2 * $Id: mdrun.c,v 1.139.2.9 2009/05/04 16:13:29 hess Exp $
4 * This source code is part of
8 * GROningen MAchine for Chemical Simulations
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.
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.
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.
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.
31 * For more info, check our website at http://www.gromacs.org
34 * Gallium Rubidium Oxygen Manganese Argon Carbon Silicon
52 #include "checkpoint.h"
77 #include "mpelogging.h"
85 #include "checkpoint.h"
86 #include "mtop_util.h"
89 #include "sighandler.h"
102 /* We use the same defines as in mvdata.c here */
103 #define block_bc(cr, d) gmx_bcast( sizeof(d), &(d),(cr))
104 #define nblock_bc(cr,nr,d) gmx_bcast((nr)*sizeof((d)[0]), (d),(cr))
105 #define snew_bc(cr,d,nr) { if (!MASTER(cr)) snew((d),(nr)); }
107 /* The following two variables and the signal_handler function
108 * is used from pme.c as well
137 int natoms; /*nr of atoms per lipid*/
138 int mol1; /*id of the first lipid molecule*/
159 int search_string(char *s,int ng,char ***gn)
163 for(i=0; (i<ng); i++)
164 if (gmx_strcasecmp(s,*gn[i]) == 0)
167 gmx_fatal(FARGS,"Group %s not found in indexfile.\nMaybe you have non-default groups in your mdp file, while not using the '-n' option of grompp.\nIn that case use the '-n' option.\n",s);
172 int get_mol_id(int at,int nmblock,gmx_molblock_t *mblock, int *type, int *block)
177 for(i=0;i<nmblock;i++)
179 if(at<(mblock[i].nmol*mblock[i].natoms_mol))
181 mol_id+=at/mblock[i].natoms_mol;
182 *type = mblock[i].type;
186 at-= mblock[i].nmol*mblock[i].natoms_mol;
187 mol_id+=mblock[i].nmol;
191 gmx_fatal(FARGS,"Something is wrong in mol ids, at %d, mol_id %d",at,mol_id);
196 int get_block(int mol_id,int nmblock,gmx_molblock_t *mblock)
201 for(i=0;i<nmblock;i++)
203 nmol+=mblock[i].nmol;
208 gmx_fatal(FARGS,"mol_id %d larger than total number of molecules %d.\n",mol_id,nmol);
213 int get_tpr_version(const char *infile)
220 fio = open_tpx(infile,"r");
221 gmx_fio_checktype(fio);
223 precision = sizeof(real);
225 gmx_fio_do_string(fio,buf);
226 if (strncmp(buf,"VERSION",7))
227 gmx_fatal(FARGS,"Can not read file %s,\n"
228 " this file is from a Gromacs version which is older than 2.0\n"
229 " Make a new one with grompp or use a gro or pdb file, if possible",
230 gmx_fio_getname(fio));
231 gmx_fio_do_int(fio,precision);
232 bDouble = (precision == sizeof(double));
233 if ((precision != sizeof(float)) && !bDouble)
234 gmx_fatal(FARGS,"Unknown precision in file %s: real is %d bytes "
235 "instead of %d or %d",
236 gmx_fio_getname(fio),precision,sizeof(float),sizeof(double));
237 gmx_fio_setprecision(fio,bDouble);
238 fprintf(stderr,"Reading file %s, %s (%s precision)\n",
239 gmx_fio_getname(fio),buf,bDouble ? "double" : "single");
241 gmx_fio_do_int(fio,fver);
248 void set_inbox(int natom, rvec *x)
253 tmp[XX]=tmp[YY]=tmp[ZZ]=0.0;
256 if(x[i][XX]<tmp[XX]) tmp[XX]=x[i][XX];
257 if(x[i][YY]<tmp[YY]) tmp[YY]=x[i][YY];
258 if(x[i][ZZ]<tmp[ZZ]) tmp[ZZ]=x[i][ZZ];
265 int get_mtype_list(t_block *at, gmx_mtop_t *mtop, t_block *tlist)
272 snew(tlist->index,at->nr);
273 for (i=0;i<at->nr;i++)
276 mol_id = get_mol_id(at->index[i],mtop->nmolblock,mtop->molblock,&type,&block);
279 if(tlist->index[j]==type)
284 tlist->index[nr]=type;
289 srenew(tlist->index,nr);
293 void check_types(t_block *ins_at,t_block *rest_at,gmx_mtop_t *mtop)
295 t_block *ins_mtype,*rest_mtype;
300 ins_mtype->nr = get_mtype_list(ins_at , mtop, ins_mtype );
301 rest_mtype->nr = get_mtype_list(rest_at, mtop, rest_mtype);
303 for(i=0;i<ins_mtype->nr;i++)
305 for(j=0;j<rest_mtype->nr;j++)
307 if(ins_mtype->index[i]==rest_mtype->index[j])
308 gmx_fatal(FARGS,"Moleculetype %s is found both in the group to insert and the rest of the system.\n"
309 "Because we need to exclude all interactions between the atoms in the group to\n"
310 "insert, the same moleculetype can not be used in both groups. Change the\n"
311 "moleculetype of the molecules %s in the inserted group. Do not forget to provide\n"
312 "an appropriate *.itp file",*(mtop->moltype[rest_mtype->index[j]].name),
313 *(mtop->moltype[rest_mtype->index[j]].name));
317 sfree(ins_mtype->index);
318 sfree(rest_mtype->index);
323 int init_ins_at(t_block *ins_at,t_block *rest_at,t_state *state, pos_ins_t *pos_ins,gmx_groups_t *groups,int ins_grp_id, real xy_max)
326 real x,xmin,xmax,y,ymin,ymax,z,zmin,zmax;
328 snew(rest_at->index,state->natoms);
330 xmin=xmax=state->x[ins_at->index[0]][XX];
331 ymin=ymax=state->x[ins_at->index[0]][YY];
332 zmin=zmax=state->x[ins_at->index[0]][ZZ];
334 for(i=0;i<state->natoms;i++)
336 gid = groups->grpnr[egcFREEZE][i];
337 if(groups->grps[egcFREEZE].nm_ind[gid]==ins_grp_id)
355 srenew(rest_at->index,c);
359 pos_ins->xmin[XX]=xmin-((xmax-xmin)*xy_max-(xmax-xmin))/2;
360 pos_ins->xmin[YY]=ymin-((ymax-ymin)*xy_max-(ymax-ymin))/2;
362 pos_ins->xmax[XX]=xmax+((xmax-xmin)*xy_max-(xmax-xmin))/2;
363 pos_ins->xmax[YY]=ymax+((ymax-ymin)*xy_max-(ymax-ymin))/2;
365 pos_ins->xmin[XX]=xmin;
366 pos_ins->xmin[YY]=ymin;
368 pos_ins->xmax[XX]=xmax;
369 pos_ins->xmax[YY]=ymax;
372 /* 6.0 is estimated thickness of bilayer */
373 if( (zmax-zmin) < 6.0 )
375 pos_ins->xmin[ZZ]=zmin+(zmax-zmin)/2.0-3.0;
376 pos_ins->xmax[ZZ]=zmin+(zmax-zmin)/2.0+3.0;
378 pos_ins->xmin[ZZ]=zmin;
379 pos_ins->xmax[ZZ]=zmax;
385 real est_prot_area(pos_ins_t *pos_ins,rvec *r,t_block *ins_at, mem_t *mem_p)
387 real x,y,dx=0.15,dy=0.15,area=0.0;
391 for(x=pos_ins->xmin[XX];x<pos_ins->xmax[XX];x+=dx)
393 for(y=pos_ins->xmin[YY];y<pos_ins->xmax[YY];y+=dy)
400 if ( (r[at][XX]>=x) && (r[at][XX]<x+dx) &&
401 (r[at][YY]>=y) && (r[at][YY]<y+dy) &&
402 (r[at][ZZ]>mem_p->zmin+1.0) && (r[at][ZZ]<mem_p->zmax-1.0) )
405 } while ( (c<ins_at->nr) && (add<0.5) );
414 void init_lip(matrix box, gmx_mtop_t *mtop, lip_t *lip)
420 mem_area = box[XX][XX]*box[YY][YY]-box[XX][YY]*box[YY][XX];
421 for(i=0;i<mtop->nmolblock;i++)
423 if(mtop->molblock[i].type == lip->id)
425 lip->nr=mtop->molblock[i].nmol;
426 lip->natoms=mtop->molblock[i].natoms_mol;
429 lip->area=2.0*mem_area/(double)lip->nr;
431 for (i=0;i<lip->id;i++)
432 mol1+=mtop->molblock[i].nmol;
436 int init_mem_at(mem_t *mem_p, gmx_mtop_t *mtop, rvec *r, matrix box, pos_ins_t *pos_ins)
438 int i,j,at,mol,nmol,nmolbox,count;
440 real z,zmin,zmax,mem_area;
446 mem_a=&(mem_p->mem_at);
447 snew(mol_id,mem_a->nr);
448 /* snew(index,mem_a->nr); */
449 zmin=pos_ins->xmax[ZZ];
450 zmax=pos_ins->xmin[ZZ];
451 for(i=0;i<mem_a->nr;i++)
454 if( (r[at][XX]>pos_ins->xmin[XX]) && (r[at][XX]<pos_ins->xmax[XX]) &&
455 (r[at][YY]>pos_ins->xmin[YY]) && (r[at][YY]<pos_ins->xmax[YY]) &&
456 (r[at][ZZ]>pos_ins->xmin[ZZ]) && (r[at][ZZ]<pos_ins->xmax[ZZ]) )
458 mol = get_mol_id(at,mtop->nmolblock,mtop->molblock,&type,&block);
475 /* index[count]=at;*/
482 mem_p->mol_id=mol_id;
483 /* srenew(index,count);*/
484 /* mem_p->mem_at.nr=count;*/
485 /* sfree(mem_p->mem_at.index);*/
486 /* mem_p->mem_at.index=index;*/
488 if((zmax-zmin)>(box[ZZ][ZZ]-0.5))
489 gmx_fatal(FARGS,"Something is wrong with your membrane. Max and min z values are %f and %f.\n"
490 "Maybe your membrane is not centered in the box, but located at the box edge in the z-direction,\n"
491 "so that one membrane is distributed over two periodic box images. Another possibility is that\n"
492 "your water layer is not thick enough.\n",zmax,zmin);
495 mem_p->zmed=(zmax-zmin)/2+zmin;
497 /*number of membrane molecules in protein box*/
498 nmolbox = count/mtop->molblock[block].natoms_mol;
499 /*mem_area = box[XX][XX]*box[YY][YY]-box[XX][YY]*box[YY][XX];
500 mem_p->lip_area = 2.0*mem_area/(double)mem_p->nmol;*/
501 mem_area = (pos_ins->xmax[XX]-pos_ins->xmin[XX])*(pos_ins->xmax[YY]-pos_ins->xmin[YY]);
502 mem_p->lip_area = 2.0*mem_area/(double)nmolbox;
504 return mem_p->mem_at.nr;
507 void init_resize(t_block *ins_at,rvec *r_ins,pos_ins_t *pos_ins,mem_t *mem_p,rvec *r, gmx_bool bALLOW_ASYMMETRY)
509 int i,j,at,c,outsidesum,gctr=0;
513 for (i=0;i<pos_ins->pieces;i++)
514 idxsum+=pos_ins->nidx[i];
515 if (idxsum!=ins_at->nr)
516 gmx_fatal(FARGS,"Piecewise sum of inserted atoms not same as size of group selected to insert.");
518 snew(pos_ins->geom_cent,pos_ins->pieces);
519 for (i=0;i<pos_ins->pieces;i++)
524 pos_ins->geom_cent[i][j]=0;
527 pos_ins->geom_cent[i][j]=0;
528 for (j=0;j<pos_ins->nidx[i];j++)
530 at=pos_ins->subindex[i][j];
531 copy_rvec(r[at],r_ins[gctr]);
532 if( (r_ins[gctr][ZZ]<mem_p->zmax) && (r_ins[gctr][ZZ]>mem_p->zmin) )
534 rvec_inc(pos_ins->geom_cent[i],r_ins[gctr]);
542 svmul(1/(double)c,pos_ins->geom_cent[i],pos_ins->geom_cent[i]);
543 if (!bALLOW_ASYMMETRY)
544 pos_ins->geom_cent[i][ZZ]=mem_p->zmed;
546 fprintf(stderr,"Embedding piece %d with center of geometry: %f %f %f\n",i,pos_ins->geom_cent[i][XX],pos_ins->geom_cent[i][YY],pos_ins->geom_cent[i][ZZ]);
548 fprintf(stderr,"\n");
551 void resize(t_block *ins_at, rvec *r_ins, rvec *r, pos_ins_t *pos_ins,rvec fac)
554 for (k=0;k<pos_ins->pieces;k++)
555 for(i=0;i<pos_ins->nidx[k];i++)
557 at=pos_ins->subindex[k][i];
559 r[at][j]=pos_ins->geom_cent[k][j]+fac[j]*(r_ins[c][j]-pos_ins->geom_cent[k][j]);
564 int gen_rm_list(rm_t *rm_p,t_block *ins_at,t_block *rest_at,t_pbc *pbc, gmx_mtop_t *mtop,
565 rvec *r, rvec *r_ins, mem_t *mem_p, pos_ins_t *pos_ins, real probe_rad, int low_up_rm, gmx_bool bALLOW_ASYMMETRY)
567 int i,j,k,l,at,at2,mol_id;
569 int nrm,nupper,nlower;
570 real r_min_rad,z_lip,min_norm;
576 r_min_rad=probe_rad*probe_rad;
577 snew(rm_p->mol,mtop->mols.nr);
578 snew(rm_p->block,mtop->mols.nr);
581 for(i=0;i<ins_at->nr;i++)
584 for(j=0;j<rest_at->nr;j++)
586 at2=rest_at->index[j];
587 pbc_dx(pbc,r[at],r[at2],dr);
589 if(norm2(dr)<r_min_rad)
591 mol_id = get_mol_id(at2,mtop->nmolblock,mtop->molblock,&type,&block);
594 if(rm_p->mol[l]==mol_id)
598 /*fprintf(stderr,"%d wordt toegevoegd\n",mol_id);*/
599 rm_p->mol[nrm]=mol_id;
600 rm_p->block[nrm]=block;
603 for(l=0;l<mem_p->nmol;l++)
605 if(mol_id==mem_p->mol_id[l])
607 for(k=mtop->mols.index[mol_id];k<mtop->mols.index[mol_id+1];k++)
609 z_lip/=mtop->molblock[block].natoms_mol;
610 if(z_lip<mem_p->zmed)
621 /*make sure equal number of lipids from upper and lower layer are removed */
622 if( (nupper!=nlower) && (!bALLOW_ASYMMETRY) )
624 snew(dist,mem_p->nmol);
625 snew(order,mem_p->nmol);
626 for(i=0;i<mem_p->nmol;i++)
628 at = mtop->mols.index[mem_p->mol_id[i]];
629 pbc_dx(pbc,r[at],pos_ins->geom_cent[0],dr);
630 if (pos_ins->pieces>1)
634 for (k=1;k<pos_ins->pieces;k++)
636 pbc_dx(pbc,r[at],pos_ins->geom_cent[k],dr_tmp);
637 if (norm2(dr_tmp) < min_norm)
639 min_norm=norm2(dr_tmp);
640 copy_rvec(dr_tmp,dr);
644 dist[i]=dr[XX]*dr[XX]+dr[YY]*dr[YY];
646 while (j>=0 && dist[i]<dist[order[j]])
655 while(nupper!=nlower)
657 mol_id=mem_p->mol_id[order[i]];
658 block=get_block(mol_id,mtop->nmolblock,mtop->molblock);
662 if(rm_p->mol[l]==mol_id)
667 for(k=mtop->mols.index[mol_id];k<mtop->mols.index[mol_id+1];k++)
669 z_lip/=mtop->molblock[block].natoms_mol;
670 if(nupper>nlower && z_lip<mem_p->zmed)
672 rm_p->mol[nrm]=mol_id;
673 rm_p->block[nrm]=block;
677 else if (nupper<nlower && z_lip>mem_p->zmed)
679 rm_p->mol[nrm]=mol_id;
680 rm_p->block[nrm]=block;
688 gmx_fatal(FARGS,"Trying to remove more lipid molecules than there are in the membrane");
695 srenew(rm_p->mol,nrm);
696 srenew(rm_p->block,nrm);
698 return nupper+nlower;
701 void rm_group(t_inputrec *ir, gmx_groups_t *groups, gmx_mtop_t *mtop, rm_t *rm_p, t_state *state, t_block *ins_at, pos_ins_t *pos_ins)
703 int i,j,k,n,rm,mol_id,at,block;
705 atom_id *list,*new_mols;
706 unsigned char *new_egrp[egcNR];
709 snew(list,state->natoms);
711 for(i=0;i<rm_p->nr;i++)
714 at=mtop->mols.index[mol_id];
715 block =rm_p->block[i];
716 mtop->molblock[block].nmol--;
717 for(j=0;j<mtop->molblock[block].natoms_mol;j++)
723 mtop->mols.index[mol_id]=-1;
726 mtop->mols.nr-=rm_p->nr;
727 mtop->mols.nalloc_index-=rm_p->nr;
728 snew(new_mols,mtop->mols.nr);
729 for(i=0;i<mtop->mols.nr+rm_p->nr;i++)
732 if(mtop->mols.index[i]!=-1)
734 new_mols[j]=mtop->mols.index[i];
738 sfree(mtop->mols.index);
739 mtop->mols.index=new_mols;
744 state->nalloc=state->natoms;
745 snew(x_tmp,state->nalloc);
746 snew(v_tmp,state->nalloc);
750 if(groups->grpnr[i]!=NULL)
752 groups->ngrpnr[i]=state->natoms;
753 snew(new_egrp[i],state->natoms);
758 for (i=0;i<state->natoms+n;i++)
774 if(groups->grpnr[j]!=NULL)
776 new_egrp[j][i-rm]=groups->grpnr[j][i];
779 copy_rvec(state->x[i],x_tmp[i-rm]);
780 copy_rvec(state->v[i],v_tmp[i-rm]);
781 for(j=0;j<ins_at->nr;j++)
783 if (i==ins_at->index[j])
784 ins_at->index[j]=i-rm;
786 for(j=0;j<pos_ins->pieces;j++)
788 for(k=0;k<pos_ins->nidx[j];k++)
790 if (i==pos_ins->subindex[j][k])
791 pos_ins->subindex[j][k]=i-rm;
803 if(groups->grpnr[i]!=NULL)
805 sfree(groups->grpnr[i]);
806 groups->grpnr[i]=new_egrp[i];
811 int rm_bonded(t_block *ins_at, gmx_mtop_t *mtop)
814 int type,natom,nmol,at,atom1=0,rm_at=0;
816 /*this routine lives dangerously by assuming that all molecules of a given type are in order in the structure*/
817 /*this routine does not live as dangerously as it seems. There is namely a check in mdrunner_membed to make
818 *sure that g_membed exits with a warning when there are molecules of the same type not in the
819 *ins_at index group. MGWolf 050710 */
822 snew(bRM,mtop->nmoltype);
823 for (i=0;i<mtop->nmoltype;i++)
828 for (i=0;i<mtop->nmolblock;i++)
830 /*loop over molecule blocks*/
831 type =mtop->molblock[i].type;
832 natom =mtop->molblock[i].natoms_mol;
833 nmol =mtop->molblock[i].nmol;
835 for(j=0;j<natom*nmol && bRM[type]==TRUE;j++)
