1 /* -*- mode: c; tab-width: 4; indent-tabs-mode: nil; c-basic-offset: 4; c-file-style: "stroustrup"; -*-
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.
15 * This program is free software; you can redistribute it and/or
16 * modify it under the terms of the GNU General Public License
17 * as published by the Free Software Foundation; either version 2
18 * of the License, or (at your option) any later version.
20 * If you want to redistribute modifications, please consider that
21 * scientific software is very special. Version control is crucial -
22 * bugs must be traceable. We will be happy to consider code for
23 * inclusion in the official distribution, but derived work must not
24 * be called official GROMACS. Details are found in the README & COPYING
25 * files - if they are missing, get the official version at www.gromacs.org.
27 * To help us fund GROMACS development, we humbly ask that you cite
28 * the papers on the package - you can find them in the top README file.
30 * For more info, check our website at http://www.gromacs.org
33 * GROwing Monsters And Cloning Shrimps
56 #include "mtop_util.h"
59 #include "groupcoord.h"
62 /* We use the same defines as in mvdata.c here */
63 #define block_bc(cr, d) gmx_bcast( sizeof(d), &(d),(cr))
64 #define nblock_bc(cr,nr,d) gmx_bcast((nr)*sizeof((d)[0]), (d),(cr))
65 #define snew_bc(cr,d,nr) { if (!MASTER(cr)) snew((d),(nr)); }
68 /* enum to identify the type of ED: none, normal ED, flooding */
69 enum {eEDnone, eEDedsam, eEDflood, eEDnr};
71 /* enum to identify operations on reference, average, origin, target structures */
72 enum {eedREF, eedAV, eedORI, eedTAR, eedNR};
77 int neig; /* nr of eigenvectors */
78 int *ieig; /* index nrs of eigenvectors */
79 real *stpsz; /* stepsizes (per eigenvector) */
80 rvec **vec; /* eigenvector components */
81 real *xproj; /* instantaneous x projections */
82 real *fproj; /* instantaneous f projections */
83 real radius; /* instantaneous radius */
84 real *refproj; /* starting or target projecions */
85 /* When using flooding as harmonic restraint: The current reference projection
86 * is at each step calculated from the initial refproj0 and the slope. */
87 real *refproj0,*refprojslope;
93 t_eigvec mon; /* only monitored, no constraints */
94 t_eigvec linfix; /* fixed linear constraints */
95 t_eigvec linacc; /* acceptance linear constraints */
96 t_eigvec radfix; /* fixed radial constraints (exp) */
97 t_eigvec radacc; /* acceptance radial constraints (exp) */
98 t_eigvec radcon; /* acceptance rad. contraction constr. */
105 gmx_bool bHarmonic; /* Use flooding for harmonic restraint on
107 gmx_bool bConstForce; /* Do not calculate a flooding potential,
108 instead flood with a constant force */
118 rvec *forces_cartesian;
119 t_eigvec vecs; /* use flooding for these */
123 /* This type is for the average, reference, target, and origin structure */
124 typedef struct gmx_edx
126 int nr; /* number of atoms this structure contains */
127 int nr_loc; /* number of atoms on local node */
128 int *anrs; /* atom index numbers */
129 int *anrs_loc; /* local atom index numbers */
130 int nalloc_loc; /* allocation size of anrs_loc */
131 int *c_ind; /* at which position of the whole anrs
132 * array is a local atom?, i.e.
133 * c_ind[0...nr_loc-1] gives the atom index
134 * with respect to the collective
135 * anrs[0...nr-1] array */
136 rvec *x; /* positions for this structure */
137 rvec *x_old; /* used to keep track of the shift vectors
138 such that the ED molecule can always be
139 made whole in the parallel case */
140 real *m; /* masses */
141 real mtot; /* total mass (only used in sref) */
142 real *sqrtm; /* sqrt of the masses used for mass-
143 * weighting of analysis (only used in sav) */
149 int nini; /* total Nr of atoms */
150 gmx_bool fitmas; /* true if trans fit with cm */
151 gmx_bool pcamas; /* true if mass-weighted PCA */
152 int presteps; /* number of steps to run without any
153 * perturbations ... just monitoring */
154 int outfrq; /* freq (in steps) of writing to edo */
155 int maxedsteps; /* max nr of steps per cycle */
157 /* all gmx_edx datasets are copied to all nodes in the parallel case */
158 struct gmx_edx sref; /* reference positions, to these fitting
160 gmx_bool bRefEqAv; /* If true, reference & average indices
161 * are the same. Used for optimization */
162 struct gmx_edx sav; /* average positions */
163 struct gmx_edx star; /* target positions */
164 struct gmx_edx sori; /* origin positions */
166 t_edvecs vecs; /* eigenvectors */
167 real slope; /* minimal slope in acceptance radexp */
169 gmx_bool bNeedDoEdsam; /* if any of the options mon, linfix, ...
