the force cycle counter.
 678      * This counter is stopped in do_forcelow_level.
 679      * No parallel communication should occur while this counter is running,
 680      * since that will interfere with the dynamic load balancing.
 681      */
 682     wallcycle_start(wcycle,ewcFORCE);
 683
 684     if (bDoForces)
 685     {
 686         /* Reset forces for which the virial is calculated separately:
 687          * PME/Ewald forces if necessary */
 688         if (fr->bF_NoVirSum)
 689         {
 690             if (flags & GMX_FORCE_VIRIAL)
 691             {
 692                 fr->f_novirsum = fr->f_novirsum_alloc;
 693                 GMX_BARRIER(cr->mpi_comm_mygroup);
 694                 if (fr->bDomDec)
 695                 {
 696                     clear_rvecs(fr->f_novirsum_n,fr->f_novirsum);
 697                 }
 698                 else
 699                 {
 700                     clear_rvecs(homenr,fr->f_novirsum+start);
 701                 }
 702                 GMX_BARRIER(cr->mpi_comm_mygroup);
 703             }
 704             else
 705             {
 706                 /* We are not calculating the pressure so we do not need
 707                  * a separate array for forces that do not contribute
 708                  * to the pressure.
 709                  */
 710                 fr->f_novirsum = f;
 711             }
 712         }
 713
 714         if (bSepLRF)
 715         {
 716             /* Add the long range forces to the short range forces */
 717             for(i=0; i<fr->natoms_force_constr; i++)
 718             {
 719                 copy_rvec(fr->f_twin[i],f[i]);
 720             }
 721         }
 722         else if (!(fr->bTwinRange && bNS))
 723         {
 724             /* Clear the short-range forces */
 725             clear_rvecs(fr->natoms_force_constr,f);
 726         }
 727
 728         clear_rvec(fr->vir_diag_posres);
 729
 730         GMX_BARRIER(cr->mpi_comm_mygroup);
 731     }
 732     if (inputrec->ePull == epullCONSTRAINT)
 733     {
 734         clear_pull_forces(inputrec->pull);
 735     }
 736
 737     /* update QMMMrec, if necessary */
 738     if(fr->bQMMM)
 739     {
 740         update_QMMMrec(cr,fr,x,mdatoms,box,top);
 741     }
 742
 743     if ((flags & GMX_FORCE_BONDED) && top->idef.il[F_POSRES].nr > 0)
 744     {
 745         /* Position restraints always require full pbc. Check if we already did it for Adress */
 746         if(!(bStateChanged && bDoAdressWF))
 747         {
 748             set_pbc(&pbc,inputrec->ePBC,box);
 749         }
 750         v = posres(top->idef.il[F_POSRES].nr,top->idef.il[F_POSRES].iatoms,
 751                    top->idef.iparams_posres,
 752                    (const rvec*)x,fr->f_novirsum,fr->vir_diag_posres,
 753                    inputrec->ePBC==epbcNONE ? NULL : &pbc,lambda[efptRESTRAINT],&(dvdlambda[efptRESTRAINT]),
 754                    fr->rc_scaling,fr->ePBC,fr->posres_com,fr->posres_comB);
 755         if (bSepDVDL)
 756         {
 757             fprintf(fplog,sepdvdlformat,
 758                     interaction_function[F_POSRES].longname,v,dvdlambda);
 759         }
 760         enerd->term[F_POSRES] += v;
 761         /* This linear lambda dependence assumption is only correct
 762          * when only k depends on lambda,
 763          * not when the reference position depends on lambda.
 764          * grompp checks for this.  (verify this is still the case?)
 765          */
 766         enerd->dvdl_nonlin[efptRESTRAINT] += dvdlambda[efptRESTRAINT]; /* if just the force constant changes, this is linear,
 767                                                                           but we can't be sure w/o additional checking that is
 768                                                                           hard to do at this level of code. Otherwise,
 769                                                                           the dvdl is not differentiable */
 770         inc_nrnb(nrnb,eNR_POSRES,top->idef.il[F_POSRES].nr/2);
 771         if ((inputrec->fepvals->n_lambda > 0) && (flags & GMX_FORCE_DHDL))
 772         {
 773             for(i=0; i<enerd->n_lambda; i++)
 774             {
 775                 lambda_dum = (i==0 ? lambda[efptRESTRAINT] : inputrec->fepvals->all_lambda[efptRESTRAINT][i-1]);
 776                 v = posres(top->idef.il[F_POSRES].nr,top->idef.il[F_POSRES].iatoms,
 777                            top->idef.iparams_posres,
 778                            (const rvec*)x,NULL,NULL,
 779                            inputrec->ePBC==epbcNONE ? NULL : &pbc,lambda_dum,&dvdl_dum,
 780                            fr->rc_scaling,fr->ePBC,fr->posres_com,fr->posres_comB);
 781                 enerd->enerpart_lambda[i] += v;
 782             }
 783         }
 784    }
 785
 786     /* Compute the bonded and non-bonded energies and optionally forces */