837 /*loop over atoms in the block*/
838 at=j+atom1; /*atom index = block index + offset*/
841 for (m=0;(m<ins_at->nr) && (bINS==FALSE);m++)
843 /*loop over atoms in insertion index group to determine if we're inserting one*/
844 if(at==ins_at->index[m])
851 atom1+=natom*nmol; /*update offset*/
854 rm_at+=natom*nmol; /*increment bonded removal counter by # atoms in block*/
858 for(i=0;i<mtop->nmoltype;i++)
864 mtop->moltype[i].ilist[j].nr=0;
866 for(j=F_POSRES;j<=F_VSITEN;j++)
868 mtop->moltype[i].ilist[j].nr=0;
877 void top_update(const char *topfile, char *ins, rm_t *rm_p, gmx_mtop_t *mtop)
879 #define TEMP_FILENM "temp.top"
882 char buf[STRLEN],buf2[STRLEN],*temp;
883 int i,*nmol_rm,nmol,line;
885 fpin = ffopen(topfile,"r");
886 fpout = ffopen(TEMP_FILENM,"w");
888 snew(nmol_rm,mtop->nmoltype);
889 for(i=0;i<rm_p->nr;i++)
890 nmol_rm[rm_p->block[i]]++;
893 while(fgets(buf,STRLEN,fpin))
899 if ((temp=strchr(buf2,'\n')) != NULL)
906 if ((temp=strchr(buf2,'\n')) != NULL)
909 if (buf2[strlen(buf2)-1]==']')
911 buf2[strlen(buf2)-1]='\0';
914 if (gmx_strcasecmp(buf2,"molecules")==0)
917 fprintf(fpout,"%s",buf);
918 } else if (bMolecules==1)
920 for(i=0;i<mtop->nmolblock;i++)
922 nmol=mtop->molblock[i].nmol;
923 sprintf(buf,"%-15s %5d\n",*(mtop->moltype[mtop->molblock[i].type].name),nmol);
924 fprintf(fpout,"%s",buf);
927 } else if (bMolecules==2)
932 fprintf(fpout,"%s",buf);
936 fprintf(fpout,"%s",buf);
941 /* use ffopen to generate backup of topinout */
942 fpout=ffopen(topfile,"w");
944 rename(TEMP_FILENM,topfile);
948 void md_print_warning(const t_commrec *cr,FILE *fplog,const char *buf)
952 fprintf(stderr,"\n%s\n",buf);
956 fprintf(fplog,"\n%s\n",buf);
960 /* simulation conditions to transmit. Keep in mind that they are
961 transmitted to other nodes through an MPI_Reduce after
962 casting them to a real (so the signals can be sent together with other
963 data). This means that the only meaningful values are positive,
965 enum { eglsNABNSB, eglsCHKPT, eglsSTOPCOND, eglsRESETCOUNTERS, eglsNR };
966 /* Is the signal in one simulation independent of other simulations? */
967 gmx_bool gs_simlocal[eglsNR] = { TRUE, FALSE, FALSE, TRUE };
970 int nstms; /* The frequency for intersimulation communication */
971 int sig[eglsNR]; /* The signal set by one process in do_md */
972 int set[eglsNR]; /* The communicated signal, equal for all processes */
976 static int multisim_min(const gmx_multisim_t *ms,int nmin,int n)
979 gmx_bool bPos,bEqual;
984 gmx_sumi_sim(ms->nsim,buf,ms);
987 for(s=0; s<ms->nsim; s++)
989 bPos = bPos && (buf[s] > 0);
990 bEqual = bEqual && (buf[s] == buf[0]);
996 nmin = min(nmin,buf[0]);
1000 /* Find the least common multiple */
1001 for(d=2; d<nmin; d++)
1004 while (s < ms->nsim && d % buf[s] == 0)
1010 /* We found the LCM and it is less than nmin */
1022 static int multisim_nstsimsync(const t_commrec *cr,
1023 const t_inputrec *ir,int repl_ex_nst)
1030 nmin = multisim_min(cr->ms,nmin,ir->nstlist);
1031 nmin = multisim_min(cr->ms,nmin,ir->nstcalcenergy);
1032 nmin = multisim_min(cr->ms,nmin,repl_ex_nst);
1033 if (nmin == INT_MAX)
1035 gmx_fatal(FARGS,"Can not find an appropriate interval for inter-simulation communication, since nstlist, nstcalcenergy and -replex are all <= 0");
1037 /* Avoid inter-simulation communication at every (second) step */
1044 gmx_bcast(sizeof(int),&nmin,cr);
1049 static void init_global_signals(globsig_t *gs,const t_commrec *cr,
1050 const t_inputrec *ir,int repl_ex_nst)
1056 gs->nstms = multisim_nstsimsync(cr,ir,repl_ex_nst);
1059 fprintf(debug,"Syncing simulations for checkpointing and termination every %d steps\n",gs->nstms);
1067 for(i=0; i<eglsNR; i++)
1074 static void copy_coupling_state(t_state *statea,t_state *stateb,
1075 gmx_ekindata_t *ekinda,gmx_ekindata_t *ekindb, t_grpopts* opts)
1078 /* MRS note -- might be able to get rid of some of the arguments. Look over it when it's all debugged */
1082 /* Make sure we have enough space for x and v */
1083 if (statea->nalloc > stateb->nalloc)
1085 stateb->nalloc = statea->nalloc;
1086 srenew(stateb->x,stateb->nalloc);
1087 srenew(stateb->v,stateb->nalloc);
1090 stateb->natoms = statea->natoms;
1091 stateb->ngtc = statea->ngtc;
1092 stateb->nnhpres = statea->nnhpres;
1093 stateb->veta = statea->veta;
1096 copy_mat(ekinda->ekin,ekindb->ekin);
1097 for (i=0; i<stateb->ngtc; i++)
1099 ekindb->tcstat[i].T = ekinda->tcstat[i].T;
1100 ekindb->tcstat[i].Th = ekinda->tcstat[i].Th;
1101 copy_mat(ekinda->tcstat[i].ekinh,ekindb->tcstat[i].ekinh);
1102 copy_mat(ekinda->tcstat[i].ekinf,ekindb->tcstat[i].ekinf);
1103 ekindb->tcstat[i].ekinscalef_nhc = ekinda->tcstat[i].ekinscalef_nhc;
1104 ekindb->tcstat[i].ekinscaleh_nhc = ekinda->tcstat[i].ekinscaleh_nhc;
1105 ekindb->tcstat[i].vscale_nhc = ekinda->tcstat[i].vscale_nhc;
1108 copy_rvecn(statea->x,stateb->x,0,stateb->natoms);
1109 copy_rvecn(statea->v,stateb->v,0,stateb->natoms);
1110 copy_mat(statea->box,stateb->box);
1111 copy_mat(statea->box_rel,stateb->box_rel);
1112 copy_mat(statea->boxv,stateb->boxv);
1114 for (i = 0; i<stateb->ngtc; i++)
1116 nc = i*opts->nhchainlength;
1117 for (j=0; j<opts->nhchainlength; j++)
1119 stateb->nosehoover_xi[nc+j] = statea->nosehoover_xi[nc+j];
1120 stateb->nosehoover_vxi[nc+j] = statea->nosehoover_vxi[nc+j];
1123 if (stateb->nhpres_xi != NULL)
1125 for (i = 0; i<stateb->nnhpres; i++)
1127 nc = i*opts->nhchainlength;
1128 for (j=0; j<opts->nhchainlength; j++)
1130 stateb->nhpres_xi[nc+j] = statea->nhpres_xi[nc+j];
1131 stateb->nhpres_vxi[nc+j] = statea->nhpres_vxi[nc+j];
1137 static void compute_globals(FILE *fplog, gmx_global_stat_t gstat, t_commrec *cr, t_inputrec *ir,
1138 t_forcerec *fr, gmx_ekindata_t *ekind,
1139 t_state *state, t_state *state_global, t_mdatoms *mdatoms,
1140 t_nrnb *nrnb, t_vcm *vcm, gmx_wallcycle_t wcycle,
1141 gmx_enerdata_t *enerd,tensor force_vir, tensor shake_vir, tensor total_vir,
1142 tensor pres, rvec mu_tot, gmx_constr_t constr,
1143 globsig_t *gs,gmx_bool bInterSimGS,
1144 matrix box, gmx_mtop_t *top_global, real *pcurr,
1145 int natoms, gmx_bool *bSumEkinhOld, int flags)
1148 real gs_buf[eglsNR];
1149 tensor corr_vir,corr_pres;
1150 gmx_bool bEner,bPres,bTemp;
1151 gmx_bool bRerunMD, bStopCM, bGStat, bIterate,
1152 bFirstIterate,bReadEkin,bEkinAveVel,bScaleEkin, bConstrain;
1153 real prescorr,enercorr,dvdlcorr;
1155 /* translate CGLO flags to gmx_booleans */
1156 bRerunMD = flags & CGLO_RERUNMD;
1157 bStopCM = flags & CGLO_STOPCM;
1158 bGStat = flags & CGLO_GSTAT;
1159 bReadEkin = (flags & CGLO_READEKIN);
1160 bScaleEkin = (flags & CGLO_SCALEEKIN);
1161 bEner = flags & CGLO_ENERGY;
1162 bTemp = flags & CGLO_TEMPERATURE;
1163 bPres = (flags & CGLO_PRESSURE);
1164 bConstrain = (flags & CGLO_CONSTRAINT);
1165 bIterate = (flags & CGLO_ITERATE);
1166 bFirstIterate = (flags & CGLO_FIRSTITERATE);
1168 /* we calculate a full state kinetic energy either with full-step velocity verlet
1169 or half step where we need the pressure */
1170 bEkinAveVel = (ir->eI==eiVV || (ir->eI==eiVVAK && IR_NPT_TROTTER(ir) && bPres) || bReadEkin);
1172 /* in initalization, it sums the shake virial in vv, and to
1173 sums ekinh_old in leapfrog (or if we are calculating ekinh_old for other reasons */
1175 /* ########## Kinetic energy ############## */
1179 /* Non-equilibrium MD: this is parallellized, but only does communication
1180 * when there really is NEMD.
1183 if (PAR(cr) && (ekind->bNEMD))
1185 accumulate_u(cr,&(ir->opts),ekind);
1190 restore_ekinstate_from_state(cr,ekind,&state_global->ekinstate);
1195 calc_ke_part(state,&(ir->opts),mdatoms,ekind,nrnb,bEkinAveVel,bIterate);
1200 /* Calculate center of mass velocity if necessary, also parallellized */
1201 if (bStopCM && !bRerunMD && bEner)
1203 calc_vcm_grp(fplog,mdatoms->start,mdatoms->homenr,mdatoms,
1204 state->x,state->v,vcm);
1208 if (bTemp || bPres || bEner || bConstrain)
1212 /* We will not sum ekinh_old,
1213 * so signal that we still have to do it.
1215 *bSumEkinhOld = TRUE;
1222 for(i=0; i<eglsNR; i++)
1224 gs_buf[i] = gs->sig[i];
1229 wallcycle_start(wcycle,ewcMoveE);
1230 GMX_MPE_LOG(ev_global_stat_start);
1231 global_stat(fplog,gstat,cr,enerd,force_vir,shake_vir,mu_tot,
1232 ir,ekind,constr,vcm,
1233 gs != NULL ? eglsNR : 0,gs_buf,
1235 *bSumEkinhOld,flags);
1236 GMX_MPE_LOG(ev_global_stat_finish);
1237 wallcycle_stop(wcycle,ewcMoveE);
1241 if (MULTISIM(cr) && bInterSimGS)
1245 /* Communicate the signals between the simulations */
1246 gmx_sum_sim(eglsNR,gs_buf,cr->ms);
1248 /* Communicate the signals form the master to the others */
1249 gmx_bcast(eglsNR*sizeof(gs_buf[0]),gs_buf,cr);
1251 for(i=0; i<eglsNR; i++)
1253 if (bInterSimGS || gs_simlocal[i])
1255 /* Set the communicated signal only when it is non-zero,
1256 * since signals might not be processed at each MD step.
1258 gsi = (gs_buf[i] >= 0 ?
1259 (int)(gs_buf[i] + 0.5) :
1260 (int)(gs_buf[i] - 0.5));
1265 /* Turn off the local signal */
1270 *bSumEkinhOld = FALSE;
1274 if (!ekind->bNEMD && debug && bTemp && (vcm->nr > 0))
1277 mdatoms->start,mdatoms->start+mdatoms->homenr,
1278 state->v,vcm->group_p[0],
1279 mdatoms->massT,mdatoms->tmass,ekind->ekin);
1283 /* Do center of mass motion removal */
1284 if (bStopCM && !bRerunMD) /* is this correct? Does it get called too often with this logic? */
1286 check_cm_grp(fplog,vcm,ir,1);
1287 do_stopcm_grp(fplog,mdatoms->start,mdatoms->homenr,mdatoms->cVCM,
1288 state->x,state->v,vcm);
1289 inc_nrnb(nrnb,eNR_STOPCM,mdatoms->homenr);
1295 /* Sum the kinetic energies of the groups & calc temp */
1296 /* compute full step kinetic energies if vv, or if vv-avek and we are computing the pressure with IR_NPT_TROTTER */
1297 /* three maincase: VV with AveVel (md-vv), vv with AveEkin (md-vv-avek), leap with AveEkin (md).
1298 Leap with AveVel is also an option for the future but not supported now.
1299 bEkinAveVel: If TRUE, we simply multiply ekin by ekinscale to get a full step kinetic energy.
1300 If FALSE, we average ekinh_old and ekinh*ekinscale_nhc to get an averaged half step kinetic energy.
1301 bSaveEkinOld: If TRUE (in the case of iteration = bIterate is TRUE), we don't reset the ekinscale_nhc.
1302 If FALSE, we go ahead and erase over it.
1304 enerd->term[F_TEMP] = sum_ekin(&(ir->opts),ekind,&(enerd->term[F_DKDL]),
1305 bEkinAveVel,bIterate,bScaleEkin);
1307 enerd->term[F_EKIN] = trace(ekind->ekin);
1310 /* ########## Long range energy information ###### */
1312 if (bEner || bPres || bConstrain)
1314 calc_dispcorr(fplog,ir,fr,0,top_global->natoms,box,state->lambda,
1315 corr_pres,corr_vir,&prescorr,&enercorr,&dvdlcorr);
1318 if (bEner && bFirstIterate)
1320 enerd->term[F_DISPCORR] = enercorr;
1321 enerd->term[F_EPOT] += enercorr;
1322 enerd->term[F_DVDL] += dvdlcorr;
1323 if (fr->efep != efepNO) {
1324 enerd->dvdl_lin += dvdlcorr;
1328 /* ########## Now pressure ############## */
1329 if (bPres || bConstrain)
1332 m_add(force_vir,shake_vir,total_vir);
1334 /* Calculate pressure and apply LR correction if PPPM is used.
1335 * Use the box from last timestep since we already called update().
1338 enerd->term[F_PRES] = calc_pres(fr->ePBC,ir->nwall,box,ekind->ekin,total_vir,pres,
1339 (fr->eeltype==eelPPPM)?enerd->term[F_COUL_RECIP]:0.0);
1341 /* Calculate long range corrections to pressure and energy */
1342 /* this adds to enerd->term[F_PRES] and enerd->term[F_ETOT],
1343 and computes enerd->term[F_DISPCORR]. Also modifies the
1344 total_vir and pres tesors */
1346 m_add(total_vir,corr_vir,total_vir);
1347 m_add(pres,corr_pres,pres);
1348 enerd->term[F_PDISPCORR] = prescorr;
1349 enerd->term[F_PRES] += prescorr;
1350 *pcurr = enerd->term[F_PRES];
1351 /* calculate temperature using virial */
1352 enerd->term[F_VTEMP] = calc_temp(trace(total_vir),ir->opts.nrdf[0]);
1358 /* Definitions for convergence of iterated constraints */
1360 /* iterate constraints up to 50 times */
1361 #define MAXITERCONST 50
1366 real f,fprev,x,xprev;
1369 real allrelerr[MAXITERCONST+2];
1370 int num_close; /* number of "close" violations, caused by limited precision. */
1374 #define CONVERGEITER 0.000000001
1375 #define CLOSE_ENOUGH 0.000001000
1377 #define CONVERGEITER 0.0001
1378 #define CLOSE_ENOUGH 0.0050
1381 /* we want to keep track of the close calls. If there are too many, there might be some other issues.
1382 so we make sure that it's either less than some predetermined number, or if more than that number,
1383 only some small fraction of the total. */
1384 #define MAX_NUMBER_CLOSE 50
1385 #define FRACTION_CLOSE 0.001
1387 /* maximum length of cyclic traps to check, emerging from limited numerical precision */
1390 static void gmx_iterate_init(gmx_iterate_t *iterate,gmx_bool bIterate)
1394 iterate->iter_i = 0;
1395 iterate->bIterate = bIterate;
1396 iterate->num_close = 0;
1397 for (i=0;i<MAXITERCONST+2;i++)
1399 iterate->allrelerr[i] = 0;
1403 static gmx_bool done_iterating(const t_commrec *cr,FILE *fplog, int nsteps, gmx_iterate_t *iterate, gmx_bool bFirstIterate, real fom, real *newf)
1405 /* monitor convergence, and use a secant search to propose new
1408 The secant method computes x_{i+1} = x_{i} - f(x_{i}) * ---------------------
1409 f(x_{i}) - f(x_{i-1})
1411 The function we are trying to zero is fom-x, where fom is the
1412 "figure of merit" which is the pressure (or the veta value) we
1413 would get by putting in an old value of the pressure or veta into
1414 the incrementor function for the step or half step. I have
1415 verified that this gives the same answer as self consistent
1416 iteration, usually in many fewer steps, especially for small tau_p.