170 * is used (i.e. apart from flooding) */
171 t_edflood flood; /* parameters especially for flooding */
172 struct t_ed_buffer *buf; /* handle to local buffers */
173 struct edpar *next_edi; /* Pointer to another ed dataset */
177 typedef struct gmx_edsam
179 int eEDtype; /* Type of ED: see enums above */
180 const char *edinam; /* name of ED sampling input file */
181 const char *edonam; /* output */
182 FILE *edo; /* output file pointer */
185 gmx_bool bStartFromCpt;
193 rvec old_transvec,older_transvec,transvec_compact;
194 rvec *xcoll; /* Positions from all nodes, this is the
195 collective set we work on.
196 These are the positions of atoms with
197 average structure indices */
198 rvec *xc_ref; /* same but with reference structure indices */
199 ivec *shifts_xcoll; /* Shifts for xcoll */
200 ivec *extra_shifts_xcoll; /* xcoll shift changes since last NS step */
201 ivec *shifts_xc_ref; /* Shifts for xc_ref */
202 ivec *extra_shifts_xc_ref; /* xc_ref shift changes since last NS step */
203 gmx_bool bUpdateShifts; /* TRUE in NS steps to indicate that the
204 ED shifts for this ED dataset need to
209 /* definition of ED buffer structure */
212 struct t_fit_to_ref * fit_to_ref;
213 struct t_do_edfit * do_edfit;
214 struct t_do_edsam * do_edsam;
215 struct t_do_radcon * do_radcon;
219 /* Function declarations */
220 static void fit_to_reference(rvec *xcoll,rvec transvec,matrix rotmat,t_edpar *edi);
222 static void translate_and_rotate(rvec *x,int nat,rvec transvec,matrix rotmat);
223 /* End function declarations */
226 /* Does not subtract average positions, projection on single eigenvector is returned
227 * used by: do_linfix, do_linacc, do_radfix, do_radacc, do_radcon
228 * Average position is subtracted in ed_apply_constraints prior to calling projectx
230 static real projectx(t_edpar *edi, rvec *xcoll, rvec *vec)
236 for (i=0; i<edi->sav.nr; i++)
237 proj += edi->sav.sqrtm[i]*iprod(vec[i], xcoll[i]);
243 /* Specialized: projection is stored in vec->refproj
244 * -> used for radacc, radfix, radcon and center of flooding potential
245 * subtracts average positions, projects vector x */
246 static void rad_project(t_edpar *edi, rvec *x, t_eigvec *vec, t_commrec *cr)
251 /* Subtract average positions */
252 for (i = 0; i < edi->sav.nr; i++)
253 rvec_dec(x[i], edi->sav.x[i]);
255 for (i = 0; i < vec->neig; i++)
257 vec->refproj[i] = projectx(edi,x,vec->vec[i]);
258 rad += pow((vec->refproj[i]-vec->xproj[i]),2);
260 vec->radius=sqrt(rad);
262 /* Add average positions */
263 for (i = 0; i < edi->sav.nr; i++)
264 rvec_inc(x[i], edi->sav.x[i]);
268 /* Project vector x, subtract average positions prior to projection and add
269 * them afterwards to retain the unchanged vector. Store in xproj. Mass-weighting
271 static void project_to_eigvectors(rvec *x, /* The positions to project to an eigenvector */
272 t_eigvec *vec, /* The eigenvectors */
278 if (!vec->neig) return;
280 /* Subtract average positions */
281 for (i=0; i<edi->sav.nr; i++)
282 rvec_dec(x[i], edi->sav.x[i]);
284 for (i=0; i<vec->neig; i++)
285 vec->xproj[i] = projectx(edi, x, vec->vec[i]);
287 /* Add average positions */
288 for (i=0; i<edi->sav.nr; i++)
289 rvec_inc(x[i], edi->sav.x[i]);
293 /* Project vector x onto all edi->vecs (mon, linfix,...) */
294 static void project(rvec *x, /* positions to project */
295 t_edpar *edi) /* edi data set */
297 /* It is not more work to subtract the average position in every
298 * subroutine again, because these routines are rarely used simultanely */
299 project_to_eigvectors(x, &edi->vecs.mon , edi);
300 project_to_eigvectors(x, &edi->vecs.linfix, edi);