 787     do_force_lowlevel(fplog,step,fr,inputrec,&(top->idef),
 788                       cr,nrnb,wcycle,mdatoms,&(inputrec->opts),
 789                       x,hist,f,enerd,fcd,mtop,top,fr->born,
 790                       &(top->atomtypes),bBornRadii,box,
 791                       inputrec->fepvals,lambda,graph,&(top->excls),fr->mu_tot,
 792                       flags,&cycles_pme);
 793
 794     cycles_force = wallcycle_stop(wcycle,ewcFORCE);
 795     GMX_BARRIER(cr->mpi_comm_mygroup);
 796
 797     if (ed)
 798     {
 799         do_flood(fplog,cr,x,f,ed,box,step);
 800     }
 801
 802     if (DOMAINDECOMP(cr))
 803     {
 804         dd_force_flop_stop(cr->dd,nrnb);
 805         if (wcycle)
 806         {
 807             dd_cycles_add(cr->dd,cycles_force-cycles_pme,ddCyclF);
 808         }
 809     }
 810
 811     if (bDoForces)
 812     {
 813         if (IR_ELEC_FIELD(*inputrec))
 814         {
 815             /* Compute forces due to electric field */
 816             calc_f_el(MASTER(cr) ? field : NULL,
 817                       start,homenr,mdatoms->chargeA,x,fr->f_novirsum,
 818                       inputrec->ex,inputrec->et,t);
 819         }
 820
 821         if (bDoAdressWF && fr->adress_icor == eAdressICThermoForce)
 822         {
 823             /* Compute thermodynamic force in hybrid AdResS region */
 824             adress_thermo_force(start,homenr,&(top->cgs),x,fr->f_novirsum,fr,mdatoms,
 825                                 inputrec->ePBC==epbcNONE ? NULL : &pbc);
 826         }
 827
 828         /* Communicate the forces */
 829         if (PAR(cr))
 830         {
 831             wallcycle_start(wcycle,ewcMOVEF);
 832             if (DOMAINDECOMP(cr))
 833             {
 834                 dd_move_f(cr->dd,f,fr->fshift);
 835                 /* Do we need to communicate the separate force array
 836                  * for terms that do not contribute to the single sum virial?
 837                  * Position restraints and electric fields do not introduce
 838                  * inter-cg forces, only full electrostatics methods do.
 839                  * When we do not calculate the virial, fr->f_novirsum = f,
 840                  * so we have already communicated these forces.
 841                  */
 842                 if (EEL_FULL(fr->eeltype) && cr->dd->n_intercg_excl &&
 843                     (flags & GMX_FORCE_VIRIAL))
 844                 {
 845                     dd_move_f(cr->dd,fr->f_novirsum,NULL);
 846                 }
 847                 if (bSepLRF)
 848                 {
 849                     /* We should not update the shift forces here,
 850                      * since f_twin is already included in f.
 851                      */
 852                     dd_move_f(cr->dd,fr->f_twin,NULL);
 853                 }
 854             }
 855             else
 856             {
 857                 pd_move_f(cr,f,nrnb);
 858                 if (bSepLRF)
 859                 {
 860                     pd_move_f(cr,fr->f_twin,nrnb);
 861                 }
 862             }
 863             wallcycle_stop(wcycle,ewcMOVEF);
 864         }
 865
 866         /* If we have NoVirSum forces, but we do not calculate the virial,
 867          * we sum fr->f_novirum=f later.
 868          */
 869         if (vsite && !(fr->bF_NoVirSum && !(flags & GMX_FORCE_VIRIAL)))
 870         {
 871             wallcycle_start(wcycle,ewcVSITESPREAD);
 872             spread_vsite_f(fplog,vsite,x,f,fr->fshift,FALSE,NULL,nrnb,
 873                            &top->idef,fr->ePBC,fr->bMolPBC,graph,box,cr);
 874             wallcycle_stop(wcycle,ewcVSITESPREAD);
 875
 876             if (bSepLRF)
 877             {
 878                 wallcycle_start(wcycle,ewcVSITESPREAD);
 879                 spread_vsite_f(fplog,vsite,x,fr->f_twin,NULL,FALSE,NULL,
 880                                nrnb,
 881                                &top->idef,fr->ePBC,fr->bMolPBC,graph,box,cr);
 882                 wallcycle_stop(wcycle,ewcVSITESPREAD);
 883             }
 884         }
 885
 886         if (flags & GMX_FORCE_VIRIAL)
 887         {
 888             /* Calculation of the virial must be done after vsites! */
 889             calc_virial(fplog,mdatoms->start,mdatoms->homenr,x,f,
 890                         vir_force,graph,box,nrnb,fr,inputrec->ePBC);
 891         }
 892     }
 893
 894     enerd->term[F_COM_PULL] = 0;
 895     if (inputrec->ePull == epullUMBRELLA || inputrec->ePull == epullCONST_F)
 896     {
 897         /* Calculate the center of mass forces, this requires communication,
 898          * which is why pull_potential is called close to other communication.
 899          * The virial contribution is calculated directly,
 900          * which is why we call pull_potential after calc_virial.