1418 We could possibly eliminate an iteration with proper use
1419 of the value from the previous step, but that would take a bit
1420 more bookkeeping, especially for veta, since tests indicate the
1421 function of veta on the last step is not sufficiently close to
1422 guarantee convergence this step. This is
1423 good enough for now. On my tests, I could use tau_p down to
1424 0.02, which is smaller that would ever be necessary in
1425 practice. Generally, 3-5 iterations will be sufficient */
1435 iterate->f = fom-iterate->x;
1442 iterate->f = fom-iterate->x; /* we want to zero this difference */
1443 if ((iterate->iter_i > 1) && (iterate->iter_i < MAXITERCONST))
1445 if (iterate->f==iterate->fprev)
1451 *newf = iterate->x - (iterate->x-iterate->xprev)*(iterate->f)/(iterate->f-iterate->fprev);
1456 /* just use self-consistent iteration the first step to initialize, or
1457 if it's not converging (which happens occasionally -- need to investigate why) */
1461 /* Consider a slight shortcut allowing us to exit one sooner -- we check the
1462 difference between the closest of x and xprev to the new
1463 value. To be 100% certain, we should check the difference between
1464 the last result, and the previous result, or
1466 relerr = (fabs((x-xprev)/fom));
1468 but this is pretty much never necessary under typical conditions.
1469 Checking numerically, it seems to lead to almost exactly the same
1470 trajectories, but there are small differences out a few decimal
1471 places in the pressure, and eventually in the v_eta, but it could
1474 if (fabs(*newf-x) < fabs(*newf - xprev)) { xmin = x;} else { xmin = xprev;}
1475 relerr = (fabs((*newf-xmin) / *newf));
1478 err = fabs((iterate->f-iterate->fprev));
1479 relerr = fabs(err/fom);
1481 iterate->allrelerr[iterate->iter_i] = relerr;
1483 if (iterate->iter_i > 0)
1487 fprintf(debug,"Iterating NPT constraints: %6i %20.12f%14.6g%20.12f\n",
1488 iterate->iter_i,fom,relerr,*newf);
1491 if ((relerr < CONVERGEITER) || (err < CONVERGEITER) || (fom==0) || ((iterate->x == iterate->xprev) && iterate->iter_i > 1))
1493 iterate->bIterate = FALSE;
1496 fprintf(debug,"Iterating NPT constraints: CONVERGED\n");
1500 if (iterate->iter_i > MAXITERCONST)
1502 if (relerr < CLOSE_ENOUGH)
1505 for (i=1;i<CYCLEMAX;i++) {
1506 if ((iterate->allrelerr[iterate->iter_i-(1+i)] == iterate->allrelerr[iterate->iter_i-1]) &&
1507 (iterate->allrelerr[iterate->iter_i-(1+i)] == iterate->allrelerr[iterate->iter_i-(1+2*i)])) {
1511 fprintf(debug,"Exiting from an NPT iterating cycle of length %d\n",i);
1518 /* step 1: trapped in a numerical attractor */
1519 /* we are trapped in a numerical attractor, and can't converge any more, and are close to the final result.
1520 Better to give up convergence here than have the simulation die.
1522 iterate->num_close++;
1527 /* Step #2: test if we are reasonably close for other reasons, then monitor the number. If not, die */
1529 /* how many close calls have we had? If less than a few, we're OK */
1530 if (iterate->num_close < MAX_NUMBER_CLOSE)
1532 sprintf(buf,"Slight numerical convergence deviation with NPT at step %d, relative error only %10.5g, likely not a problem, continuing\n",nsteps,relerr);
1533 md_print_warning(cr,fplog,buf);
1534 iterate->num_close++;
1536 /* if more than a few, check the total fraction. If too high, die. */
1537 } else if (iterate->num_close/(double)nsteps > FRACTION_CLOSE) {
1538 gmx_fatal(FARGS,"Could not converge NPT constraints, too many exceptions (%d%%\n",iterate->num_close/(double)nsteps);
1544 gmx_fatal(FARGS,"Could not converge NPT constraints\n");
1549 iterate->xprev = iterate->x;
1551 iterate->fprev = iterate->f;
1557 static void check_nst_param(FILE *fplog,t_commrec *cr,
1558 const char *desc_nst,int nst,
1559 const char *desc_p,int *p)
1563 if (*p > 0 && *p % nst != 0)
1565 /* Round up to the next multiple of nst */
1566 *p = ((*p)/nst + 1)*nst;
1567 sprintf(buf,"NOTE: %s changes %s to %d\n",desc_nst,desc_p,*p);
1568 md_print_warning(cr,fplog,buf);
1572 static void reset_all_counters(FILE *fplog,t_commrec *cr,
1573 gmx_large_int_t step,
1574 gmx_large_int_t *step_rel,t_inputrec *ir,
1575 gmx_wallcycle_t wcycle,t_nrnb *nrnb,
1576 gmx_runtime_t *runtime)
1578 char buf[STRLEN],sbuf[STEPSTRSIZE];
1580 /* Reset all the counters related to performance over the run */
1581 sprintf(buf,"Step %s: resetting all time and cycle counters\n",
1582 gmx_step_str(step,sbuf));
1583 md_print_warning(cr,fplog,buf);
1585 wallcycle_stop(wcycle,ewcRUN);
1586 wallcycle_reset_all(wcycle);
1587 if (DOMAINDECOMP(cr))
1589 reset_dd_statistics_counters(cr->dd);
1592 ir->init_step += *step_rel;
1593 ir->nsteps -= *step_rel;
1595 wallcycle_start(wcycle,ewcRUN);
1596 runtime_start(runtime);
1597 print_date_and_time(fplog,cr->nodeid,"Restarted time",runtime);
1600 static int check_nstglobalcomm(FILE *fplog,t_commrec *cr,
1601 int nstglobalcomm,t_inputrec *ir)
1605 if (!EI_DYNAMICS(ir->eI))
1610 if (nstglobalcomm == -1)
1612 if (ir->nstcalcenergy == 0 && ir->nstlist == 0)
1615 if (ir->nstenergy > 0 && ir->nstenergy < nstglobalcomm)
1617 nstglobalcomm = ir->nstenergy;
1622 /* We assume that if nstcalcenergy > nstlist,
1623 * nstcalcenergy is a multiple of nstlist.
1625 if (ir->nstcalcenergy == 0 ||
1626 (ir->nstlist > 0 && ir->nstlist < ir->nstcalcenergy))
1628 nstglobalcomm = ir->nstlist;
1632 nstglobalcomm = ir->nstcalcenergy;
1638 if (ir->nstlist > 0 &&
1639 nstglobalcomm > ir->nstlist && nstglobalcomm % ir->nstlist != 0)
1641 nstglobalcomm = (nstglobalcomm / ir->nstlist)*ir->nstlist;
1642 sprintf(buf,"WARNING: nstglobalcomm is larger than nstlist, but not a multiple, setting it to %d\n",nstglobalcomm);
1643 md_print_warning(cr,fplog,buf);
1645 if (nstglobalcomm > ir->nstcalcenergy)
1647 check_nst_param(fplog,cr,"-gcom",nstglobalcomm,
1648 "nstcalcenergy",&ir->nstcalcenergy);
1651 check_nst_param(fplog,cr,"-gcom",nstglobalcomm,
1652 "nstenergy",&ir->nstenergy);
1654 check_nst_param(fplog,cr,"-gcom",nstglobalcomm,
1655 "nstlog",&ir->nstlog);
1658 if (ir->comm_mode != ecmNO && ir->nstcomm < nstglobalcomm)
1660 sprintf(buf,"WARNING: Changing nstcomm from %d to %d\n",
1661 ir->nstcomm,nstglobalcomm);
1662 md_print_warning(cr,fplog,buf);
1663 ir->nstcomm = nstglobalcomm;
1666 return nstglobalcomm;
1669 void check_ir_old_tpx_versions(t_commrec *cr,FILE *fplog,
1670 t_inputrec *ir,gmx_mtop_t *mtop)
1672 /* Check required for old tpx files */
1673 if (IR_TWINRANGE(*ir) && ir->nstlist > 1 &&
1674 ir->nstcalcenergy % ir->nstlist != 0)
1676 md_print_warning(cr,fplog,"Old tpr file with twin-range settings: modifying energy calculation and/or T/P-coupling frequencies");
1678 if (gmx_mtop_ftype_count(mtop,F_CONSTR) +
1679 gmx_mtop_ftype_count(mtop,F_CONSTRNC) > 0 &&
1680 ir->eConstrAlg == econtSHAKE)
1682 md_print_warning(cr,fplog,"With twin-range cut-off's and SHAKE the virial and pressure are incorrect");
1683 if (ir->epc != epcNO)
1685 gmx_fatal(FARGS,"Can not do pressure coupling with twin-range cut-off's and SHAKE");
1688 check_nst_param(fplog,cr,"nstlist",ir->nstlist,
1689 "nstcalcenergy",&ir->nstcalcenergy);
1690 check_nst_param(fplog,cr,"nstcalcenergy",ir->nstcalcenergy,
1691 "nstenergy",&ir->nstenergy);
1692 check_nst_param(fplog,cr,"nstcalcenergy",ir->nstcalcenergy,
1693 "nstlog",&ir->nstlog);
1694 if (ir->efep != efepNO)
1696 check_nst_param(fplog,cr,"nstcalcenergy",ir->nstcalcenergy,
1697 "nstdhdl",&ir->nstdhdl);
1703 gmx_bool bGStatEveryStep;
1704 gmx_large_int_t step_ns;
1705 gmx_large_int_t step_nscheck;
1706 gmx_large_int_t nns;
1716 static void reset_nlistheuristics(gmx_nlheur_t *nlh,gmx_large_int_t step)
1720 nlh->step_nscheck = step;
1723 static void init_nlistheuristics(gmx_nlheur_t *nlh,
1724 gmx_bool bGStatEveryStep,gmx_large_int_t step)
1726 nlh->bGStatEveryStep = bGStatEveryStep;
1733 reset_nlistheuristics(nlh,step);
1736 static void update_nliststatistics(gmx_nlheur_t *nlh,gmx_large_int_t step)
1738 gmx_large_int_t nl_lt;
1739 char sbuf[STEPSTRSIZE],sbuf2[STEPSTRSIZE];
1741 /* Determine the neighbor list life time */
1742 nl_lt = step - nlh->step_ns;
1745 fprintf(debug,"%d atoms beyond ns buffer, updating neighbor list after %s steps\n",nlh->nabnsb,gmx_step_str(nl_lt,sbuf));
1749 nlh->s2 += nl_lt*nl_lt;
1750 nlh->ab += nlh->nabnsb;
1751 if (nlh->lt_runav == 0)
1753 nlh->lt_runav = nl_lt;
1754 /* Initialize the fluctuation average
1755 * such that at startup we check after 0 steps.
1757 nlh->lt_runav2 = sqr(nl_lt/2.0);
1759 /* Running average with 0.9 gives an exp. history of 9.5 */
1760 nlh->lt_runav2 = 0.9*nlh->lt_runav2 + 0.1*sqr(nlh->lt_runav - nl_lt);
1761 nlh->lt_runav = 0.9*nlh->lt_runav + 0.1*nl_lt;
1762 if (nlh->bGStatEveryStep)
1764 /* Always check the nlist validity */
1765 nlh->step_nscheck = step;
1769 /* We check after: <life time> - 2*sigma
1770 * The factor 2 is quite conservative,
1771 * but we assume that with nstlist=-1 the user
1772 * prefers exact integration over performance.
1774 nlh->step_nscheck = step
1775 + (int)(nlh->lt_runav - 2.0*sqrt(nlh->lt_runav2)) - 1;
1779 fprintf(debug,"nlist life time %s run av. %4.1f sig %3.1f check %s check with -gcom %d\n",
1780 gmx_step_str(nl_lt,sbuf),nlh->lt_runav,sqrt(nlh->lt_runav2),
1781 gmx_step_str(nlh->step_nscheck-step+1,sbuf2),
1782 (int)(nlh->lt_runav - 2.0*sqrt(nlh->lt_runav2)));
1786 static void set_nlistheuristics(gmx_nlheur_t *nlh,gmx_bool bReset,gmx_large_int_t step)
1792 reset_nlistheuristics(nlh,step);
1796 update_nliststatistics(nlh,step);
1799 nlh->step_ns = step;
1800 /* Initialize the cumulative coordinate scaling matrix */
1801 clear_mat(nlh->scale_tot);
1802 for(d=0; d<DIM; d++)
1804 nlh->scale_tot[d][d] = 1.0;
1808 double do_md_membed(FILE *fplog,t_commrec *cr,int nfile,const t_filenm fnm[],
1809 const output_env_t oenv, gmx_bool bVerbose,gmx_bool bCompact,
1811 gmx_vsite_t *vsite,gmx_constr_t constr,
1812 int stepout,t_inputrec *ir,
1813 gmx_mtop_t *top_global,
1815 t_state *state_global,
1817 t_nrnb *nrnb,gmx_wallcycle_t wcycle,
1818 gmx_edsam_t ed,t_forcerec *fr,
1819 int repl_ex_nst,int repl_ex_seed,
1820 real cpt_period,real max_hours,
1821 const char *deviceOptions,
1822 unsigned long Flags,
1823 gmx_runtime_t *runtime,
1824 rvec fac, rvec *r_ins, pos_ins_t *pos_ins, t_block *ins_at,
1825 real xy_step, real z_step, int it_xy, int it_z)
1828 gmx_large_int_t step,step_rel;
1831 gmx_bool bGStatEveryStep,bGStat,bNstEner,bCalcEnerPres;
1832 gmx_bool bNS,bNStList,bSimAnn,bStopCM,bRerunMD,bNotLastFrame=FALSE,
1833 bFirstStep,bStateFromTPX,bInitStep,bLastStep,
1834 bBornRadii,bStartingFromCpt;
1835 gmx_bool bDoDHDL=FALSE;
1836 gmx_bool do_ene,do_log,do_verbose,bRerunWarnNoV=TRUE,
1837 bForceUpdate=FALSE,bCPT;
1839 gmx_bool bMasterState;
1840 int force_flags,cglo_flags;
1841 tensor force_vir,shake_vir,total_vir,tmp_vir,pres;
1843 t_trxstatus *status;
1846 t_state *bufstate=NULL;
1847 matrix *scale_tot,pcoupl_mu,M,ebox;
1849 t_trxframe rerun_fr;
1850 /* gmx_repl_ex_t repl_ex=NULL;*/
1853 gmx_localtop_t *top;
1854 t_mdebin *mdebin=NULL;
1855 t_state *state=NULL;
1856 rvec *f_global=NULL;
1859 gmx_enerdata_t *enerd;
1861 gmx_global_stat_t gstat;
1862 gmx_update_t upd=NULL;
1863 t_graph *graph=NULL;
1867 gmx_groups_t *groups;
1868 gmx_ekindata_t *ekind, *ekind_save;
1869 gmx_shellfc_t shellfc;
1870 int count,nconverged=0;
1873 gmx_bool bIonize=FALSE;
1874 gmx_bool bTCR=FALSE,bConverged=TRUE,bOK,bSumEkinhOld,bExchanged;
1876 gmx_bool bResetCountersHalfMaxH=FALSE;
1877 gmx_bool bVV,bIterations,bIterate,bFirstIterate,bTemp,bPres,bTrotter;
1880 rvec *xcopy=NULL,*vcopy=NULL,*cbuf=NULL;
1881 matrix boxcopy={{0}},lastbox;
1882 real veta_save,pcurr,scalevir,tracevir;
1885 real last_conserved = 0;
1889 char sbuf[STEPSTRSIZE],sbuf2[STEPSTRSIZE];
1890 int handled_stop_condition=gmx_stop_cond_none; /* compare to get_stop_condition*/
1891 gmx_iterate_t iterate;
1893 /* Temporary addition for FAHCORE checkpointing */
1897 /* Check for special mdrun options */
1898 bRerunMD = (Flags & MD_RERUN);
1899 bIonize = (Flags & MD_IONIZE);
1900 bFFscan = (Flags & MD_FFSCAN);
1901 bAppend = (Flags & MD_APPENDFILES);
1902 bGStatEveryStep = FALSE;
1903 if (Flags & MD_RESETCOUNTERSHALFWAY)
1907 /* Signal to reset the counters half the simulation steps. */
1908 wcycle_set_reset_counters(wcycle,ir->nsteps/2);
1910 /* Signal to reset the counters halfway the simulation time. */
1911 bResetCountersHalfMaxH = (max_hours > 0);
1914 /* md-vv uses averaged full step velocities for T-control
1915 md-vv-avek uses averaged half step velocities for T-control (but full step ekin for P control)
1916 md uses averaged half step kinetic energies to determine temperature unless defined otherwise by GMX_EKIN_AVE_VEL; */
1917 bVV = EI_VV(ir->eI);
1918 if (bVV) /* to store the initial velocities while computing virial */
1920 snew(cbuf,top_global->natoms);
1922 /* all the iteratative cases - only if there are constraints */
1923 bIterations = ((IR_NPT_TROTTER(ir)) && (constr) && (!bRerunMD));
1924 bTrotter = (bVV && (IR_NPT_TROTTER(ir) || (IR_NVT_TROTTER(ir))));
1928 /* Since we don't know if the frames read are related in any way,
1929 * rebuild the neighborlist at every step.
1932 ir->nstcalcenergy = 1;
1936 check_ir_old_tpx_versions(cr,fplog,ir,top_global);
1938 nstglobalcomm = check_nstglobalcomm(fplog,cr,nstglobalcomm,ir);
1939 /*bGStatEveryStep = (nstglobalcomm == 1);*/
1940 bGStatEveryStep = FALSE;
1942 if (!bGStatEveryStep && ir->nstlist == -1 && fplog != NULL)
1945 "To reduce the energy communication with nstlist = -1\n"
1946 "the neighbor list validity should not be checked at every step,\n"
1947 "this means that exact integration is not guaranteed.\n"
1948 "The neighbor list validity is checked after:\n"
1949 " <n.list life time> - 2*std.dev.(n.list life time) steps.\n"
1950 "In most cases this will result in exact integration.\n"
1951 "This reduces the energy communication by a factor of 2 to 3.\n"
1952 "If you want less energy communication, set nstlist > 3.\n\n");
1955 if (bRerunMD || bFFscan)
1959 groups = &top_global->groups;
1961 /* Initial values */
1962 init_md(fplog,cr,ir,oenv,&t,&t0,&state_global->lambda,&lam0,
1963 nrnb,top_global,&upd,
1964 nfile,fnm,&outf,&mdebin,
1965 force_vir,shake_vir,mu_tot,&bSimAnn,&vcm,state_global,Flags);
1967 clear_mat(total_vir);
1969 /* Energy terms and groups */
1971 init_enerdata(top_global->groups.grps[egcENER].nr,ir->n_flambda,enerd);
1972 if (DOMAINDECOMP(cr))
1978 snew(f,top_global->natoms);
1981 /* Kinetic energy data */
1983 init_ekindata(fplog,top_global,&(ir->opts),ekind);
1984 /* needed for iteration of constraints */
1986 init_ekindata(fplog,top_global,&(ir->opts),ekind_save);
1987 /* Copy the cos acceleration to the groups struct */
1988 ekind->cosacc.cos_accel = ir->cos_accel;
1990 gstat = global_stat_init(ir);
1993 /* Check for polarizable models and flexible constraints */
1994 shellfc = init_shell_flexcon(fplog,
1995 top_global,n_flexible_constraints(constr),
1996 (ir->bContinuation ||
1997 (DOMAINDECOMP(cr) && !MASTER(cr))) ?
1998 NULL : state_global->x);
2003 tMPI_Thread_mutex_lock(&deform_init_box_mutex);
2005 set_deform_reference_box(upd,
2006 deform_init_init_step_tpx,
2007 deform_init_box_tpx);
2009 tMPI_Thread_mutex_unlock(&deform_init_box_mutex);
2014 double io = compute_io(ir,top_global->natoms,groups,mdebin->ebin->nener,1);
2015 if ((io > 2000) && MASTER(cr))
2017 "\nWARNING: This run will generate roughly %.0f Mb of data\n\n",
2021 if (DOMAINDECOMP(cr)) {
2022 top = dd_init_local_top(top_global);
2025 dd_init_local_state(cr->dd,state_global,state);
2027 if (DDMASTER(cr->dd) && ir->nstfout) {
2028 snew(f_global,state_global->natoms);
2032 /* Initialize the particle decomposition and split the topology */
2033 top = split_system(fplog,top_global,ir,cr);
2035 pd_cg_range(cr,&fr->cg0,&fr->hcg);
2036 pd_at_range(cr,&a0,&a1);
2038 top = gmx_mtop_generate_local_top(top_global,ir);
2041 a1 = top_global->natoms;
2044 state = partdec_init_local_state(cr,state_global);
2047 atoms2md(top_global,ir,0,NULL,a0,a1-a0,mdatoms);
2050 set_vsite_top(vsite,top,mdatoms,cr);
2053 if (ir->ePBC != epbcNONE && !ir->bPeriodicMols) {
2054 graph = mk_graph(fplog,&(top->idef),0,top_global->natoms,FALSE,FALSE);
2058 make_local_shells(cr,mdatoms,shellfc);
2061 if (ir->pull && PAR(cr)) {
2062 dd_make_local_pull_groups(NULL,ir->pull,mdatoms);
2066 if (DOMAINDECOMP(cr))
2068 /* Distribute the charge groups over the nodes from the master node */
2069 dd_partition_system(fplog,ir->init_step,cr,TRUE,1,
2070 state_global,top_global,ir,
2071 state,&f,mdatoms,top,fr,
2072 vsite,shellfc,constr,
2076 update_mdatoms(mdatoms,state->lambda);
2080 /* Update mdebin with energy history if appending to output files */
2081 if ( Flags & MD_APPENDFILES )
2083 restore_energyhistory_from_state(mdebin,&state_global->enerhist);
2085 /* Set the initial energy history in state to zero by updating once */
2086 update_energyhistory(&state_global->enerhist,mdebin);
2089 if ((state->flags & (1<<estLD_RNG)) && (Flags & MD_READ_RNG)) {
2090 /* Set the random state if we read a checkpoint file */
2091 set_stochd_state(upd,state);
2094 /* Initialize constraints */
2096 if (!DOMAINDECOMP(cr))
2097 set_constraints(constr,top,ir,mdatoms,cr);
2100 /* Check whether we have to GCT stuff */
2101 /* bTCR = ftp2bSet(efGCT,nfile,fnm);
2104 fprintf(stderr,"Will do General Coupling Theory!\n");
2106 gnx = top_global->mols.nr;
2108 for(i=0; (i<gnx); i++) {
2113 /* if (repl_ex_nst > 0 && MASTER(cr))
2114 repl_ex = init_replica_exchange(fplog,cr->ms,state_global,ir,
2115 repl_ex_nst,repl_ex_seed);*/
2117 if (!ir->bContinuation && !bRerunMD)
2119 if (mdatoms->cFREEZE && (state->flags & (1<<estV)))
2121 /* Set the velocities of frozen particles to zero */
2122 for(i=mdatoms->start; i<mdatoms->start+mdatoms->homenr; i++)
2124 for(m=0; m<DIM; m++)
2126 if (ir->opts.nFreeze[mdatoms->cFREEZE[i]][m])
2136 /* Constrain the initial coordinates and velocities */
2137 do_constrain_first(fplog,constr,ir,mdatoms,state,f,
2138 graph,cr,nrnb,fr,top,shake_vir);
2142 /* Construct the virtual sites for the initial configuration */
2143 construct_vsites(fplog,vsite,state->x,nrnb,ir->delta_t,NULL,
2144 top->idef.iparams,top->idef.il,
2145 fr->ePBC,fr->bMolPBC,graph,cr,state->box);
2151 /* I'm assuming we need global communication the first time! MRS */
2152 cglo_flags = (CGLO_TEMPERATURE | CGLO_GSTAT
2153 | (bVV ? CGLO_PRESSURE:0)
2154 | (bVV ? CGLO_CONSTRAINT:0)
2155 | (bRerunMD ? CGLO_RERUNMD:0)
2156 | ((Flags & MD_READ_EKIN) ? CGLO_READEKIN:0));
2158 bSumEkinhOld = FALSE;
2159 compute_globals(fplog,gstat,cr,ir,fr,ekind,state,state_global,mdatoms,nrnb,vcm,
2160 wcycle,enerd,force_vir,shake_vir,total_vir,pres,mu_tot,
2161 constr,NULL,FALSE,state->box,
2162 top_global,&pcurr,top_global->natoms,&bSumEkinhOld,cglo_flags);
2163 if (ir->eI == eiVVAK) {
2164 /* a second call to get the half step temperature initialized as well */
2165 /* we do the same call as above, but turn the pressure off -- internally, this
2166 is recognized as a velocity verlet half-step kinetic energy calculation.
2167 This minimized excess variables, but perhaps loses some logic?*/
2169 compute_globals(fplog,gstat,cr,ir,fr,ekind,state,state_global,mdatoms,nrnb,vcm,
2170 wcycle,enerd,force_vir,shake_vir,total_vir,pres,mu_tot,
2171 constr,NULL,FALSE,state->box,
2172 top_global,&pcurr,top_global->natoms,&bSumEkinhOld,
2173 cglo_flags &~ CGLO_PRESSURE);
2176 /* Calculate the initial half step temperature, and save the ekinh_old */
2177 if (!(Flags & MD_STARTFROMCPT))
2179 for(i=0; (i<ir->opts.ngtc); i++)
2181 copy_mat(ekind->tcstat[i].ekinh,ekind->tcstat[i].ekinh_old);
2186 enerd->term[F_TEMP] *= 2; /* result of averages being done over previous and current step,
2187 and there is no previous step */
2189 temp0 = enerd->term[F_TEMP];
2191 /* if using an iterative algorithm, we need to create a working directory for the state. */
2194 bufstate = init_bufstate(state);
2198 snew(xcopy,state->natoms);
2199 snew(vcopy,state->natoms);
2200 copy_rvecn(state->x,xcopy,0,state->natoms);
2201 copy_rvecn(state->v,vcopy,0,state->natoms);
2202 copy_mat(state->box,boxcopy);
2205 /* need to make an initiation call to get the Trotter variables set, as well as other constants for non-trotter
2206 temperature control */
2207 trotter_seq = init_npt_vars(ir,state,&MassQ,bTrotter);
2211 if (constr && !ir->bContinuation && ir->eConstrAlg == econtLINCS)
2214 "RMS relative constraint deviation after constraining: %.2e\n",
2215 constr_rmsd(constr,FALSE));
2217 fprintf(fplog,"Initial temperature: %g K\n",enerd->term[F_TEMP]);
2220 fprintf(stderr,"starting md rerun '%s', reading coordinates from"
2221 " input trajectory '%s'\n\n",
2222 *(top_global->name),opt2fn("-rerun",nfile,fnm));
2225 fprintf(stderr,"Calculated time to finish depends on nsteps from "
2226 "run input file,\nwhich may not correspond to the time "
2227 "needed to process input trajectory.\n\n");
2233 fprintf(stderr,"starting mdrun '%s'\n",
2234 *(top_global->name));
2235 if (ir->nsteps >= 0)
2237 sprintf(tbuf,"%8.1f",(ir->init_step+ir->nsteps)*ir->delta_t);
2241 sprintf(tbuf,"%s","infinite");
2243 if (ir->init_step > 0)
2245 fprintf(stderr,"%s steps, %s ps (continuing from step %s, %8.1f ps).\n",
2246 gmx_step_str(ir->init_step+ir->nsteps,sbuf),tbuf,
2247 gmx_step_str(ir->init_step,sbuf2),
2248 ir->init_step*ir->delta_t);
2252 fprintf(stderr,"%s steps, %s ps.\n",
2253 gmx_step_str(ir->nsteps,sbuf),tbuf);
2256 fprintf(fplog,"\n");
2259 /* Set and write start time */
2260 runtime_start(runtime);
2261 print_date_and_time(fplog,cr->nodeid,"Started mdrun",runtime);
2262 wallcycle_start(wcycle,ewcRUN);
2264 fprintf(fplog,"\n");
2266 /* safest point to do file checkpointing is here. More general point would be immediately before integrator call */
2267 /*#ifdef GMX_FAHCORE
2268 chkpt_ret=fcCheckPointParallel( cr->nodeid,
2270 if ( chkpt_ret == 0 )
2271 gmx_fatal( 3,__FILE__,__LINE__, "Checkpoint error on step %d\n", 0 );
2275 /***********************************************************
2277 * Loop over MD steps
2279 ************************************************************/
2281 /* if rerunMD then read coordinates and velocities from input trajectory */
2284 if (getenv("GMX_FORCE_UPDATE"))
2286 bForceUpdate = TRUE;
2289 bNotLastFrame = read_first_frame(oenv,&status,
2290 opt2fn("-rerun",nfile,fnm),
2291 &rerun_fr,TRX_NEED_X | TRX_READ_V);
2292 if (rerun_fr.natoms != top_global->natoms)
2295 "Number of atoms in trajectory (%d) does not match the "
2296 "run input file (%d)\n",
2297 rerun_fr.natoms,top_global->natoms);
2299 if (ir->ePBC != epbcNONE)
2303 gmx_fatal(FARGS,"Rerun trajectory frame step %d time %f does not contain a box, while pbc is used",rerun_fr.step,rerun_fr.time);
2305 if (max_cutoff2(ir->ePBC,rerun_fr.box) < sqr(fr->rlistlong))
2307 gmx_fatal(FARGS,"Rerun trajectory frame step %d time %f has too small box dimensions",rerun_fr.step,rerun_fr.time);
2310 /* Set the shift vectors.
2311 * Necessary here when have a static box different from the tpr box.
2313 calc_shifts(rerun_fr.box,fr->shift_vec);
2317 /* loop over MD steps or if rerunMD to end of input trajectory */
2319 /* Skip the first Nose-Hoover integration when we get the state from tpx */
2320 bStateFromTPX = !opt2bSet("-cpi",nfile,fnm);
2321 bInitStep = bFirstStep && (bStateFromTPX || bVV);
2322 bStartingFromCpt = (Flags & MD_STARTFROMCPT) && bInitStep;
2324 bSumEkinhOld = FALSE;
2327 init_global_signals(&gs,cr,ir,repl_ex_nst);
2329 step = ir->init_step;
2332 if (ir->nstlist == -1)
2334 init_nlistheuristics(&nlh,bGStatEveryStep,step);
2337 bLastStep = (bRerunMD || (ir->nsteps >= 0 && step_rel > ir->nsteps));
2338 while (!bLastStep || (bRerunMD && bNotLastFrame)) {
2340 wallcycle_start(wcycle,ewcSTEP);
2342 GMX_MPE_LOG(ev_timestep1);
2345 if (rerun_fr.bStep) {
2346 step = rerun_fr.step;
2347 step_rel = step - ir->init_step;
2349 if (rerun_fr.bTime) {
2359 bLastStep = (step_rel == ir->nsteps);
2360 t = t0 + step*ir->delta_t;
2363 if (ir->efep != efepNO)
2365 if (bRerunMD && rerun_fr.bLambda && (ir->delta_lambda!=0))
2367 state_global->lambda = rerun_fr.lambda;
2371 state_global->lambda = lam0 + step*ir->delta_lambda;
2373 state->lambda = state_global->lambda;
2374 bDoDHDL = do_per_step(step,ir->nstdhdl);
2379 update_annealing_target_temp(&(ir->opts),t);
2384 if (!(DOMAINDECOMP(cr) && !MASTER(cr)))
2386 for(i=0; i<state_global->natoms; i++)
2388 copy_rvec(rerun_fr.x[i],state_global->x[i]);
2392 for(i=0; i<state_global->natoms; i++)
2394 copy_rvec(rerun_fr.v[i],state_global->v[i]);
2399 for(i=0; i<state_global->natoms; i++)
2401 clear_rvec(state_global->v[i]);
2405 fprintf(stderr,"\nWARNING: Some frames do not contain velocities.\n"
2406 " Ekin, temperature and pressure are incorrect,\n"
2407 " the virial will be incorrect when constraints are present.\n"
2409 bRerunWarnNoV = FALSE;
2413 copy_mat(rerun_fr.box,state_global->box);
2414 copy_mat(state_global->box,state->box);
2416 if (vsite && (Flags & MD_RERUN_VSITE))
2418 if (DOMAINDECOMP(cr))
2420 gmx_fatal(FARGS,"Vsite recalculation with -rerun is not implemented for domain decomposition, use particle decomposition");
2424 /* Following is necessary because the graph may get out of sync
2425 * with the coordinates if we only have every N'th coordinate set
2427 mk_mshift(fplog,graph,fr->ePBC,state->box,state->x);
2428 shift_self(graph,state->box,state->x);
2430 construct_vsites(fplog,vsite,state->x,nrnb,ir->delta_t,state->v,
2431 top->idef.iparams,top->idef.il,
2432 fr->ePBC,fr->bMolPBC,graph,cr,state->box);
2435 unshift_self(graph,state->box,state->x);
2440 /* Stop Center of Mass motion */
2441 bStopCM = (ir->comm_mode != ecmNO && do_per_step(step,ir->nstcomm));
2443 /* Copy back starting coordinates in case we're doing a forcefield scan */
2446 for(ii=0; (ii<state->natoms); ii++)
2448 copy_rvec(xcopy[ii],state->x[ii]);
2449 copy_rvec(vcopy[ii],state->v[ii]);
2451 copy_mat(boxcopy,state->box);
2456 /* for rerun MD always do Neighbour Searching */
2457 bNS = (bFirstStep || ir->nstlist != 0);
2462 /* Determine whether or not to do Neighbour Searching and LR */
2463 bNStList = (ir->nstlist > 0 && step % ir->nstlist == 0);
2465 bNS = (bFirstStep || bExchanged || bNStList ||
2466 (ir->nstlist == -1 && nlh.nabnsb > 0));
2468 if (bNS && ir->nstlist == -1)
2470 set_nlistheuristics(&nlh,bFirstStep || bExchanged,step);
2474 /* < 0 means stop at next step, > 0 means stop at next NS step */
2475 if ( (gs.set[eglsSTOPCOND] < 0 ) ||
2476 ( (gs.set[eglsSTOPCOND] > 0 ) && ( bNS || ir->nstlist==0)) )
2481 /* Determine whether or not to update the Born radii if doing GB */
2482 bBornRadii=bFirstStep;
2483 if (ir->implicit_solvent && (step % ir->nstgbradii==0))
2488 do_log = do_per_step(step,ir->nstlog) || bFirstStep || bLastStep;
2489 do_verbose = bVerbose &&
2490 (step % stepout == 0 || bFirstStep || bLastStep);
2492 if (bNS && !(bFirstStep && ir->bContinuation && !bRerunMD))
2496 bMasterState = TRUE;
2500 bMasterState = FALSE;
2501 /* Correct the new box if it is too skewed */
2502 if (DYNAMIC_BOX(*ir))
2504 if (correct_box(fplog,step,state->box,graph))
2506 bMasterState = TRUE;
2509 if (DOMAINDECOMP(cr) && bMasterState)
2511 dd_collect_state(cr->dd,state,state_global);
2515 if (DOMAINDECOMP(cr))
2517 /* Repartition the domain decomposition */
2518 wallcycle_start(wcycle,ewcDOMDEC);
2519 dd_partition_system(fplog,step,cr,
2520 bMasterState,nstglobalcomm,
2521 state_global,top_global,ir,
2522 state,&f,mdatoms,top,fr,
2523 vsite,shellfc,constr,
2524 nrnb,wcycle,do_verbose);
2525 wallcycle_stop(wcycle,ewcDOMDEC);
2526 /* If using an iterative integrator, reallocate space to match the decomposition */
2530 if (MASTER(cr) && do_log && !bFFscan)
2532 print_ebin_header(fplog,step,t,state->lambda);
2535 if (ir->efep != efepNO)
2537 update_mdatoms(mdatoms,state->lambda);
2540 if (bRerunMD && rerun_fr.bV)
2543 /* We need the kinetic energy at minus the half step for determining
2544 * the full step kinetic energy and possibly for T-coupling.*/
2545 /* This may not be quite working correctly yet . . . . */
2546 compute_globals(fplog,gstat,cr,ir,fr,ekind,state,state_global,mdatoms,nrnb,vcm,
2547 wcycle,enerd,NULL,NULL,NULL,NULL,mu_tot,
2548 constr,NULL,FALSE,state->box,
2549 top_global,&pcurr,top_global->natoms,&bSumEkinhOld,
2550 CGLO_RERUNMD | CGLO_GSTAT | CGLO_TEMPERATURE);
2552 clear_mat(force_vir);
2554 /* Ionize the atoms if necessary */
2557 ionize(fplog,oenv,mdatoms,top_global,t,ir,state->x,state->v,
2558 mdatoms->start,mdatoms->start+mdatoms->homenr,state->box,cr);
2561 /* Update force field in ffscan program */
2564 if (update_forcefield(fplog,
2566 mdatoms->nr,state->x,state->box)) {
2567 if (gmx_parallel_env_initialized())
2575 GMX_MPE_LOG(ev_timestep2);
2577 /* We write a checkpoint at this MD step when:
2578 * either at an NS step when we signalled through gs,
2579 * or at the last step (but not when we do not want confout),
2580 * but never at the first step or with rerun.