 901          */
 902         set_pbc(&pbc,inputrec->ePBC,box);
 903         dvdlambda[efptRESTRAINT] = 0;
 904         enerd->term[F_COM_PULL] +=
 905             pull_potential(inputrec->ePull,inputrec->pull,mdatoms,&pbc,
 906                            cr,t,lambda[efptRESTRAINT],x,f,vir_force,&(dvdlambda[efptRESTRAINT]));
 907         if (bSepDVDL)
 908         {
 909             fprintf(fplog,sepdvdlformat,"Com pull",enerd->term[F_COM_PULL],dvdlambda[efptRESTRAINT]);
 910         }
 911         enerd->dvdl_lin[efptRESTRAINT] += dvdlambda[efptRESTRAINT];
 912     }
 913
 914     /* Add the forces from enforced rotation potentials (if any) */
 915     if (inputrec->bRot)
 916     {
 917         wallcycle_start(wcycle,ewcROTadd);
 918         enerd->term[F_COM_PULL] += add_rot_forces(inputrec->rot, f, cr,step,t);
 919         wallcycle_stop(wcycle,ewcROTadd);
 920     }
 921
 922     if (PAR(cr) && !(cr->duty & DUTY_PME))
 923     {
 924         cycles_ppdpme = wallcycle_stop(wcycle,ewcPPDURINGPME);
 925         dd_cycles_add(cr->dd,cycles_ppdpme,ddCyclPPduringPME);
 926
 927         /* In case of node-splitting, the PP nodes receive the long-range
 928          * forces, virial and energy from the PME nodes here.
 929          */
 930         wallcycle_start(wcycle,ewcPP_PMEWAITRECVF);
 931         dvdlambda[efptCOUL] = 0;
 932         gmx_pme_receive_f(cr,fr->f_novirsum,fr->vir_el_recip,&e,&dvdlambda[efptCOUL],
 933                           &cycles_seppme);
 934         if (bSepDVDL)
 935         {
 936             fprintf(fplog,sepdvdlformat,"PME mesh",e,dvdlambda[efptCOUL]);
 937         }
 938         enerd->term[F_COUL_RECIP] += e;
 939         enerd->dvdl_lin[efptCOUL] += dvdlambda[efptCOUL];
 940         if (wcycle)
 941         {
 942             dd_cycles_add(cr->dd,cycles_seppme,ddCyclPME);
 943         }
 944         wallcycle_stop(wcycle,ewcPP_PMEWAITRECVF);
 945     }
 946
 947     if (bDoForces && fr->bF_NoVirSum)
 948     {
 949         if (vsite)
 950         {
 951             /* Spread the mesh force on virtual sites to the other particles...
 952              * This is parallellized. MPI communication is performed
 953              * if the constructing atoms aren't local.
 954              */
 955             wallcycle_start(wcycle,ewcVSITESPREAD);
 956             spread_vsite_f(fplog,vsite,x,fr->f_novirsum,NULL,
 957                            (flags & GMX_FORCE_VIRIAL),fr->vir_el_recip,
 958                            nrnb,
 959                            &top->idef,fr->ePBC,fr->bMolPBC,graph,box,cr);
 960             wallcycle_stop(wcycle,ewcVSITESPREAD);
 961         }
 962         if (flags & GMX_FORCE_VIRIAL)
 963         {
 964             /* Now add the forces, this is local */
 965             if (fr->bDomDec)
 966             {
 967                 sum_forces(0,fr->f_novirsum_n,f,fr->f_novirsum);
 968             }
 969             else
 970             {
 971                 sum_forces(start,start+homenr,f,fr->f_novirsum);
 972             }
 973             if (EEL_FULL(fr->eeltype))
 974             {
 975                 /* Add the mesh contribution to the virial */
 976                 m_add(vir_force,fr->vir_el_recip,vir_force);
 977             }
 978             if (debug)
 979             {
 980                 pr_rvecs(debug,0,"vir_force",vir_force,DIM);
 981             }
 982         }
 983     }
 984
 985     /* Sum the potential energy terms from group contributions */
 986     sum_epot(&(inputrec->opts),enerd);
 987
 988     if (fr->print_force >= 0 && bDoForces)
 989     {
 990         print_large_forces(stderr,mdatoms,cr,step,fr->print_force,x,f);
 991     }
 992 }
 993
 994 void do_constrain_first(FILE *fplog,gmx_constr_t constr,
 995                         t_inputrec *ir,t_mdatoms *md,
 996                         t_state *state,rvec *f,
 997                         t_graph *graph,t_commrec *cr,t_nrnb *nrnb,
 998                         t_forcerec *fr, gmx_localtop_t *top, tensor shake_vir)
 999 {
1000     int    i,m,start,end;
1001     gmx_large_int_t step;
1002     real   dt=ir->delta_t;
1003     real   dvdl_dum;
1004     rvec   *savex;
1005
1006     snew(savex,state->natoms);
1007
1008     start = md->start;
1009     end   = md->homenr + start;
1010
1011     if (debug)
1012         fprintf(debug,"vcm: start=%d, homenr=%d, end=%d\n",
1013                 start,md->homenr,end);
1014     /* Do a first constrain to reset particles... */