2582 /* bCPT = (((gs.set[eglsCHKPT] && bNS) ||
2583 (bLastStep && (Flags & MD_CONFOUT))) &&
2584 step > ir->init_step && !bRerunMD);
2587 gs.set[eglsCHKPT] = 0;
2590 /* Determine the energy and pressure:
2591 * at nstcalcenergy steps and at energy output steps (set below).
2593 bNstEner = (bGStatEveryStep || do_per_step(step,ir->nstcalcenergy));
2594 bCalcEnerPres = bNstEner;
2596 /* Do we need global communication ? */
2597 bGStat = (bCalcEnerPres || bStopCM ||
2598 (ir->nstlist == -1 && !bRerunMD && step >= nlh.step_nscheck));
2600 do_ene = (do_per_step(step,ir->nstenergy) || bLastStep);
2602 if (do_ene || do_log)
2604 bCalcEnerPres = TRUE;
2608 /* these CGLO_ options remain the same throughout the iteration */
2609 cglo_flags = ((bRerunMD ? CGLO_RERUNMD : 0) |
2610 (bStopCM ? CGLO_STOPCM : 0) |
2611 (bGStat ? CGLO_GSTAT : 0)
2614 force_flags = (GMX_FORCE_STATECHANGED |
2615 ((DYNAMIC_BOX(*ir) || bRerunMD) ? GMX_FORCE_DYNAMICBOX : 0) |
2616 GMX_FORCE_ALLFORCES |
2617 (bNStList ? GMX_FORCE_DOLR : 0) |
2619 (bCalcEnerPres ? GMX_FORCE_VIRIAL : 0) |
2620 (bDoDHDL ? GMX_FORCE_DHDL : 0)
2625 /* Now is the time to relax the shells */
2626 count=relax_shell_flexcon(fplog,cr,bVerbose,bFFscan ? step+1 : step,
2628 bStopCM,top,top_global,
2630 state,f,force_vir,mdatoms,
2631 nrnb,wcycle,graph,groups,
2632 shellfc,fr,bBornRadii,t,mu_tot,
2633 state->natoms,&bConverged,vsite,
2644 /* The coordinates (x) are shifted (to get whole molecules)
2646 * This is parallellized as well, and does communication too.
2647 * Check comments in sim_util.c
2650 do_force(fplog,cr,ir,step,nrnb,wcycle,top,top_global,groups,
2651 state->box,state->x,&state->hist,
2652 f,force_vir,mdatoms,enerd,fcd,
2653 state->lambda,graph,
2654 fr,vsite,mu_tot,t,outf->fp_field,ed,bBornRadii,
2655 (bNS ? GMX_FORCE_NS : 0) | force_flags);
2658 GMX_BARRIER(cr->mpi_comm_mygroup);
2662 mu_aver = calc_mu_aver(cr,state->x,mdatoms->chargeA,
2663 mu_tot,&top_global->mols,mdatoms,gnx,grpindex);
2666 if (bTCR && bFirstStep)
2668 tcr=init_coupling(fplog,nfile,fnm,cr,fr,mdatoms,&(top->idef));
2669 fprintf(fplog,"Done init_coupling\n");
2673 /* ############### START FIRST UPDATE HALF-STEP ############### */
2675 if (bVV && !bStartingFromCpt && !bRerunMD)
2681 /* if using velocity verlet with full time step Ekin,
2682 * take the first half step only to compute the
2683 * virial for the first step. From there,
2684 * revert back to the initial coordinates
2685 * so that the input is actually the initial step.
2687 copy_rvecn(state->v,cbuf,0,state->natoms); /* should make this better for parallelizing? */
2690 /* this is for NHC in the Ekin(t+dt/2) version of vv */
2693 trotter_update(ir,step,ekind,enerd,state,total_vir,mdatoms,&MassQ,trotter_seq,ettTSEQ2);
2696 if (ir->eI == eiVVAK)
2698 update_tcouple(fplog,step,ir,state,ekind,wcycle,upd,&MassQ,mdatoms);
2701 update_coords(fplog,step,ir,mdatoms,state,
2702 f,fr->bTwinRange && bNStList,fr->f_twin,fcd,
2703 ekind,M,wcycle,upd,bInitStep,etrtVELOCITY1,
2704 cr,nrnb,constr,&top->idef);
2708 gmx_iterate_init(&iterate,bIterations && !bInitStep);
2710 /* for iterations, we save these vectors, as we will be self-consistently iterating
2712 /*#### UPDATE EXTENDED VARIABLES IN TROTTER FORMULATION */
2714 /* save the state */
2715 if (bIterations && iterate.bIterate) {
2716 copy_coupling_state(state,bufstate,ekind,ekind_save,&(ir->opts));
2720 bFirstIterate = TRUE;
2721 while (bFirstIterate || (bIterations && iterate.bIterate))
2723 if (bIterations && iterate.bIterate)
2725 copy_coupling_state(bufstate,state,ekind_save,ekind,&(ir->opts));
2726 if (bFirstIterate && bTrotter)
2728 /* The first time through, we need a decent first estimate
2729 of veta(t+dt) to compute the constraints. Do
2730 this by computing the box volume part of the
2731 trotter integration at this time. Nothing else
2732 should be changed by this routine here. If
2733 !(first time), we start with the previous value
2736 veta_save = state->veta;
2737 trotter_update(ir,step,ekind,enerd,state,total_vir,mdatoms,&MassQ,trotter_seq,ettTSEQ0);
2738 vetanew = state->veta;
2739 state->veta = veta_save;
2744 if ( !bRerunMD || rerun_fr.bV || bForceUpdate) { /* Why is rerun_fr.bV here? Unclear. */
2747 update_constraints(fplog,step,&dvdl,ir,ekind,mdatoms,state,graph,f,
2748 &top->idef,shake_vir,NULL,
2749 cr,nrnb,wcycle,upd,constr,
2750 bInitStep,TRUE,bCalcEnerPres,vetanew);
2752 if (!bOK && !bFFscan)
2754 gmx_fatal(FARGS,"Constraint error: Shake, Lincs or Settle could not solve the constrains");
2759 { /* Need to unshift here if a do_force has been
2760 called in the previous step */
2761 unshift_self(graph,state->box,state->x);
2766 /* if VV, compute the pressure and constraints */
2767 /* if VV2, the pressure and constraints only if using pressure control.*/
2768 bPres = (ir->eI==eiVV || IR_NPT_TROTTER(ir));
2769 bTemp = ((ir->eI==eiVV &&(!bInitStep)) || (ir->eI==eiVVAK && IR_NPT_TROTTER(ir)));
2770 compute_globals(fplog,gstat,cr,ir,fr,ekind,state,state_global,mdatoms,nrnb,vcm,
2771 wcycle,enerd,force_vir,shake_vir,total_vir,pres,mu_tot,
2772 constr,NULL,FALSE,state->box,
2773 top_global,&pcurr,top_global->natoms,&bSumEkinhOld,
2776 | (bTemp ? CGLO_TEMPERATURE:0)
2777 | (bPres ? CGLO_PRESSURE : 0)
2778 | (bPres ? CGLO_CONSTRAINT : 0)
2779 | (iterate.bIterate ? CGLO_ITERATE : 0)
2780 | (bFirstIterate ? CGLO_FIRSTITERATE : 0)
2784 /* explanation of above:
2785 a) We compute Ekin at the full time step
2786 if 1) we are using the AveVel Ekin, and it's not the
2787 initial step, or 2) if we are using AveEkin, but need the full
2788 time step kinetic energy for the pressure.
2789 b) If we are using EkinAveEkin for the kinetic energy for the temperture control, we still feed in
2790 EkinAveVel because it's needed for the pressure */
2792 /* temperature scaling and pressure scaling to produce the extended variables at t+dt */
2793 if (bVV && !bInitStep)
2795 trotter_update(ir,step,ekind,enerd,state,total_vir,mdatoms,&MassQ, trotter_seq,ettTSEQ2);
2799 done_iterating(cr,fplog,step,&iterate,bFirstIterate,
2800 state->veta,&vetanew))
2804 bFirstIterate = FALSE;
2807 if (bTrotter && !bInitStep) {
2808 copy_mat(shake_vir,state->svir_prev);
2809 copy_mat(force_vir,state->fvir_prev);
2810 if (IR_NVT_TROTTER(ir) && ir->eI==eiVV) {
2811 /* update temperature and kinetic energy now that step is over - this is the v(t+dt) point */
2812 enerd->term[F_TEMP] = sum_ekin(&(ir->opts),ekind,NULL,(ir->eI==eiVV),FALSE,FALSE);
2813 enerd->term[F_EKIN] = trace(ekind->ekin);
2816 /* if it's the initial step, we performed this first step just to get the constraint virial */
2817 if (bInitStep && ir->eI==eiVV) {
2818 copy_rvecn(cbuf,state->v,0,state->natoms);
2821 if (fr->bSepDVDL && fplog && do_log)
2823 fprintf(fplog,sepdvdlformat,"Constraint",0.0,dvdl);
2825 enerd->term[F_DHDL_CON] += dvdl;
2827 GMX_MPE_LOG(ev_timestep1);
2831 /* MRS -- now done iterating -- compute the conserved quantity */
2834 if (IR_NVT_TROTTER(ir) || IR_NPT_TROTTER(ir))
2837 NPT_energy(ir,state,&MassQ);
2838 if ((ir->eDispCorr != edispcEnerPres) && (ir->eDispCorr != edispcAllEnerPres))
2840 last_conserved -= enerd->term[F_DISPCORR];
2844 last_ekin = enerd->term[F_EKIN]; /* does this get preserved through checkpointing? */
2848 /* ######## END FIRST UPDATE STEP ############## */
2849 /* ######## If doing VV, we now have v(dt) ###### */
2851 /* ################## START TRAJECTORY OUTPUT ################# */
2853 /* Now we have the energies and forces corresponding to the
2854 * coordinates at time t. We must output all of this before
2856 * for RerunMD t is read from input trajectory
2858 GMX_MPE_LOG(ev_output_start);
2861 if (do_per_step(step,ir->nstxout)) { mdof_flags |= MDOF_X; }
2862 if (do_per_step(step,ir->nstvout)) { mdof_flags |= MDOF_V; }
2863 if (do_per_step(step,ir->nstfout)) { mdof_flags |= MDOF_F; }
2864 if (do_per_step(step,ir->nstxtcout)) { mdof_flags |= MDOF_XTC; }
2865 /* if (bCPT) { mdof_flags |= MDOF_CPT; };*/
2869 fcReportProgress( ir->nsteps, step );
2873 /* Enforce writing positions and velocities at end of run */
2874 mdof_flags |= (MDOF_X | MDOF_V);
2876 /* sync bCPT and fc record-keeping */
2877 /* if (bCPT && MASTER(cr))
2878 fcRequestCheckPoint();*/
2881 if (mdof_flags != 0)
2883 wallcycle_start(wcycle,ewcTRAJ);
2886 if (state->flags & (1<<estLD_RNG))
2888 get_stochd_state(upd,state);
2894 state_global->ekinstate.bUpToDate = FALSE;
2898 update_ekinstate(&state_global->ekinstate,ekind);
2899 state_global->ekinstate.bUpToDate = TRUE;
2901 update_energyhistory(&state_global->enerhist,mdebin);
2904 write_traj(fplog,cr,outf,mdof_flags,top_global,
2905 step,t,state,state_global,f,f_global,&n_xtc,&x_xtc);
2912 if (bLastStep && step_rel == ir->nsteps &&
2913 (Flags & MD_CONFOUT) && MASTER(cr) &&
2914 !bRerunMD && !bFFscan)
2916 /* x and v have been collected in write_traj,
2917 * because a checkpoint file will always be written
2920 fprintf(stderr,"\nWriting final coordinates.\n");
2921 if (ir->ePBC != epbcNONE && !ir->bPeriodicMols &&
2924 /* Make molecules whole only for confout writing */
2925 do_pbc_mtop(fplog,ir->ePBC,state->box,top_global,state_global->x);
2927 /* write_sto_conf_mtop(ftp2fn(efSTO,nfile,fnm),
2928 *top_global->name,top_global,
2929 state_global->x,state_global->v,
2930 ir->ePBC,state->box);*/
2933 wallcycle_stop(wcycle,ewcTRAJ);
2935 GMX_MPE_LOG(ev_output_finish);
2937 /* kludge -- virial is lost with restart for NPT control. Must restart */
2938 if (bStartingFromCpt && bVV)
2940 copy_mat(state->svir_prev,shake_vir);
2941 copy_mat(state->fvir_prev,force_vir);
2943 /* ################## END TRAJECTORY OUTPUT ################ */
2945 /* Determine the pressure:
2946 * always when we want exact averages in the energy file,
2947 * at ns steps when we have pressure coupling,
2948 * otherwise only at energy output steps (set below).
2951 bNstEner = (bGStatEveryStep || do_per_step(step,ir->nstcalcenergy));
2952 bCalcEnerPres = bNstEner;
2954 /* Do we need global communication ? */
2955 bGStat = (bGStatEveryStep || bStopCM || bNS ||
2956 (ir->nstlist == -1 && !bRerunMD && step >= nlh.step_nscheck));
2958 do_ene = (do_per_step(step,ir->nstenergy) || bLastStep);
2960 if (do_ene || do_log)
2962 bCalcEnerPres = TRUE;
2966 /* Determine the wallclock run time up till now */
2967 run_time = gmx_gettime() - (double)runtime->real;
2969 /* Check whether everything is still allright */
2970 if (((int)gmx_get_stop_condition() > handled_stop_condition)
2976 /* this is just make gs.sig compatible with the hack
2977 of sending signals around by MPI_Reduce with together with
2979 if ( gmx_get_stop_condition() == gmx_stop_cond_next_ns )
2980 gs.sig[eglsSTOPCOND]=1;
2981 if ( gmx_get_stop_condition() == gmx_stop_cond_next )
2982 gs.sig[eglsSTOPCOND]=-1;
2983 /* < 0 means stop at next step, > 0 means stop at next NS step */
2987 "\n\nReceived the %s signal, stopping at the next %sstep\n\n",
2988 gmx_get_signal_name(),
2989 gs.sig[eglsSTOPCOND]==1 ? "NS " : "");
2993 "\n\nReceived the %s signal, stopping at the next %sstep\n\n",
2994 gmx_get_signal_name(),
2995 gs.sig[eglsSTOPCOND]==1 ? "NS " : "");
2997 handled_stop_condition=(int)gmx_get_stop_condition();
2999 else if (MASTER(cr) && (bNS || ir->nstlist <= 0) &&
3000 (max_hours > 0 && run_time > max_hours*60.0*60.0*0.99) &&
3001 gs.sig[eglsSTOPCOND] == 0 && gs.set[eglsSTOPCOND] == 0)
3003 /* Signal to terminate the run */
3004 gs.sig[eglsSTOPCOND] = 1;
3007 fprintf(fplog,"\nStep %s: Run time exceeded %.3f hours, will terminate the run\n",gmx_step_str(step,sbuf),max_hours*0.99);
3009 fprintf(stderr, "\nStep %s: Run time exceeded %.3f hours, will terminate the run\n",gmx_step_str(step,sbuf),max_hours*0.99);
3012 if (bResetCountersHalfMaxH && MASTER(cr) &&
3013 run_time > max_hours*60.0*60.0*0.495)
3015 gs.sig[eglsRESETCOUNTERS] = 1;
3018 if (ir->nstlist == -1 && !bRerunMD)
3020 /* When bGStatEveryStep=FALSE, global_stat is only called
3021 * when we check the atom displacements, not at NS steps.
3022 * This means that also the bonded interaction count check is not
3023 * performed immediately after NS. Therefore a few MD steps could
3024 * be performed with missing interactions.
3025 * But wrong energies are never written to file,
3026 * since energies are only written after global_stat
3029 if (step >= nlh.step_nscheck)
3031 nlh.nabnsb = natoms_beyond_ns_buffer(ir,fr,&top->cgs,
3032 nlh.scale_tot,state->x);
3036 /* This is not necessarily true,
3037 * but step_nscheck is determined quite conservatively.
3043 /* In parallel we only have to check for checkpointing in steps
3044 * where we do global communication,
3045 * otherwise the other nodes don't know.
3047 if (MASTER(cr) && ((bGStat || !PAR(cr)) &&
3050 run_time >= nchkpt*cpt_period*60.0)) &&
3051 gs.set[eglsCHKPT] == 0)
3053 gs.sig[eglsCHKPT] = 1;
3058 gmx_iterate_init(&iterate,bIterations);
3061 /* for iterations, we save these vectors, as we will be redoing the calculations */
3062 if (bIterations && iterate.bIterate)
3064 copy_coupling_state(state,bufstate,ekind,ekind_save,&(ir->opts));
3066 bFirstIterate = TRUE;
3067 while (bFirstIterate || (bIterations && iterate.bIterate))
3069 /* We now restore these vectors to redo the calculation with improved extended variables */
3072 copy_coupling_state(bufstate,state,ekind_save,ekind,&(ir->opts));
3075 /* We make the decision to break or not -after- the calculation of Ekin and Pressure,
3076 so scroll down for that logic */
3078 /* ######### START SECOND UPDATE STEP ################# */
3079 GMX_MPE_LOG(ev_update_start);
3081 if (!bRerunMD || rerun_fr.bV || bForceUpdate)
3083 wallcycle_start(wcycle,ewcUPDATE);
3085 /* Box is changed in update() when we do pressure coupling,
3086 * but we should still use the old box for energy corrections and when
3087 * writing it to the energy file, so it matches the trajectory files for
3088 * the same timestep above. Make a copy in a separate array.
3090 copy_mat(state->box,lastbox);
3091 /* UPDATE PRESSURE VARIABLES IN TROTTER FORMULATION WITH CONSTRAINTS */
3094 if (bIterations && iterate.bIterate)
3102 /* we use a new value of scalevir to converge the iterations faster */
3103 scalevir = tracevir/trace(shake_vir);
3105 msmul(shake_vir,scalevir,shake_vir);
3106 m_add(force_vir,shake_vir,total_vir);
3107 clear_mat(shake_vir);
3109 trotter_update(ir,step,ekind,enerd,state,total_vir,mdatoms,&MassQ, trotter_seq,ettTSEQ3);
3111 /* We can only do Berendsen coupling after we have summed
3112 * the kinetic energy or virial. Since the happens
3113 * in global_state after update, we should only do it at
3114 * step % nstlist = 1 with bGStatEveryStep=FALSE.
3117 if (ir->eI != eiVVAK)
3119 update_tcouple(fplog,step,ir,state,ekind,wcycle,upd,&MassQ,mdatoms);
3121 update_pcouple(fplog,step,ir,state,pcoupl_mu,M,wcycle,
3126 /* velocity half-step update */
3127 update_coords(fplog,step,ir,mdatoms,state,f,fr->bTwinRange && bNStList,fr->f_twin,fcd,
3128 ekind,M,wcycle,upd,FALSE,etrtVELOCITY2,cr,nrnb,constr,&top->idef);
3131 /* Above, initialize just copies ekinh into ekin,
3132 * it doesn't copy position (for VV),
3133 * and entire integrator for MD.
3138 copy_rvecn(state->x,cbuf,0,state->natoms);
3141 update_coords(fplog,step,ir,mdatoms,state,f,fr->bTwinRange && bNStList,fr->f_twin,fcd,
3142 ekind,M,wcycle,upd,bInitStep,etrtPOSITION,cr,nrnb,constr,&top->idef);
3143 wallcycle_stop(wcycle,ewcUPDATE);
3145 update_constraints(fplog,step,&dvdl,ir,ekind,mdatoms,state,graph,f,
3146 &top->idef,shake_vir,force_vir,
3147 cr,nrnb,wcycle,upd,constr,
3148 bInitStep,FALSE,bCalcEnerPres,state->veta);
3152 /* erase F_EKIN and F_TEMP here? */
3153 /* just compute the kinetic energy at the half step to perform a trotter step */
3154 compute_globals(fplog,gstat,cr,ir,fr,ekind,state,state_global,mdatoms,nrnb,vcm,
3155 wcycle,enerd,force_vir,shake_vir,total_vir,pres,mu_tot,
3156 constr,NULL,FALSE,lastbox,
3157 top_global,&pcurr,top_global->natoms,&bSumEkinhOld,
3158 cglo_flags | CGLO_TEMPERATURE | CGLO_CONSTRAINT
3160 wallcycle_start(wcycle,ewcUPDATE);
3161 trotter_update(ir,step,ekind,enerd,state,total_vir,mdatoms,&MassQ, trotter_seq,ettTSEQ4);
3162 /* now we know the scaling, we can compute the positions again again */
3163 copy_rvecn(cbuf,state->x,0,state->natoms);
3165 update_coords(fplog,step,ir,mdatoms,state,f,fr->bTwinRange && bNStList,fr->f_twin,fcd,
3166 ekind,M,wcycle,upd,bInitStep,etrtPOSITION,cr,nrnb,constr,&top->idef);
3167 wallcycle_stop(wcycle,ewcUPDATE);
3169 /* do we need an extra constraint here? just need to copy out of state->v to upd->xp? */
3170 /* are the small terms in the shake_vir here due
3171 * to numerical errors, or are they important
3172 * physically? I'm thinking they are just errors, but not completely sure.
3173 * For now, will call without actually constraining, constr=NULL*/
3174 update_constraints(fplog,step,&dvdl,ir,ekind,mdatoms,state,graph,f,
3175 &top->idef,tmp_vir,force_vir,
3176 cr,nrnb,wcycle,upd,NULL,
3177 bInitStep,FALSE,bCalcEnerPres,state->veta);
3179 if (!bOK && !bFFscan)
3181 gmx_fatal(FARGS,"Constraint error: Shake, Lincs or Settle could not solve the constrains");
3184 if (fr->bSepDVDL && fplog && do_log)
3186 fprintf(fplog,sepdvdlformat,"Constraint",0.0,dvdl);
3188 enerd->term[F_DHDL_CON] += dvdl;
3192 /* Need to unshift here */
3193 unshift_self(graph,state->box,state->x);
3196 GMX_BARRIER(cr->mpi_comm_mygroup);
3197 GMX_MPE_LOG(ev_update_finish);
3201 wallcycle_start(wcycle,ewcVSITECONSTR);
3204 shift_self(graph,state->box,state->x);
3206 construct_vsites(fplog,vsite,state->x,nrnb,ir->delta_t,state->v,
3207 top->idef.iparams,top->idef.il,
3208 fr->ePBC,fr->bMolPBC,graph,cr,state->box);
3212 unshift_self(graph,state->box,state->x);
3214 wallcycle_stop(wcycle,ewcVSITECONSTR);
3217 /* ############## IF NOT VV, Calculate globals HERE, also iterate constraints ############ */
3218 if (ir->nstlist == -1 && bFirstIterate)
3220 gs.sig[eglsNABNSB] = nlh.nabnsb;
3222 compute_globals(fplog,gstat,cr,ir,fr,ekind,state,state_global,mdatoms,nrnb,vcm,
3223 wcycle,enerd,force_vir,shake_vir,total_vir,pres,mu_tot,
3225 bFirstIterate ? &gs : NULL,(step % gs.nstms == 0),
3227 top_global,&pcurr,top_global->natoms,&bSumEkinhOld,
3229 | (!EI_VV(ir->eI) ? CGLO_ENERGY : 0)
3230 | (!EI_VV(ir->eI) ? CGLO_TEMPERATURE : 0)
3231 | (!EI_VV(ir->eI) || bRerunMD ? CGLO_PRESSURE : 0)
3232 | (bIterations && iterate.bIterate ? CGLO_ITERATE : 0)
3233 | (bFirstIterate ? CGLO_FIRSTITERATE : 0)
3236 if (ir->nstlist == -1 && bFirstIterate)
3238 nlh.nabnsb = gs.set[eglsNABNSB];
3239 gs.set[eglsNABNSB] = 0;
3241 /* bIterate is set to keep it from eliminating the old ekin kinetic energy terms */
3242 /* ############# END CALC EKIN AND PRESSURE ################# */
3244 /* Note: this is OK, but there are some numerical precision issues with using the convergence of
3245 the virial that should probably be addressed eventually. state->veta has better properies,
3246 but what we actually need entering the new cycle is the new shake_vir value. Ideally, we could
3247 generate the new shake_vir, but test the veta value for convergence. This will take some thought. */
3250 done_iterating(cr,fplog,step,&iterate,bFirstIterate,
3251 trace(shake_vir),&tracevir))
3255 bFirstIterate = FALSE;
3258 update_box(fplog,step,ir,mdatoms,state,graph,f,
3259 ir->nstlist==-1 ? &nlh.scale_tot : NULL,pcoupl_mu,nrnb,wcycle,upd,bInitStep,FALSE);
3261 /* ################# END UPDATE STEP 2 ################# */
3262 /* #### We now have r(t+dt) and v(t+dt/2) ############# */
3264 /* The coordinates (x) were unshifted in update */
3265 /* if (bFFscan && (shellfc==NULL || bConverged))
3267 if (print_forcefield(fplog,enerd->term,mdatoms->homenr,
3269 &(top_global->mols),mdatoms->massT,pres))
3271 if (gmx_parallel_env_initialized())
3275 fprintf(stderr,"\n");
3281 /* We will not sum ekinh_old,
3282 * so signal that we still have to do it.
3284 bSumEkinhOld = TRUE;
3289 /* Only do GCT when the relaxation of shells (minimization) has converged,
3290 * otherwise we might be coupling to bogus energies.
3291 * In parallel we must always do this, because the other sims might
3295 /* Since this is called with the new coordinates state->x, I assume
3296 * we want the new box state->box too. / EL 20040121
3298 /* do_coupling(fplog,oenv,nfile,fnm,tcr,t,step,enerd->term,fr,
3300 mdatoms,&(top->idef),mu_aver,
3301 top_global->mols.nr,cr,
3302 state->box,total_vir,pres,
3303 mu_tot,state->x,f,bConverged);
3307 /* ######### BEGIN PREPARING EDR OUTPUT ########### */
3309 sum_dhdl(enerd,state->lambda,ir);
3310 /* use the directly determined last velocity, not actually the averaged half steps */
3311 if (bTrotter && ir->eI==eiVV)
3313 enerd->term[F_EKIN] = last_ekin;
3315 enerd->term[F_ETOT] = enerd->term[F_EPOT] + enerd->term[F_EKIN];
3324 if (IR_NVT_TROTTER(ir)) {
3325 enerd->term[F_ECONSERVED] = enerd->term[F_ETOT] + last_conserved;
3327 enerd->term[F_ECONSERVED] = enerd->term[F_ETOT] +
3328 NPT_energy(ir,state,&MassQ);
3332 enerd->term[F_ECONSERVED] =
3333 enerd->term[F_ETOT] + vrescale_energy(&(ir->opts),
3334 state->therm_integral);
3340 /* Check for excessively large energies */
3344 real etot_max = 1e200;
3346 real etot_max = 1e30;
3348 if (fabs(enerd->term[F_ETOT]) > etot_max)
3350 fprintf(stderr,"Energy too large (%g), giving up\n",
3351 enerd->term[F_ETOT]);
3354 /* ######### END PREPARING EDR OUTPUT ########### */
3356 /* Time for performance */
3357 if (((step % stepout) == 0) || bLastStep)
3359 runtime_upd_proc(runtime);
3365 gmx_bool do_dr,do_or;
3367 if (!(bStartingFromCpt && (EI_VV(ir->eI))))
3371 upd_mdebin(mdebin,bDoDHDL,TRUE,
3372 t,mdatoms->tmass,enerd,state,lastbox,
3373 shake_vir,force_vir,total_vir,pres,
3374 ekind,mu_tot,constr);
3378 upd_mdebin_step(mdebin);
3381 do_dr = do_per_step(step,ir->nstdisreout);
3382 do_or = do_per_step(step,ir->nstorireout);
3384 print_ebin(outf->fp_ene,do_ene,do_dr,do_or,do_log?fplog:NULL,
3386 eprNORMAL,bCompact,mdebin,fcd,groups,&(ir->opts));
3388 if (ir->ePull != epullNO)
3390 pull_print_output(ir->pull,step,t);
3393 if (do_per_step(step,ir->nstlog))
3395 if(fflush(fplog) != 0)
3397 gmx_fatal(FARGS,"Cannot flush logfile - maybe you are out of quota?");
3403 /* Remaining runtime */
3404 if (MULTIMASTER(cr) && (do_verbose || gmx_got_usr_signal() ))
3408 fprintf(stderr,"\n");
3410 print_time(stderr,runtime,step,ir,cr);
3413 /* Set new positions for the group to embed */
3419 } else if (step_rel<=(it_xy+it_z))
3423 resize(ins_at,r_ins,state_global->x,pos_ins,fac);
3426 /* Replica exchange */
3427 /* bExchanged = FALSE;
3428 if ((repl_ex_nst > 0) && (step > 0) && !bLastStep &&
3429 do_per_step(step,repl_ex_nst))
3431 bExchanged = replica_exchange(fplog,cr,repl_ex,
3432 state_global,enerd->term,
3435 if (bExchanged && PAR(cr))
3437 if (DOMAINDECOMP(cr))
3439 dd_partition_system(fplog,step,cr,TRUE,1,
3440 state_global,top_global,ir,
3441 state,&f,mdatoms,top,fr,
3442 vsite,shellfc,constr,
3447 bcast_state(cr,state,FALSE);
3453 bStartingFromCpt = FALSE;
3455 /* ####### SET VARIABLES FOR NEXT ITERATION IF THEY STILL NEED IT ###### */
3456 /* Complicated conditional when bGStatEveryStep=FALSE.
3457 * We can not just use bGStat, since then the simulation results
3458 * would depend on nstenergy and nstlog or step_nscheck.
3460 if (((state->flags & (1<<estPRES_PREV)) ||
3461 (state->flags & (1<<estSVIR_PREV)) ||
3462 (state->flags & (1<<estFVIR_PREV))) &&
3464 (ir->nstlist > 0 && step % ir->nstlist == 0) ||
3465 (ir->nstlist < 0 && nlh.nabnsb > 0) ||
3466 (ir->nstlist == 0 && bGStat)))
3468 /* Store the pressure in t_state for pressure coupling
3469 * at the next MD step.
3471 if (state->flags & (1<<estPRES_PREV))
3473 copy_mat(pres,state->pres_prev);
3477 /* ####### END SET VARIABLES FOR NEXT ITERATION ###### */
3481 /* read next frame from input trajectory */
3482 bNotLastFrame = read_next_frame(oenv,status,&rerun_fr);
3485 if (!bRerunMD || !rerun_fr.bStep)
3487 /* increase the MD step number */
3492 cycles = wallcycle_stop(wcycle,ewcSTEP);
3493 if (DOMAINDECOMP(cr) && wcycle)
3495 dd_cycles_add(cr->dd,cycles,ddCyclStep);
3498 if (step_rel == wcycle_get_reset_counters(wcycle) ||
3499 gs.set[eglsRESETCOUNTERS] != 0)
3501 /* Reset all the counters related to performance over the run */
3502 reset_all_counters(fplog,cr,step,&step_rel,ir,wcycle,nrnb,runtime);
3503 wcycle_set_reset_counters(wcycle,-1);
3504 bResetCountersHalfMaxH = FALSE;
3505 gs.set[eglsRESETCOUNTERS] = 0;
3508 /* End of main MD loop */
3510 write_sto_conf_mtop(ftp2fn(efSTO,nfile,fnm),
3511 *top_global->name,top_global,
3512 state_global->x,state_global->v,
3513 ir->ePBC,state->box);
3516 runtime_end(runtime);
3523 if (!(cr->duty & DUTY_PME))
3525 /* Tell the PME only node to finish */
3531 if (ir->nstcalcenergy > 0 && !bRerunMD)
3533 print_ebin(outf->fp_ene,FALSE,FALSE,FALSE,fplog,step,t,
3534 eprAVER,FALSE,mdebin,fcd,groups,&(ir->opts));
3542 if (ir->nstlist == -1 && nlh.nns > 0 && fplog)
3544 fprintf(fplog,"Average neighborlist lifetime: %.1f steps, std.dev.: %.1f steps\n",nlh.s1/nlh.nns,sqrt(nlh.s2/nlh.nns - sqr(nlh.s1/nlh.nns)));
3545 fprintf(fplog,"Average number of atoms that crossed the half buffer length: %.1f\n\n",nlh.ab/nlh.nns);
3548 if (shellfc && fplog)
3550 fprintf(fplog,"Fraction of iterations that converged: %.2f %%\n",
3551 (nconverged*100.0)/step_rel);
3552 fprintf(fplog,"Average number of force evaluations per MD step: %.2f\n\n",
3556 /* if (repl_ex_nst > 0 && MASTER(cr))
3558 print_replica_exchange_statistics(fplog,repl_ex);
3561 runtime->nsteps_done = step_rel;
3567 int mdrunner_membed(FILE *fplog,t_commrec *cr,int nfile,const t_filenm fnm[],
3568 const output_env_t oenv, gmx_bool bVerbose,gmx_bool bCompact,
3570 ivec ddxyz,int dd_node_order,real rdd,real rconstr,
3571 const char *dddlb_opt,real dlb_scale,
3572 const char *ddcsx,const char *ddcsy,const char *ddcsz,
3573 int nstepout,int resetstep,int nmultisim,int repl_ex_nst,int repl_ex_seed,
3574 real pforce,real cpt_period,real max_hours,
3575 const char *deviceOptions,
3576 unsigned long Flags,
3577 real xy_fac, real xy_max, real z_fac, real z_max,
3578 int it_xy, int it_z, real probe_rad, int low_up_rm,
3579 int pieces, gmx_bool bALLOW_ASYMMETRY, int maxwarn)
3581 double nodetime=0,realtime;
3582 t_inputrec *inputrec;
3583 t_state *state=NULL;
3586 int npme_major,npme_minor;
3589 gmx_mtop_t *mtop=NULL;
3590 t_mdatoms *mdatoms=NULL;
3591 t_forcerec *fr=NULL;
3594 gmx_pme_t *pmedata=NULL;
3595 gmx_vsite_t *vsite=NULL;
3596 gmx_constr_t constr;
3597 int i,m,nChargePerturbed=-1,status,nalloc;
3599 gmx_wallcycle_t wcycle;
3600 gmx_bool bReadRNG,bReadEkin;
3602 gmx_runtime_t runtime;
3604 gmx_large_int_t reset_counters;
3605 gmx_edsam_t ed=NULL;
3606 t_commrec *cr_old=cr;
3607 int nthreads=1,nthreads_requested=1;
3611 int rm_bonded_at,fr_id,fr_i=0,tmp_id,warn=0;
3612 int ng,j,max_lip_rm,ins_grp_id,ins_nat,mem_nat,ntype,lip_rm,tpr_version;
3613 real xy_step=0,z_step=0;
3615 rvec *r_ins=NULL,fac;
3616 t_block *ins_at,*rest_at;
3620 gmx_groups_t *groups;
3621 gmx_bool bExcl=FALSE;
3624 char **piecename=NULL;
3626 /* CAUTION: threads may be started later on in this function, so
3627 cr doesn't reflect the final parallel state right now */
3631 if (bVerbose && SIMMASTER(cr))
3633 fprintf(stderr,"Getting Loaded...\n");
3636 if (Flags & MD_APPENDFILES)
3644 /* Read (nearly) all data required for the simulation */
3645 read_tpx_state(ftp2fn(efTPX,nfile,fnm),inputrec,state,NULL,mtop);
3647 /* NOW the threads will be started: */
3651 /* END OF CAUTION: cr is now reliable */
3655 /* now broadcast everything to the non-master nodes/threads: */
3656 init_parallel(fplog, cr, inputrec, mtop);
3658 /* now make sure the state is initialized and propagated */
3659 set_state_entries(state,inputrec,cr->nnodes);
3661 if (can_use_allvsall(inputrec,mtop,TRUE,cr,fplog))
3663 /* All-vs-all loops do not work with domain decomposition */
3664 Flags |= MD_PARTDEC;
3667 if (!EEL_PME(inputrec->coulombtype) || (Flags & MD_PARTDEC))
3676 tpr_version = get_tpr_version(ftp2fn(efTPX,nfile,fnm));
3678 gmx_fatal(FARGS,"Version of *.tpr file to old (%d). Rerun grompp with gromacs VERSION 4.0.3 or newer.\n",tpr_version);
3680 if( inputrec->eI != eiMD )
3681 gmx_input("Change integrator to md in mdp file.");
3684 gmx_input("Sorry, parallel g_membed is not yet fully functrional.");
3686 groups=&(mtop->groups);
3688 atoms=gmx_mtop_global_atoms(mtop);