1015     step = ir->init_step;
1016     if (fplog)
1017     {
1018         char buf[STEPSTRSIZE];
1019         fprintf(fplog,"\nConstraining the starting coordinates (step %s)\n",
1020                 gmx_step_str(step,buf));
1021     }
1022     dvdl_dum = 0;
1023
1024     /* constrain the current position */
1025     constrain(NULL,TRUE,FALSE,constr,&(top->idef),
1026               ir,NULL,cr,step,0,md,
1027               state->x,state->x,NULL,
1028               state->box,state->lambda[efptBONDED],&dvdl_dum,
1029               NULL,NULL,nrnb,econqCoord,ir->epc==epcMTTK,state->veta,state->veta);
1030     if (EI_VV(ir->eI))
1031     {
1032         /* constrain the inital velocity, and save it */
1033         /* also may be useful if we need the ekin from the halfstep for velocity verlet */
1034         /* might not yet treat veta correctly */
1035         constrain(NULL,TRUE,FALSE,constr,&(top->idef),
1036                   ir,NULL,cr,step,0,md,
1037                   state->x,state->v,state->v,
1038                   state->box,state->lambda[efptBONDED],&dvdl_dum,
1039                   NULL,NULL,nrnb,econqVeloc,ir->epc==epcMTTK,state->veta,state->veta);
1040     }
1041     /* constrain the inital velocities at t-dt/2 */
1042     if (EI_STATE_VELOCITY(ir->eI) && ir->eI!=eiVV)
1043     {
1044         for(i=start; (i<end); i++)
1045         {
1046             for(m=0; (m<DIM); m++)
1047             {
1048                 /* Reverse the velocity */
1049                 state->v[i][m] = -state->v[i][m];
1050                 /* Store the position at t-dt in buf */
1051                 savex[i][m] = state->x[i][m] + dt*state->v[i][m];
1052             }
1053         }
1054     /* Shake the positions at t=-dt with the positions at t=0
1055      * as reference coordinates.
1056          */
1057         if (fplog)
1058         {
1059             char buf[STEPSTRSIZE];
1060             fprintf(fplog,"\nConstraining the coordinates at t0-dt (step %s)\n",
1061                     gmx_step_str(step,buf));
1062         }
1063         dvdl_dum = 0;
1064         constrain(NULL,TRUE,FALSE,constr,&(top->idef),
1065                   ir,NULL,cr,step,-1,md,
1066                   state->x,savex,NULL,
1067                   state->box,state->lambda[efptBONDED],&dvdl_dum,
1068                   state->v,NULL,nrnb,econqCoord,ir->epc==epcMTTK,state->veta,state->veta);
1069
1070         for(i=start; i<end; i++) {
1071             for(m=0; m<DIM; m++) {
1072                 /* Re-reverse the velocities */
1073                 state->v[i][m] = -state->v[i][m];
1074             }
1075         }
1076     }
1077     sfree(savex);
1078 }
1079
1080 void calc_enervirdiff(FILE *fplog,int eDispCorr,t_forcerec *fr)
1081 {
1082   double eners[2],virs[2],enersum,virsum,y0,f,g,h;
1083   double r0,r1,r,rc3,rc9,ea,eb,ec,pa,pb,pc,pd;
1084   double invscale,invscale2,invscale3;
1085   int    ri0,ri1,ri,i,offstart,offset;
1086   real   scale,*vdwtab;
1087
1088   fr->enershiftsix = 0;
1089   fr->enershifttwelve = 0;
1090   fr->enerdiffsix = 0;
1091   fr->enerdifftwelve = 0;
1092   fr->virdiffsix = 0;
1093   fr->virdifftwelve = 0;
1094
1095   if (eDispCorr != edispcNO) {
1096     for(i=0; i<2; i++) {
1097       eners[i] = 0;
1098       virs[i]  = 0;
1099     }
1100     if ((fr->vdwtype == evdwSWITCH) || (fr->vdwtype == evdwSHIFT)) {
1101       if (fr->rvdw_switch == 0)
1102         gmx_fatal(FARGS,
1103                   "With dispersion correction rvdw-switch can not be zero "
1104                   "for vdw-type = %s",evdw_names[fr->vdwtype]);
1105
1106       scale  = fr->nblists[0].tab.scale;
1107       vdwtab = fr->nblists[0].vdwtab;
1108
1109       /* Round the cut-offs to exact table values for precision */
1110       ri0 = floor(fr->rvdw_switch*scale);
1111       ri1 = ceil(fr->rvdw*scale);
1112       r0  = ri0/scale;
1113       r1  = ri1/scale;
1114       rc3 = r0*r0*r0;
1115       rc9  = rc3*rc3*rc3;
1116
1117       if (fr->vdwtype == evdwSHIFT) {
1118         /* Determine the constant energy shift below rvdw_switch */
1119         fr->enershiftsix    = (real)(-1.0/(rc3*rc3)) - vdwtab[8*ri0];
1120         fr->enershifttwelve = (real)( 1.0/(rc9*rc3)) - vdwtab[8*ri0 + 4];
1121       }
1122       /* Add the constant part from 0 to rvdw_switch.
1123        * This integration from 0 to rvdw_switch overcounts the number
1124        * of interactions by 1, as it also counts the self interaction.
1125        * We will correct for this later.