3690 fprintf(stderr,"\nSelect a group to embed in the membrane:\n");
3691 get_index(&atoms,ftp2fn_null(efNDX,nfile,fnm),1,&(ins_at->nr),&(ins_at->index),&ins);
3692 ins_grp_id = search_string(ins,groups->ngrpname,(groups->grpname));
3693 fprintf(stderr,"\nSelect a group to embed %s into (e.g. the membrane):\n",ins);
3694 get_index(&atoms,ftp2fn_null(efNDX,nfile,fnm),1,&(mem_p->mem_at.nr),&(mem_p->mem_at.index),&(mem_p->name));
3696 pos_ins->pieces=pieces;
3697 snew(pos_ins->nidx,pieces);
3698 snew(pos_ins->subindex,pieces);
3699 snew(piecename,pieces);
3702 fprintf(stderr,"\nSelect pieces to embed:\n");
3703 get_index(&atoms,ftp2fn_null(efNDX,nfile,fnm),pieces,pos_ins->nidx,pos_ins->subindex,piecename);
3707 /*use whole embedded group*/
3708 snew(pos_ins->nidx,1);
3709 snew(pos_ins->subindex,1);
3710 pos_ins->nidx[0]=ins_at->nr;
3711 pos_ins->subindex[0]=ins_at->index;
3714 if(probe_rad<0.2199999)
3717 fprintf(stderr,"\nWarning %d:\nA probe radius (-rad) smaller than 0.2 can result in overlap between waters "
3718 "and the group to embed, which will result in Lincs errors etc.\nIf you are sure, you can increase maxwarn.\n\n",warn);
3721 if(xy_fac<0.09999999)
3724 fprintf(stderr,"\nWarning %d:\nThe initial size of %s is probably too smal.\n"
3725 "If you are sure, you can increase maxwarn.\n\n",warn,ins);
3731 fprintf(stderr,"\nWarning %d;\nThe number of steps used to grow the xy-coordinates of %s (%d) is probably too small.\n"
3732 "Increase -nxy or, if you are sure, you can increase maxwarn.\n\n",warn,ins,it_xy);
3735 if( (it_z<100) && ( z_fac<0.99999999 || z_fac>1.0000001) )
3738 fprintf(stderr,"\nWarning %d;\nThe number of steps used to grow the z-coordinate of %s (%d) is probably too small.\n"
3739 "Increase -nz or, if you are sure, you can increase maxwarn.\n\n",warn,ins,it_z);
3742 if(it_xy+it_z>inputrec->nsteps)
3745 fprintf(stderr,"\nWarning %d:\nThe number of growth steps (-nxy + -nz) is larger than the number of steps in the tpr.\n"
3746 "If you are sure, you can increase maxwarn.\n\n",warn);
3750 if( inputrec->opts.ngfrz==1)
3751 gmx_fatal(FARGS,"You did not specify \"%s\" as a freezegroup.",ins);
3752 for(i=0;i<inputrec->opts.ngfrz;i++)
3754 tmp_id = mtop->groups.grps[egcFREEZE].nm_ind[i];
3755 if(ins_grp_id==tmp_id)
3762 gmx_fatal(FARGS,"\"%s\" not as freezegroup defined in the mdp-file.",ins);
3765 if( inputrec->opts.nFreeze[fr_i][i] != 1)
3766 gmx_fatal(FARGS,"freeze dimensions for %s are not Y Y Y\n",ins);
3768 ng = groups->grps[egcENER].nr;
3770 gmx_input("No energy groups defined. This is necessary for energy exclusion in the freeze group");
3776 if (inputrec->opts.egp_flags[ng*i+j] == EGP_EXCL)
3779 if ( (groups->grps[egcENER].nm_ind[i] != ins_grp_id) || (groups->grps[egcENER].nm_ind[j] != ins_grp_id) )
3780 gmx_fatal(FARGS,"Energy exclusions \"%s\" and \"%s\" do not match the group to embed \"%s\"",
3781 *groups->grpname[groups->grps[egcENER].nm_ind[i]],
3782 *groups->grpname[groups->grps[egcENER].nm_ind[j]],ins);
3787 gmx_input("No energy exclusion groups defined. This is necessary for energy exclusion in the freeze group");
3789 /* Set all atoms in box*/
3790 /*set_inbox(state->natoms,state->x);*/
3792 /* Guess the area the protein will occupy in the membrane plane Calculate area per lipid*/
3794 ins_nat = init_ins_at(ins_at,rest_at,state,pos_ins,groups,ins_grp_id,xy_max);
3795 /* Check moleculetypes in insertion group */
3796 check_types(ins_at,rest_at,mtop);
3798 mem_nat = init_mem_at(mem_p,mtop,state->x,state->box,pos_ins);
3800 prot_area = est_prot_area(pos_ins,state->x,ins_at,mem_p);
3801 if ( (prot_area>7.5) && ( (state->box[XX][XX]*state->box[YY][YY]-state->box[XX][YY]*state->box[YY][XX])<50) )
3804 fprintf(stderr,"\nWarning %d:\nThe xy-area is very small compared to the area of the protein.\n"
3805 "This might cause pressure problems during the growth phase. Just try with\n"
3806 "current setup (-maxwarn + 1), but if pressure problems occur, lower the\n"
3807 "compressibility in the mdp-file or use no pressure coupling at all.\n\n",warn);
3810 gmx_fatal(FARGS,"Too many warnings.\n");
3812 printf("The estimated area of the protein in the membrane is %.3f nm^2\n",prot_area);
3813 printf("\nThere are %d lipids in the membrane part that overlaps the protein.\nThe area per lipid is %.4f nm^2.\n",mem_p->nmol,mem_p->lip_area);
3815 /* Maximum number of lipids to be removed*/
3816 max_lip_rm=(int)(2*prot_area/mem_p->lip_area);
3817 printf("Maximum number of lipids that will be removed is %d.\n",max_lip_rm);
3819 printf("\nWill resize the protein by a factor of %.3f in the xy plane and %.3f in the z direction.\n"
3820 "This resizing will be done with respect to the geometrical center of all protein atoms\n"
3821 "that span the membrane region, i.e. z between %.3f and %.3f\n\n",xy_fac,z_fac,mem_p->zmin,mem_p->zmax);
3823 /* resize the protein by xy and by z if necessary*/
3824 snew(r_ins,ins_at->nr);
3825 init_resize(ins_at,r_ins,pos_ins,mem_p,state->x,bALLOW_ASYMMETRY);
3826 fac[0]=fac[1]=xy_fac;
3829 xy_step =(xy_max-xy_fac)/(double)(it_xy);
3830 z_step =(z_max-z_fac)/(double)(it_z-1);
3832 resize(ins_at,r_ins,state->x,pos_ins,fac);
3834 /* remove overlapping lipids and water from the membrane box*/
3835 /*mark molecules to be removed*/
3837 set_pbc(pbc,inputrec->ePBC,state->box);
3840 lip_rm = gen_rm_list(rm_p,ins_at,rest_at,pbc,mtop,state->x, r_ins, mem_p,pos_ins,probe_rad,low_up_rm,bALLOW_ASYMMETRY);
3841 lip_rm -= low_up_rm;
3844 for(i=0;i<rm_p->nr;i++)
3845 fprintf(fplog,"rm mol %d\n",rm_p->mol[i]);
3847 for(i=0;i<mtop->nmolblock;i++)
3850 for(j=0;j<rm_p->nr;j++)
3851 if(rm_p->block[j]==i)
3853 printf("Will remove %d %s molecules\n",ntype,*(mtop->moltype[mtop->molblock[i].type].name));
3856 if(lip_rm>max_lip_rm)
3859 fprintf(stderr,"\nWarning %d:\nTrying to remove a larger lipid area than the estimated protein area\n"
3860 "Try making the -xyinit resize factor smaller. If you are sure about this increase maxwarn.\n\n",warn);
3863 /*remove all lipids and waters overlapping and update all important structures*/
3864 rm_group(inputrec,groups,mtop,rm_p,state,ins_at,pos_ins);
3866 rm_bonded_at = rm_bonded(ins_at,mtop);
3867 if (rm_bonded_at != ins_at->nr)
3869 fprintf(stderr,"Warning: The number of atoms for which the bonded interactions are removed is %d, "
3870 "while %d atoms are embedded. Make sure that the atoms to be embedded are not in the same"
3871 "molecule type as atoms that are not to be embedded.\n",rm_bonded_at,ins_at->nr);
3875 gmx_fatal(FARGS,"Too many warnings.\nIf you are sure these warnings are harmless, you can increase -maxwarn");
3879 if (ftp2bSet(efTOP,nfile,fnm))
3880 top_update(opt2fn("-p",nfile,fnm),ins,rm_p,mtop);
3888 fcRegisterSteps(inputrec->nsteps,inputrec->init_step);
3891 /* NMR restraints must be initialized before load_checkpoint,
3892 * since with time averaging the history is added to t_state.
3893 * For proper consistency check we therefore need to extend
3895 * So the PME-only nodes (if present) will also initialize
3896 * the distance restraints.
3900 /* This needs to be called before read_checkpoint to extend the state */
3901 init_disres(fplog,mtop,inputrec,cr,Flags & MD_PARTDEC,fcd,state);
3903 if (gmx_mtop_ftype_count(mtop,F_ORIRES) > 0)
3905 if (PAR(cr) && !(Flags & MD_PARTDEC))
3907 gmx_fatal(FARGS,"Orientation restraints do not work (yet) with domain decomposition, use particle decomposition (mdrun option -pd)");
3909 /* Orientation restraints */
3912 init_orires(fplog,mtop,state->x,inputrec,cr->ms,&(fcd->orires),
3917 if (DEFORM(*inputrec))
3919 /* Store the deform reference box before reading the checkpoint */
3922 copy_mat(state->box,box);
3926 gmx_bcast(sizeof(box),box,cr);
3928 /* Because we do not have the update struct available yet
3929 * in which the reference values should be stored,
3930 * we store them temporarily in static variables.
3931 * This should be thread safe, since they are only written once
3932 * and with identical values.
3934 /* deform_init_init_step_tpx = inputrec->init_step;*/
3935 /* copy_mat(box,deform_init_box_tpx);*/
3938 if (opt2bSet("-cpi",nfile,fnm))
3940 /* Check if checkpoint file exists before doing continuation.
3941 * This way we can use identical input options for the first and subsequent runs...
3943 if( gmx_fexist_master(opt2fn_master("-cpi",nfile,fnm,cr),cr) )
3945 load_checkpoint(opt2fn_master("-cpi",nfile,fnm,cr),&fplog,
3946 cr,Flags & MD_PARTDEC,ddxyz,
3947 inputrec,state,&bReadRNG,&bReadEkin,
3948 (Flags & MD_APPENDFILES));
3952 Flags |= MD_READ_RNG;
3956 Flags |= MD_READ_EKIN;
3961 if ((MASTER(cr) || (Flags & MD_SEPPOT)) && (Flags & MD_APPENDFILES))
3963 gmx_log_open(ftp2fn(efLOG,nfile,fnm),cr,!(Flags & MD_SEPPOT),
3969 copy_mat(state->box,box);
3974 gmx_bcast(sizeof(box),box,cr);
3977 if (bVerbose && SIMMASTER(cr))
3979 fprintf(stderr,"Loaded with Money\n\n");
3982 if (PAR(cr) && !((Flags & MD_PARTDEC) || EI_TPI(inputrec->eI)))
3984 cr->dd = init_domain_decomposition(fplog,cr,Flags,ddxyz,rdd,rconstr,
3985 dddlb_opt,dlb_scale,
3989 &ddbox,&npme_major,&npme_minor);
3991 make_dd_communicators(fplog,cr,dd_node_order);
3993 /* Set overallocation to avoid frequent reallocation of arrays */
3994 set_over_alloc_dd(TRUE);
3998 /* PME, if used, is done on all nodes with 1D decomposition */
4000 cr->duty = (DUTY_PP | DUTY_PME);
4001 npme_major = cr->nnodes;
4004 if (inputrec->ePBC == epbcSCREW)
4007 "pbc=%s is only implemented with domain decomposition",
4008 epbc_names[inputrec->ePBC]);
4014 /* After possible communicator splitting in make_dd_communicators.
4015 * we can set up the intra/inter node communication.
4017 gmx_setup_nodecomm(fplog,cr);
4020 wcycle = wallcycle_init(fplog,resetstep,cr);
4023 /* Master synchronizes its value of reset_counters with all nodes
4024 * including PME only nodes */
4025 reset_counters = wcycle_get_reset_counters(wcycle);
4026 gmx_bcast_sim(sizeof(reset_counters),&reset_counters,cr);
4027 wcycle_set_reset_counters(wcycle, reset_counters);
4032 if (cr->duty & DUTY_PP)
4034 /* For domain decomposition we allocate dynamically
4035 * in dd_partition_system.
4037 if (DOMAINDECOMP(cr))
4039 bcast_state_setup(cr,state);
4049 bcast_state(cr,state,TRUE);
4053 /* Dihedral Restraints */
4054 if (gmx_mtop_ftype_count(mtop,F_DIHRES) > 0)
4056 init_dihres(fplog,mtop,inputrec,fcd);
4059 /* Initiate forcerecord */
4061 init_forcerec(fplog,oenv,fr,fcd,inputrec,mtop,cr,box,FALSE,
4062 opt2fn("-table",nfile,fnm),
4063 opt2fn("-tablep",nfile,fnm),
4064 opt2fn("-tableb",nfile,fnm),FALSE,pforce);
4066 /* version for PCA_NOT_READ_NODE (see md.c) */
4067 /*init_forcerec(fplog,fr,fcd,inputrec,mtop,cr,box,FALSE,
4068 "nofile","nofile","nofile",FALSE,pforce);
4070 fr->bSepDVDL = ((Flags & MD_SEPPOT) == MD_SEPPOT);
4072 /* Initialize QM-MM */
4075 init_QMMMrec(cr,box,mtop,inputrec,fr);
4078 /* Initialize the mdatoms structure.
4079 * mdatoms is not filled with atom data,
4080 * as this can not be done now with domain decomposition.
4082 mdatoms = init_mdatoms(fplog,mtop,inputrec->efep!=efepNO);
4084 /* Initialize the virtual site communication */
4085 vsite = init_vsite(mtop,cr);
4087 calc_shifts(box,fr->shift_vec);
4089 /* With periodic molecules the charge groups should be whole at start up
4090 * and the virtual sites should not be far from their proper positions.
4092 if (!inputrec->bContinuation && MASTER(cr) &&
4093 !(inputrec->ePBC != epbcNONE && inputrec->bPeriodicMols))
4095 /* Make molecules whole at start of run */
4096 if (fr->ePBC != epbcNONE)
4098 do_pbc_first_mtop(fplog,inputrec->ePBC,box,mtop,state->x);
4102 /* Correct initial vsite positions are required
4103 * for the initial distribution in the domain decomposition
4104 * and for the initial shell prediction.
4106 construct_vsites_mtop(fplog,vsite,mtop,state->x);
4110 /* Initiate PPPM if necessary */
4111 if (fr->eeltype == eelPPPM)
4113 if (mdatoms->nChargePerturbed)
4115 gmx_fatal(FARGS,"Free energy with %s is not implemented",
4116 eel_names[fr->eeltype]);
4118 status = gmx_pppm_init(fplog,cr,oenv,FALSE,TRUE,box,
4119 getenv("GMXGHAT"),inputrec, (Flags & MD_REPRODUCIBLE));
4122 gmx_fatal(FARGS,"Error %d initializing PPPM",status);
4126 if (EEL_PME(fr->eeltype))
4128 ewaldcoeff = fr->ewaldcoeff;
4129 pmedata = &fr->pmedata;
4138 /* This is a PME only node */
4140 /* We don't need the state */
4143 ewaldcoeff = calc_ewaldcoeff(inputrec->rcoulomb, inputrec->ewald_rtol);
4147 /* Initiate PME if necessary,
4148 * either on all nodes or on dedicated PME nodes only. */
4149 if (EEL_PME(inputrec->coulombtype))
4153 nChargePerturbed = mdatoms->nChargePerturbed;
4155 if (cr->npmenodes > 0)
4157 /* The PME only nodes need to know nChargePerturbed */
4158 gmx_bcast_sim(sizeof(nChargePerturbed),&nChargePerturbed,cr);
4160 if (cr->duty & DUTY_PME)
4162 status = gmx_pme_init(pmedata,cr,npme_major,npme_minor,inputrec,
4163 mtop ? mtop->natoms : 0,nChargePerturbed,
4164 (Flags & MD_REPRODUCIBLE));
4167 gmx_fatal(FARGS,"Error %d initializing PME",status);
4173 /* if (integrator[inputrec->eI].func == do_md
4176 integrator[inputrec->eI].func == do_md_openmm
4180 /* Turn on signal handling on all nodes */
4182 * (A user signal from the PME nodes (if any)
4183 * is communicated to the PP nodes.