1126        */
1127       eners[0] += 4.0*M_PI*fr->enershiftsix*rc3/3.0;
1128       eners[1] += 4.0*M_PI*fr->enershifttwelve*rc3/3.0;
1129
1130       invscale = 1.0/(scale);
1131       invscale2 = invscale*invscale;
1132       invscale3 = invscale*invscale2;
1133
1134       /* following summation derived from cubic spline definition,
1135         Numerical Recipies in C, second edition, p. 113-116.  Exact
1136         for the cubic spline.  We first calculate the negative of
1137         the energy from rvdw to rvdw_switch, assuming that g(r)=1,
1138         and then add the more standard, abrupt cutoff correction to
1139         that result, yielding the long-range correction for a
1140         switched function.  We perform both the pressure and energy
1141         loops at the same time for simplicity, as the computational
1142         cost is low. */
1143
1144       for (i=0;i<2;i++) {
1145         enersum = 0.0; virsum = 0.0;
1146         if (i==0)
1147           offstart = 0;
1148         else
1149           offstart = 4;
1150         for (ri=ri0; ri<ri1; ri++) {
1151           r = ri*invscale;
1152           ea = invscale3;
1153           eb = 2.0*invscale2*r;
1154           ec = invscale*r*r;
1155
1156           pa = invscale3;
1157           pb = 3.0*invscale2*r;
1158           pc = 3.0*invscale*r*r;
1159           pd = r*r*r;
1160
1161           /* this "8" is from the packing in the vdwtab array - perhaps
1162             should be #define'ed? */
1163           offset = 8*ri + offstart;
1164           y0 = vdwtab[offset];
1165           f = vdwtab[offset+1];
1166           g = vdwtab[offset+2];
1167           h = vdwtab[offset+3];
1168
1169           enersum += y0*(ea/3 + eb/2 + ec) + f*(ea/4 + eb/3 + ec/2)+
1170             g*(ea/5 + eb/4 + ec/3) + h*(ea/6 + eb/5 + ec/4);
1171           virsum  +=  f*(pa/4 + pb/3 + pc/2 + pd) +
1172             2*g*(pa/5 + pb/4 + pc/3 + pd/2) + 3*h*(pa/6 + pb/5 + pc/4 + pd/3);
1173
1174         }
1175         enersum *= 4.0*M_PI;
1176         virsum  *= 4.0*M_PI;
1177         eners[i] -= enersum;
1178         virs[i]  -= virsum;
1179       }
1180
1181       /* now add the correction for rvdw_switch to infinity */
1182       eners[0] += -4.0*M_PI/(3.0*rc3);
1183       eners[1] +=  4.0*M_PI/(9.0*rc9);
1184       virs[0]  +=  8.0*M_PI/rc3;
1185       virs[1]  += -16.0*M_PI/(3.0*rc9);
1186     }
1187     else if ((fr->vdwtype == evdwCUT) || (fr->vdwtype == evdwUSER)) {
1188       if (fr->vdwtype == evdwUSER && fplog)
1189         fprintf(fplog,
1190                 "WARNING: using dispersion correction with user tables\n");
1191       rc3  = fr->rvdw*fr->rvdw*fr->rvdw;
1192       rc9  = rc3*rc3*rc3;
1193       eners[0] += -4.0*M_PI/(3.0*rc3);
1194       eners[1] +=  4.0*M_PI/(9.0*rc9);
1195       virs[0]  +=  8.0*M_PI/rc3;
1196       virs[1]  += -16.0*M_PI/(3.0*rc9);
1197     } else {
1198       gmx_fatal(FARGS,
1199                 "Dispersion correction is not implemented for vdw-type = %s",
1200                 evdw_names[fr->vdwtype]);
1201     }
1202     fr->enerdiffsix    = eners[0];
1203     fr->enerdifftwelve = eners[1];
1204     /* The 0.5 is due to the Gromacs definition of the virial */
1205     fr->virdiffsix     = 0.5*virs[0];
1206     fr->virdifftwelve  = 0.5*virs[1];
1207   }
1208 }
1209
1210 void calc_dispcorr(FILE *fplog,t_inputrec *ir,t_forcerec *fr,
1211                    gmx_large_int_t step,int natoms,
1212                    matrix box,real lambda,tensor pres,tensor virial,
1213                    real *prescorr, real *enercorr, real *dvdlcorr)
1214 {
1215     gmx_bool bCorrAll,bCorrPres;
1216     real dvdlambda,invvol,dens,ninter,avcsix,avctwelve,enerdiff,svir=0,spres=0;
1217     int  m;
1218
1219     *prescorr = 0;
1220     *enercorr = 0;
1221     *dvdlcorr = 0;
1222
1223     clear_mat(virial);
1224     clear_mat(pres);
1225
1226     if (ir->eDispCorr != edispcNO) {
1227         bCorrAll  = (ir->eDispCorr == edispcAllEner ||
1228                      ir->eDispCorr == edispcAllEnerPres);
1229         bCorrPres = (ir->eDispCorr == edispcEnerPres ||
1230                      ir->eDispCorr == edispcAllEnerPres);
1231
1232         invvol = 1/det(box);
1233         if (fr->n_tpi)
1234         {
1235             /* Only correct for the interactions with the inserted molecule */
1236             dens = (natoms - fr->n_tpi)*invvol;
1237             ninter = fr->n_tpi;
1238         }
1239         else
1240         {
1241             dens = natoms*invvol;
1242             ninter = 0.5*natoms;
1243         }
1244
1245         if (ir->efep == efepNO)
1246         {
1247             avcsix    = fr->avcsix[0];
1248             avctwelve = fr->avctwelve[0];
1249         }
1250         else
1251         {
1252             avcsix    = (1 - lambda)*fr->avcsix[0]    + lambda*fr->avcsix[1];
1253             avctwelve = (1 - lambda)*fr->avctwelve[0] + lambda*fr->avctwelve[1];
1254         }
1255
1256         enerdiff = ninter*(dens*fr->enerdiffsix - fr->enershiftsix);
1257         *enercorr += avcsix*enerdiff;
1258         dvdlambda = 0.0;
1259         if (ir->efep != efepNO)
1260         {
1261             dvdlambda += (fr->avcsix[1] - fr->avcsix[0])*enerdiff;
1262         }
1263         if (bCorrAll)
1264         {
1265             enerdiff = ninter*(dens*fr->enerdifftwelve - fr->enershifttwelve);
1266             *enercorr += avctwelve*enerdiff;
1267             if (fr->efep != efepNO)
1268             {
1269                 dvdlambda += (fr->avctwelve[1] - fr->avctwelve[0])*enerdiff;
1270             }
1271         }
1272
1273         if (bCorrPres)
1274         {
1275             svir = ninter*dens*avcsix*fr->virdiffsix/3.0;
1276             if (ir->eDispCorr == edispcAllEnerPres)
1277             {
1278                 svir += ninter*dens*avctwelve*fr->virdifftwelve/3.0;
1279             }
1280             /* The factor 2 is because of the Gromacs virial definition */
1281             spres = -2.0*invvol*svir*PRESFAC;
1282
1283             for(m=0; m<DIM; m++) {
1284                 virial[m][m] += svir;
1285                 pres[m][m] += spres;
1286             }
1287             *prescorr += spres;
1288         }
1289
1290         /* Can't currently control when it prints, for now, just print when degugging */
1291         if (debug)
1292         {
1293             if (bCorrAll) {
1294                 fprintf(debug,"Long Range LJ corr.: <C6> %10.4e, <C12> %10.4e\n",
1295                         avcsix,avctwelve);
1296             }
1297             if (bCorrPres)
1298             {
1299                 fprintf(debug,
1300                         "Long Range LJ corr.: Epot %10g, Pres: %10g, Vir: %10g\n",
1301                         *enercorr,spres,svir);
1302             }
1303             else
1304             {
1305                 fprintf(debug,"Long Range LJ corr.: Epot %10g\n",*enercorr);
1306             }
1307         }
1308
1309         if (fr->bSepDVDL && do_per_step(step,ir->nstlog))
1310         {
1311             fprintf(fplog,sepdvdlformat,"Dispersion correction",
1312                     *enercorr,dvdlambda);
1313         }
1314         if (fr->efep != efepNO)
1315         {
1316             *dvdlcorr += dvdlambda;
1317         }
1318     }
1319 }
1320
1321 void do_pbc_first(FILE *fplog,matrix box,t_forcerec *fr,
1322                   t_graph *graph,rvec x[])
1323 {
1324   if (fplog)
1325     fprintf(fplog,"Removing pbc first time\n");
1326   calc_shifts(box,fr->shift_vec);
1327   if (graph) {
1328     mk_mshift(fplog,graph,fr->ePBC,box,x);
1329     if (gmx_debug_at)
1330       p_graph(debug,"do_pbc_first 1",graph);
1331     shift_self(graph,box,x);
1332     /* By doing an extra mk_mshift the molecules that are broken
1333      * because they were e.g. imported from another software
1334      * will be made whole again. Such are the healing powers
1335      * of GROMACS.
1336      */
1337     mk_mshift(fplog,graph,fr->ePBC,box,x);
1338     if (gmx_debug_at)
1339       p_graph(debug,"do_pbc_first 2",graph);
1340   }
1341   if (fplog)
1342     fprintf(fplog,"Done rmpbc\n");
1343 }
1344
1345 static void low_do_pbc_mtop(FILE *fplog,int ePBC,matrix box,
1346                             gmx_mtop_t *mtop,rvec x[],
1347                             gmx_bool bFirst)
1348 {
1349   t_graph *graph;
1350   int mb,as,mol;
1351   gmx_molblock_t *molb;
1352
1353   if (bFirst && fplog)
1354     fprintf(fplog,"Removing pbc first time\n");
1355
1356   snew(graph,1);
1357   as = 0;
1358   for(mb=0; mb<mtop->nmolblock; mb++) {
1359     molb = &mtop->molblock[mb];
1360     if (molb->natoms_mol == 1 ||
1361         (!bFirst && mtop->moltype[molb->type].cgs.nr == 1)) {
1362       /* Just one atom or charge group in the molecule, no PBC required */
1363       as += molb->nmol*molb->natoms_mol;
1364     } else {
1365       /* Pass NULL iso fplog to avoid graph prints for each molecule type */
1366       mk_graph_ilist(NULL,mtop->moltype[molb->type].ilist,
1367                      0,molb->natoms_mol,FALSE,FALSE,graph);
1368
1369       for(mol=0; mol<molb->nmol; mol++) {
1370         mk_mshift(fplog,graph,ePBC,box,x+as);
1371
1372         shift_self(graph,box,x+as);
1373         /* The molecule is whole now.