4185 signal_handler_install();
4188 if (cr->duty & DUTY_PP)
4190 if (inputrec->ePull != epullNO)
4192 /* Initialize pull code */
4193 init_pull(fplog,inputrec,nfile,fnm,mtop,cr,oenv,
4194 EI_DYNAMICS(inputrec->eI) && MASTER(cr),Flags);
4197 constr = init_constraints(fplog,mtop,inputrec,ed,state,cr);
4199 if (DOMAINDECOMP(cr))
4201 dd_init_bondeds(fplog,cr->dd,mtop,vsite,constr,inputrec,
4202 Flags & MD_DDBONDCHECK,fr->cginfo_mb);
4204 set_dd_parameters(fplog,cr->dd,dlb_scale,inputrec,fr,&ddbox);
4206 setup_dd_grid(fplog,cr->dd);
4209 /* Now do whatever the user wants us to do (how flexible...) */
4210 do_md_membed(fplog,cr,nfile,fnm,
4211 oenv,bVerbose,bCompact,
4214 nstepout,inputrec,mtop,
4216 mdatoms,nrnb,wcycle,ed,fr,
4217 repl_ex_nst,repl_ex_seed,
4218 cpt_period,max_hours,
4222 fac, r_ins, pos_ins, ins_at,
4223 xy_step, z_step, it_xy, it_z);
4225 if (inputrec->ePull != epullNO)
4227 finish_pull(fplog,inputrec->pull);
4233 gmx_pmeonly(*pmedata,cr,nrnb,wcycle,ewaldcoeff,FALSE,inputrec);
4236 if (EI_DYNAMICS(inputrec->eI) || EI_TPI(inputrec->eI))
4238 /* Some timing stats */
4241 if (runtime.proc == 0)
4243 runtime.proc = runtime.real;
4252 wallcycle_stop(wcycle,ewcRUN);
4254 /* Finish up, write some stuff
4255 * if rerunMD, don't write last frame again
4257 finish_run(fplog,cr,ftp2fn(efSTO,nfile,fnm),
4258 inputrec,nrnb,wcycle,&runtime,
4259 EI_DYNAMICS(inputrec->eI) && !MULTISIM(cr));
4261 /* Does what it says */
4262 print_date_and_time(fplog,cr->nodeid,"Finished mdrun",&runtime);
4264 /* Close logfile already here if we were appending to it */
4265 if (MASTER(cr) && (Flags & MD_APPENDFILES))
4267 gmx_log_close(fplog);
4275 rc=(int)gmx_get_stop_condition();
4280 int gmx_membed(int argc,char *argv[])
4282 const char *desc[] = {
4283 "g_membed embeds a membrane protein into an equilibrated lipid bilayer at the position",
4284 "and orientation specified by the user.\n",
4286 "SHORT MANUAL\n------------\n",
4287 "The user should merge the structure files of the protein and membrane (+solvent), creating a",
4288 "single structure file with the protein overlapping the membrane at the desired position and",
4289 "orientation. Box size should be taken from the membrane structure file. The corresponding topology",
4290 "files should also be merged. Consecutively, create a tpr file (input for g_membed) from these files,"
4291 "with the following options included in the mdp file.\n",
4292 " - integrator = md\n",
4293 " - energygrp = Protein (or other group that you want to insert)\n",
4294 " - freezegrps = Protein\n",
4295 " - freezedim = Y Y Y\n",
4296 " - energygrp_excl = Protein Protein\n",
4297 "The output is a structure file containing the protein embedded in the membrane. If a topology",
4298 "file is provided, the number of lipid and ",
4299 "solvent molecules will be updated to match the new structure file.\n",
4300 "For a more extensive manual see Wolf et al, J Comp Chem 31 (2010) 2169-2174, Appendix.\n",
4302 "SHORT METHOD DESCRIPTION\n",
4303 "------------------------\n",
4304 "1. The protein is resized around its center of mass by a factor -xy in the xy-plane",
4305 "(the membrane plane) and a factor -z in the z-direction (if the size of the",
4306 "protein in the z-direction is the same or smaller than the width of the membrane, a",
4307 "-z value larger than 1 can prevent that the protein will be enveloped by the lipids).\n",
4308 "2. All lipid and solvent molecules overlapping with the resized protein are removed. All",
4309 "intraprotein interactions are turned off to prevent numerical issues for small values of -xy",
4311 "3. One md step is performed.\n",
4312 "4. The resize factor (-xy or -z) is incremented by a small amount ((1-xy)/nxy or (1-z)/nz) and the",
4313 "protein is resized again around its center of mass. The resize factor for the xy-plane",
4314 "is incremented first. The resize factor for the z-direction is not changed until the -xy factor",
4315 "is 1 (thus after -nxy iteration).\n",
4316 "5. Repeat step 3 and 4 until the protein reaches its original size (-nxy + -nz iterations).\n",
4317 "For a more extensive method descrition see Wolf et al, J Comp Chem, 31 (2010) 2169-2174.\n",
4320 " - Protein can be any molecule you want to insert in the membrane.\n",
4321 " - It is recommended to perform a short equilibration run after the embedding",
4322 "(see Wolf et al, J Comp Chem 31 (2010) 2169-2174, to re-equilibrate the membrane. Clearly",
4323 "protein equilibration might require longer.\n",
4328 { efTPX, "-f", "into_mem", ffREAD },
4329 { efNDX, "-n", "index", ffOPTRD },
4330 { efTOP, "-p", "topol", ffOPTRW },
4331 { efTRN, "-o", NULL, ffWRITE },
4332 { efXTC, "-x", NULL, ffOPTWR },
4333 { efCPT, "-cpi", NULL, ffOPTRD },
4334 { efCPT, "-cpo", NULL, ffOPTWR },
4335 { efSTO, "-c", "membedded", ffWRITE },
4336 { efEDR, "-e", "ener", ffWRITE },
4337 { efLOG, "-g", "md", ffWRITE },
4338 { efEDI, "-ei", "sam", ffOPTRD },
4339 { efTRX, "-rerun", "rerun", ffOPTRD },
4340 { efXVG, "-table", "table", ffOPTRD },
4341 { efXVG, "-tablep", "tablep", ffOPTRD },
4342 { efXVG, "-tableb", "table", ffOPTRD },
4343 { efXVG, "-dhdl", "dhdl", ffOPTWR },
4344 { efXVG, "-field", "field", ffOPTWR },
4345 { efXVG, "-table", "table", ffOPTRD },
4346 { efXVG, "-tablep", "tablep", ffOPTRD },
4347 { efXVG, "-tableb", "table", ffOPTRD },
4348 { efTRX, "-rerun", "rerun", ffOPTRD },
4349 { efXVG, "-tpi", "tpi", ffOPTWR },
4350 { efXVG, "-tpid", "tpidist", ffOPTWR },
4351 { efEDI, "-ei", "sam", ffOPTRD },
4352 { efEDO, "-eo", "sam", ffOPTWR },
4353 { efGCT, "-j", "wham", ffOPTRD },
4354 { efGCT, "-jo", "bam", ffOPTWR },
4355 { efXVG, "-ffout", "gct", ffOPTWR },
4356 { efXVG, "-devout", "deviatie", ffOPTWR },
4357 { efXVG, "-runav", "runaver", ffOPTWR },
4358 { efXVG, "-px", "pullx", ffOPTWR },
4359 { efXVG, "-pf", "pullf", ffOPTWR },
4360 { efMTX, "-mtx", "nm", ffOPTWR },
4361 { efNDX, "-dn", "dipole", ffOPTWR }
4363 #define NFILE asize(fnm)
4365 /* Command line options ! */
4366 gmx_bool bCart = FALSE;
4367 gmx_bool bPPPME = FALSE;
4368 gmx_bool bPartDec = FALSE;
4369 gmx_bool bDDBondCheck = TRUE;
4370 gmx_bool bDDBondComm = TRUE;
4371 gmx_bool bVerbose = FALSE;
4372 gmx_bool bCompact = TRUE;
4373 gmx_bool bSepPot = FALSE;
4374 gmx_bool bRerunVSite = FALSE;
4375 gmx_bool bIonize = FALSE;
4376 gmx_bool bConfout = TRUE;
4377 gmx_bool bReproducible = FALSE;
4381 int nstglobalcomm=-1;
4383 int repl_ex_seed=-1;
4385 int nthreads=0; /* set to determine # of threads automatically */
4388 rvec realddxyz={0,0,0};
4389 const char *ddno_opt[ddnoNR+1] =
4390 { NULL, "interleave", "pp_pme", "cartesian", NULL };
4391 const char *dddlb_opt[] =
4392 { NULL, "auto", "no", "yes", NULL };
4393 real rdd=0.0,rconstr=0.0,dlb_scale=0.8,pforce=-1;
4394 char *ddcsx=NULL,*ddcsy=NULL,*ddcsz=NULL;
4395 real cpt_period=15.0,max_hours=-1;
4396 gmx_bool bAppendFiles=TRUE,bAddPart=TRUE;
4397 gmx_bool bResetCountersHalfWay=FALSE;
4398 output_env_t oenv=NULL;
4399 const char *deviceOptions = "";
4407 real probe_rad = 0.22;
4411 gmx_bool bALLOW_ASYMMETRY=FALSE;
4414 /* arguments relevant to OPENMM only*/
4416 gmx_input("g_membed not functional in openmm");
4420 { "-xyinit", FALSE, etREAL, {&xy_fac}, "Resize factor for the protein in the xy dimension before starting embedding" },
4421 { "-xyend", FALSE, etREAL, {&xy_max}, "Final resize factor in the xy dimension" },
4422 { "-zinit", FALSE, etREAL, {&z_fac}, "Resize factor for the protein in the z dimension before starting embedding" },
4423 { "-zend", FALSE, etREAL, {&z_max}, "Final resize faction in the z dimension" },
4424 { "-nxy", FALSE, etINT, {&it_xy}, "Number of iteration for the xy dimension" },
4425 { "-nz", FALSE, etINT, {&it_z}, "Number of iterations for the z dimension" },
4426 { "-rad", FALSE, etREAL, {&probe_rad}, "Probe radius to check for overlap between the group to embed and the membrane"},
4427 { "-pieces", FALSE, etINT, {&pieces}, "Perform piecewise resize. Select parts of the group to insert and resize these with respect to their own geometrical center." },
4428 { "-asymmetry",FALSE, etBOOL,{&bALLOW_ASYMMETRY}, "Allow asymmetric insertion, i.e. the number of lipids removed from the upper and lower leaflet will not be checked." },
4429 { "-ndiff" , FALSE, etINT, {&low_up_rm}, "Number of lipids that will additionally be removed from the lower (negative number) or upper (positive number) membrane leaflet." },
4430 { "-maxwarn", FALSE, etINT, {&maxwarn}, "Maximum number of warning allowed" },
4431 { "-pd", FALSE, etBOOL,{&bPartDec},
4432 "HIDDENUse particle decompostion" },
4433 { "-dd", FALSE, etRVEC,{&realddxyz},
4434 "HIDDENDomain decomposition grid, 0 is optimize" },
4435 { "-nt", FALSE, etINT, {&nthreads},
4436 "HIDDENNumber of threads to start (0 is guess)" },
4437 { "-npme", FALSE, etINT, {&npme},
4438 "HIDDENNumber of separate nodes to be used for PME, -1 is guess" },
4439 { "-ddorder", FALSE, etENUM, {ddno_opt},
4440 "HIDDENDD node order" },
4441 { "-ddcheck", FALSE, etBOOL, {&bDDBondCheck},
4442 "HIDDENCheck for all bonded interactions with DD" },
4443 { "-ddbondcomm", FALSE, etBOOL, {&bDDBondComm},
4444 "HIDDENUse special bonded atom communication when -rdd > cut-off" },
4445 { "-rdd", FALSE, etREAL, {&rdd},
4446 "HIDDENThe maximum distance for bonded interactions with DD (nm), 0 is determine from initial coordinates" },
4447 { "-rcon", FALSE, etREAL, {&rconstr},
4448 "HIDDENMaximum distance for P-LINCS (nm), 0 is estimate" },
4449 { "-dlb", FALSE, etENUM, {dddlb_opt},
4450 "HIDDENDynamic load balancing (with DD)" },
4451 { "-dds", FALSE, etREAL, {&dlb_scale},
4452 "HIDDENMinimum allowed dlb scaling of the DD cell size" },
4453 { "-ddcsx", FALSE, etSTR, {&ddcsx},
4454 "HIDDENThe DD cell sizes in x" },
4455 { "-ddcsy", FALSE, etSTR, {&ddcsy},
4456 "HIDDENThe DD cell sizes in y" },
4457 { "-ddcsz", FALSE, etSTR, {&ddcsz},
4458 "HIDDENThe DD cell sizes in z" },
4459 { "-gcom", FALSE, etINT,{&nstglobalcomm},
4460 "HIDDENGlobal communication frequency" },
4461 { "-compact", FALSE, etBOOL,{&bCompact},
4462 "Write a compact log file" },
4463 { "-seppot", FALSE, etBOOL, {&bSepPot},
4464 "HIDDENWrite separate V and dVdl terms for each interaction type and node to the log file(s)" },
4465 { "-pforce", FALSE, etREAL, {&pforce},
4466 "HIDDENPrint all forces larger than this (kJ/mol nm)" },
4467 { "-reprod", FALSE, etBOOL,{&bReproducible},
4468 "HIDDENTry to avoid optimizations that affect binary reproducibility" },
4469 { "-multi", FALSE, etINT,{&nmultisim},
4470 "HIDDENDo multiple simulations in parallel" },
4471 { "-replex", FALSE, etINT, {&repl_ex_nst},
4472 "HIDDENAttempt replica exchange every # steps" },
4473 { "-reseed", FALSE, etINT, {&repl_ex_seed},
4474 "HIDDENSeed for replica exchange, -1 is generate a seed" },
4475 { "-rerunvsite", FALSE, etBOOL, {&bRerunVSite},
4476 "HIDDENRecalculate virtual site coordinates with -rerun" },
4477 { "-ionize", FALSE, etBOOL,{&bIonize},
4478 "HIDDENDo a simulation including the effect of an X-Ray bombardment on your system" },
4479 { "-confout", TRUE, etBOOL, {&bConfout},
4480 "HIDDENWrite the last configuration with -c and force checkpointing at the last step" },
4481 { "-stepout", FALSE, etINT, {&nstepout},
4482 "HIDDENFrequency of writing the remaining runtime" },
4483 { "-resetstep", FALSE, etINT, {&resetstep},
4484 "HIDDENReset cycle counters after these many time steps" },
4485 { "-resethway", FALSE, etBOOL, {&bResetCountersHalfWay},
4486 "HIDDENReset the cycle counters after half the number of steps or halfway -maxh" },
4487 { "-v", FALSE, etBOOL,{&bVerbose},
4488 "Be loud and noisy" },
4489 { "-maxh", FALSE, etREAL, {&max_hours},
4490 "HIDDENTerminate after 0.99 times this time (hours)" },
4491 { "-cpt", FALSE, etREAL, {&cpt_period},
4492 "HIDDENCheckpoint interval (minutes)" },
4493 { "-append", FALSE, etBOOL, {&bAppendFiles},
4494 "HIDDENAppend to previous output files when continuing from checkpoint" },
4495 { "-addpart", FALSE, etBOOL, {&bAddPart},
4496 "HIDDENAdd the simulation part number to all output files when continuing from checkpoint" },
4499 unsigned long Flags, PCA_Flags;
4502 gmx_bool HaveCheckpoint;
4503 FILE *fplog,*fptest;
4504 int sim_part,sim_part_fn;
4505 const char *part_suffix=".part";
4506 char suffix[STRLEN];
4510 cr = init_par(&argc,&argv);
4512 PCA_Flags = (PCA_KEEP_ARGS | PCA_NOEXIT_ON_ARGS | PCA_CAN_SET_DEFFNM
4513 | (MASTER(cr) ? 0 : PCA_QUIET));
4516 /* Comment this in to do fexist calls only on master
4517 * works not with rerun or tables at the moment
4518 * also comment out the version of init_forcerec in md.c
4519 * with NULL instead of opt2fn
4524 PCA_Flags |= PCA_NOT_READ_NODE;
4528 parse_common_args(&argc,argv,PCA_Flags, NFILE,fnm,asize(pa),pa,
4529 asize(desc),desc,0,NULL, &oenv);
4531 /* we set these early because they might be used in init_multisystem()
4532 Note that there is the potential for npme>nnodes until the number of
4533 threads is set later on, if there's thread parallelization. That shouldn't
4534 lead to problems. */
4535 dd_node_order = nenum(ddno_opt);
4536 cr->npmenodes = npme;
4539 /* now determine the number of threads automatically. The threads are
4540 only started at mdrunner_threads, though. */
4543 nthreads=tMPI_Get_recommended_nthreads();
4550 if (repl_ex_nst != 0 && nmultisim < 2)
4551 gmx_fatal(FARGS,"Need at least two replicas for replica exchange (option -multi)");
4553 if (nmultisim > 1) {
4555 init_multisystem(cr,nmultisim,NFILE,fnm,TRUE);
4557 gmx_fatal(FARGS,"mdrun -multi is not supported with the thread library.Please compile GROMACS with MPI support");
4561 /* Check if there is ANY checkpoint file available */
4563 sim_part_fn = sim_part;
4564 if (opt2bSet("-cpi",NFILE,fnm))
4567 read_checkpoint_simulation_part(opt2fn_master("-cpi", NFILE,fnm,cr),
4568 &sim_part_fn,NULL,cr,
4569 bAppendFiles,NFILE,fnm,
4570 part_suffix,&bAddPart);
4571 if (sim_part_fn==0 && MASTER(cr))
4573 fprintf(stdout,"No previous checkpoint file present, assuming this is a new run.\n");
4577 sim_part = sim_part_fn + 1;
4582 bAppendFiles = FALSE;
4587 sim_part_fn = sim_part;
4590 if (bAddPart && sim_part_fn > 1)
4592 /* This is a continuation run, rename trajectory output files
4593 (except checkpoint files) */
4594 /* create new part name first (zero-filled) */
4595 sprintf(suffix,"%s%04d",part_suffix,sim_part_fn);
4597 add_suffix_to_output_names(fnm,NFILE,suffix);
4598 fprintf(stdout,"Checkpoint file is from part %d, new output files will be suffixed '%s'.\n",sim_part-1,suffix);
4601 Flags = opt2bSet("-rerun",NFILE,fnm) ? MD_RERUN : 0;
4602 Flags = Flags | (bSepPot ? MD_SEPPOT : 0);
4603 Flags = Flags | (bIonize ? MD_IONIZE : 0);
4604 Flags = Flags | (bPartDec ? MD_PARTDEC : 0);
4605 Flags = Flags | (bDDBondCheck ? MD_DDBONDCHECK : 0);
4606 Flags = Flags | (bDDBondComm ? MD_DDBONDCOMM : 0);
4607 Flags = Flags | (bConfout ? MD_CONFOUT : 0);
4608 Flags = Flags | (bRerunVSite ? MD_RERUN_VSITE : 0);
4609 Flags = Flags | (bReproducible ? MD_REPRODUCIBLE : 0);
4610 Flags = Flags | (bAppendFiles ? MD_APPENDFILES : 0);
4611 Flags = Flags | (sim_part>1 ? MD_STARTFROMCPT : 0);
4612 Flags = Flags | (bResetCountersHalfWay ? MD_RESETCOUNTERSHALFWAY : 0);
4615 /* We postpone opening the log file if we are appending, so we can
4616 first truncate the old log file and append to the correct position
4618 if ((MASTER(cr) || bSepPot) && !bAppendFiles)
4620 gmx_log_open(ftp2fn(efLOG,NFILE,fnm),cr,!bSepPot,Flags,&fplog);
4621 CopyRight(fplog,argv[0]);
4622 please_cite(fplog,"Hess2008b");
4623 please_cite(fplog,"Spoel2005a");
4624 please_cite(fplog,"Lindahl2001a");
4625 please_cite(fplog,"Berendsen95a");
4632 ddxyz[XX] = (int)(realddxyz[XX] + 0.5);
4633 ddxyz[YY] = (int)(realddxyz[YY] + 0.5);
4634 ddxyz[ZZ] = (int)(realddxyz[ZZ] + 0.5);
4636 /* even if nthreads = 1, we still call this one */
4638 rc = mdrunner_membed(fplog, cr, NFILE, fnm, oenv, bVerbose, bCompact,
4640 ddxyz, dd_node_order, rdd, rconstr, dddlb_opt[0], dlb_scale,
4641 ddcsx, ddcsy, ddcsz, nstepout, resetstep, nmultisim, repl_ex_nst,
4642 repl_ex_seed, pforce, cpt_period, max_hours, deviceOptions, Flags,
4643 xy_fac,xy_max,z_fac,z_max,
4644 it_xy,it_z,probe_rad,low_up_rm,
4645 pieces,bALLOW_ASYMMETRY,maxwarn);
4647 if (gmx_parallel_env_initialized())
4650 if (MULTIMASTER(cr)) {
4654 /* Log file has to be closed in mdrunner if we are appending to it
4655 (fplog not set here) */
4656 fprintf(stderr,"Please cite:\nWolf et al, J Comp Chem 31 (2010) 2169-2174.\n");
4658 if (MASTER(cr) && !bAppendFiles)
4660 gmx_log_close(fplog);