1374          * We don't need the second mk_mshift call as in do_pbc_first,
1375          * since we no longer need this graph.
1376          */
1377
1378         as += molb->natoms_mol;
1379       }
1380       done_graph(graph);
1381     }
1382   }
1383   sfree(graph);
1384 }
1385
1386 void do_pbc_first_mtop(FILE *fplog,int ePBC,matrix box,
1387                        gmx_mtop_t *mtop,rvec x[])
1388 {
1389   low_do_pbc_mtop(fplog,ePBC,box,mtop,x,TRUE);
1390 }
1391
1392 void do_pbc_mtop(FILE *fplog,int ePBC,matrix box,
1393                  gmx_mtop_t *mtop,rvec x[])
1394 {
1395   low_do_pbc_mtop(fplog,ePBC,box,mtop,x,FALSE);
1396 }
1397
1398 void finish_run(FILE *fplog,t_commrec *cr,const char *confout,
1399                 t_inputrec *inputrec,
1400                 t_nrnb nrnb[],gmx_wallcycle_t wcycle,
1401                 gmx_runtime_t *runtime,
1402                 gmx_bool bWriteStat)
1403 {
1404   int    i,j;
1405   t_nrnb *nrnb_tot=NULL;
1406   real   delta_t;
1407   double nbfs,mflop;
1408   double cycles[ewcNR];
1409
1410   wallcycle_sum(cr,wcycle,cycles);
1411
1412   if (cr->nnodes > 1) {
1413     if (SIMMASTER(cr))
1414       snew(nrnb_tot,1);
1415 #ifdef GMX_MPI
1416     MPI_Reduce(nrnb->n,nrnb_tot->n,eNRNB,MPI_DOUBLE,MPI_SUM,
1417                MASTERRANK(cr),cr->mpi_comm_mysim);
1418 #endif
1419   } else {
1420     nrnb_tot = nrnb;
1421   }
1422
1423   if (SIMMASTER(cr)) {
1424     print_flop(fplog,nrnb_tot,&nbfs,&mflop);
1425     if (cr->nnodes > 1) {
1426       sfree(nrnb_tot);
1427     }
1428   }
1429
1430   if ((cr->duty & DUTY_PP) && DOMAINDECOMP(cr)) {
1431     print_dd_statistics(cr,inputrec,fplog);
1432   }
1433
1434 #ifdef GMX_MPI
1435     if (PARTDECOMP(cr))
1436     {
1437         if (MASTER(cr))
1438         {
1439             t_nrnb     *nrnb_all;
1440             int        s;
1441             MPI_Status stat;
1442
1443             snew(nrnb_all,cr->nnodes);
1444             nrnb_all[0] = *nrnb;
1445             for(s=1; s<cr->nnodes; s++)
1446             {
1447                 MPI_Recv(nrnb_all[s].n,eNRNB,MPI_DOUBLE,s,0,
1448                          cr->mpi_comm_mysim,&stat);
1449             }
1450             pr_load(fplog,cr,nrnb_all);
1451             sfree(nrnb_all);
1452         }
1453         else
1454         {
1455             MPI_Send(nrnb->n,eNRNB,MPI_DOUBLE,MASTERRANK(cr),0,
1456                      cr->mpi_comm_mysim);
1457         }
1458     }
1459 #endif
1460
1461   if (SIMMASTER(cr)) {
1462     wallcycle_print(fplog,cr->nnodes,cr->npmenodes,runtime->realtime,
1463                     wcycle,cycles);
1464
1465     if (EI_DYNAMICS(inputrec->eI)) {
1466       delta_t = inputrec->delta_t;
1467     } else {
1468       delta_t = 0;
1469     }
1470
1471     if (fplog) {
1472         print_perf(fplog,runtime->proctime,runtime->realtime,
1473                    cr->nnodes-cr->npmenodes,
1474                    runtime->nsteps_done,delta_t,nbfs,mflop);
1475     }
1476     if (bWriteStat) {
1477         print_perf(stderr,runtime->proctime,runtime->realtime,
1478                    cr->nnodes-cr->npmenodes,
1479                    runtime->nsteps_done,delta_t,nbfs,mflop);
1480     }
1481
1482     /*
1483     runtime=inputrec->nsteps*inputrec->delta_t;
1484     if (bWriteStat) {
1485       if (cr->nnodes == 1)
1486         fprintf(stderr,"\n\n");
1487       print_perf(stderr,nodetime,realtime,runtime,&ntot,
1488                  cr->nnodes-cr->npmenodes,FALSE);
1489     }
1490     wallcycle_print(fplog,cr->nnodes,cr->npmenodes,realtime,wcycle,cycles);
1491     print_perf(fplog,nodetime,realtime,runtime,&ntot,cr->nnodes-cr->npmenodes,
1492                TRUE);
1493     if (PARTDECOMP(cr))
1494       pr_load(fplog,cr,nrnb_all);
1495     if (cr->nnodes > 1)
1496       sfree(nrnb_all);
1497     */
1498   }
1499 }
1500
1501 extern void initialize_lambdas(FILE *fplog,t_inputrec *ir,int *fep_state,real *lambda,double *lam0)
1502 {
1503     /* this function works, but could probably use a logic rewrite to keep all the different
1504        types of efep straight. */
1505
1506     int i;
1507     t_lambda *fep = ir->fepvals;
1508
1509     if ((ir->efep==efepNO) && (ir->bSimTemp == FALSE)) {
1510         for (i=0;i<efptNR;i++)  {
1511             lambda[i] = 0.0;
1512             if (lam0)
1513             {
1514                 lam0[i] = 0.0;
1515             }
1516         }
1517         return;
1518     } else {
1519         *fep_state = fep->init_fep_state; /* this might overwrite the checkpoint
1520                                              if checkpoint is set -- a kludge is in for now
1521                                              to prevent this.*/
1522         for (i=0;i<efptNR;i++)
1523         {
1524             /* overwrite lambda state with init_lambda for now for backwards compatibility */
1525             if (fep->init_lambda>=0) /* if it's -1, it was never initializd */
1526             {
1527                 lambda[i] = fep->init_lambda;
1528                 if (lam0) {
1529                     lam0[i] = lambda[i];
1530                 }
1531             }
1532             else
1533             {
1534                 lambda[i] = fep->all_lambda[i][*fep_state];
1535                 if (lam0) {
1536                     lam0[i] = lambda[i];
1537                 }
1538             }
1539         }
1540         if (ir->bSimTemp) {
1541             /* need to rescale control temperatures to match current state */
1542             for (i=0;i<ir->opts.ngtc;i++) {
1543                 if (ir->opts.ref_t[i] > 0) {
1544                     ir->opts.ref_t[i] = ir->simtempvals->temperatures[*fep_state];
1545                 }
1546             }
1547         }
1548     }
1549
1550     /* Send to the log the information on the current lambdas */
1551     if (fplog != NULL)
1552     {
1553         fprintf(fplog,"Initial vector of lambda components:[ ");
1554         for (i=0;i<efptNR;i++)
1555         {
1556             fprintf(fplog,"%10.4f ",lambda[i]);
1557         }
1558         fprintf(fplog,"]\n");
1559     }
1560     return;
1561 }
1562
1563
1564 void init_md(FILE *fplog,
1565              t_commrec *cr,t_inputrec *ir,const output_env_t oenv,
1566              double *t,double *t0,
1567              real *lambda, int *fep_state, double *lam0,
1568              t_nrnb *nrnb,gmx_mtop_t *mtop,
1569              gmx_update_t *upd,
1570              int nfile,const t_filenm fnm[],
1571              gmx_mdoutf_t **outf,t_mdebin **mdebin,
1572              tensor force_vir,tensor shake_vir,rvec mu_tot,
1573              gmx_bool *bSimAnn,t_vcm **vcm, t_state *state, unsigned long Flags)
1574 {
1575     int  i,j,n;
1576     real tmpt,mod;
1577
1578     /* Initial values */
1579     *t = *t0       = ir->init_t;
1580
1581     *bSimAnn=FALSE;
1582     for(i=0;i<ir->opts.ngtc;i++)
1583     {
1584         /* set bSimAnn if any group is being annealed */
1585         if(ir->opts.annealing[i]!=eannNO)
1586         {
1587             *bSimAnn = TRUE;
1588         }
1589     }
1590     if (*bSimAnn)
1591     {
1592         update_annealing_target_temp(&(ir->opts),ir->init_t);
1593     }
1594
1595     /* Initialize lambda variables */
1596     initialize_lambdas(fplog,ir,fep_state,lambda,lam0);
1597
1598     if (upd)
1599     {
1600         *upd = init_update(fplog,ir);
1601     }
1602
1603
1604     if (vcm != NULL)
1605     {
1606         *vcm = init_vcm(fplog,&mtop->groups,ir);
1607     }
1608
1609     if (EI_DYNAMICS(ir->eI) && !(Flags & MD_APPENDFILES))
1610     {
1611         if (ir->etc == etcBERENDSEN)
1612         {
1613             please_cite(fplog,"Berendsen84a");
1614         }
1615         if (ir->etc == etcVRESCALE)
1616         {
1617             please_cite(fplog,"Bussi2007a");
1618         }
1619     }
1620
1621     init_nrnb(nrnb);
1622
1623     if (nfile != -1)
1624     {
1625         *outf = init_mdoutf(nfile,fnm,Flags,cr,ir,oenv);
1626
1627         *mdebin = init_mdebin((Flags & MD_APPENDFILES) ? NULL : (*outf)->fp_ene,
1628                               mtop,ir, (*outf)->fp_dhdl);
1629     }
1630
1631     if (ir->bAdress)
1632     {
1633       please_cite(fplog,"Fritsch12");
1634       please_cite(fplog,"Junghans10");
1635     }
1636     /* Initiate variables */
1637     clear_mat(force_vir);
1638     clear_mat(shake_vir);
1639     clear_rvec(mu_tot);
1640
1641     debug_gmx();
1642 }
1643
1644
1645
1646