301 project_to_eigvectors(x, &edi->vecs.linacc, edi);
302 project_to_eigvectors(x, &edi->vecs.radfix, edi);
303 project_to_eigvectors(x, &edi->vecs.radacc, edi);
304 project_to_eigvectors(x, &edi->vecs.radcon, edi);
308 static real calc_radius(t_eigvec *vec)
314 for (i=0; i<vec->neig; i++)
315 rad += pow((vec->refproj[i]-vec->xproj[i]),2);
317 return rad=sqrt(rad);
323 static void dump_xcoll(t_edpar *edi, struct t_do_edsam *buf, t_commrec *cr,
330 ivec *shifts, *eshifts;
337 shifts = buf->shifts_xcoll;
338 eshifts = buf->extra_shifts_xcoll;
340 sprintf(fn, "xcolldump_step%d.txt", step);
343 for (i=0; i<edi->sav.nr; i++)
344 fprintf(fp, "%d %9.5f %9.5f %9.5f %d %d %d %d %d %d\n",
346 xcoll[i][XX] , xcoll[i][YY] , xcoll[i][ZZ],
347 shifts[i][XX] , shifts[i][YY] , shifts[i][ZZ],
348 eshifts[i][XX], eshifts[i][YY], eshifts[i][ZZ]);
355 static void dump_edi_positions(FILE *out, struct gmx_edx *s, const char name[])
360 fprintf(out, "#%s positions:\n%d\n", name, s->nr);
364 fprintf(out, "#index, x, y, z");
366 fprintf(out, ", sqrt(m)");
367 for (i=0; i<s->nr; i++)
369 fprintf(out, "\n%6d %11.6f %11.6f %11.6f",s->anrs[i], s->x[i][XX], s->x[i][YY], s->x[i][ZZ]);
371 fprintf(out,"%9.3f",s->sqrtm[i]);
378 static void dump_edi_eigenvecs(FILE *out, t_eigvec *ev,
379 const char name[], int length)
384 fprintf(out, "#%s eigenvectors:\n%d\n", name, ev->neig);
385 /* Dump the data for every eigenvector: */
386 for (i=0; i<ev->neig; i++)
388 fprintf(out, "EV %4d\ncomponents %d\nstepsize %f\nxproj %f\nfproj %f\nrefproj %f\nradius %f\nComponents:\n",
389 ev->ieig[i], length, ev->stpsz[i], ev->xproj[i], ev->fproj[i], ev->refproj[i], ev->radius);
390 for (j=0; j<length; j++)
391 fprintf(out, "%11.6f %11.6f %11.6f\n", ev->vec[i][j][XX], ev->vec[i][j][YY], ev->vec[i][j][ZZ]);
397 static void dump_edi(t_edpar *edpars, t_commrec *cr, int nr_edi)
403 sprintf(fn, "EDdump_node%d_edi%d", cr->nodeid, nr_edi);
404 out = ffopen(fn, "w");
406 fprintf(out,"#NINI\n %d\n#FITMAS\n %d\n#ANALYSIS_MAS\n %d\n",
407 edpars->nini,edpars->fitmas,edpars->pcamas);
408 fprintf(out,"#OUTFRQ\n %d\n#MAXLEN\n %d\n#SLOPECRIT\n %f\n",
409 edpars->outfrq,edpars->maxedsteps,edpars->slope);
410 fprintf(out,"#PRESTEPS\n %d\n#DELTA_F0\n %f\n#TAU\n %f\n#EFL_NULL\n %f\n#ALPHA2\n %f\n",
411 edpars->presteps,edpars->flood.deltaF0,edpars->flood.tau,
412 edpars->flood.constEfl,edpars->flood.alpha2);
414 /* Dump reference, average, target, origin positions */
415 dump_edi_positions(out, &edpars->sref, "REFERENCE");
416 dump_edi_positions(out, &edpars->sav , "AVERAGE" );
417 dump_edi_positions(out, &edpars->star, "TARGET" );
418 dump_edi_positions(out, &edpars->sori, "ORIGIN" );
420 /* Dump eigenvectors */
421 dump_edi_eigenvecs(out, &edpars->vecs.mon , "MONITORED", edpars->sav.nr);
422 dump_edi_eigenvecs(out, &edpars->vecs.linfix, "LINFIX" , edpars->sav.nr);
423 dump_edi_eigenvecs(out, &edpars->vecs.linacc, "LINACC" , edpars->sav.nr);
424 dump_edi_eigenvecs(out, &edpars->vecs.radfix, "RADFIX" , edpars->sav.nr);
425 dump_edi_eigenvecs(out, &edpars->vecs.radacc, "RADACC" , edpars->sav.nr);
426 dump_edi_eigenvecs(out, &edpars->vecs.radcon, "RADCON" , edpars->sav.nr);
428 /* Dump flooding eigenvectors */
429 dump_edi_eigenvecs(out, &edpars->flood.vecs, "FLOODING" , edpars->sav.nr);
431 /* Dump ed local buffer */
432 fprintf(out, "buf->do_edfit =%p\n", (void*)edpars->buf->do_edfit );
433 fprintf(out, "buf->do_edsam =%p\n", (void*)edpars->buf->do_edsam );
434 fprintf(out, "buf->do_radcon =%p\n", (void*)edpars->buf->do_radcon );
441 static void dump_rotmat(FILE* out,matrix rotmat)
443 fprintf(out,"ROTMAT: %12.8f %12.8f %12.8f\n",rotmat[XX][XX],rotmat[XX][YY],rotmat[XX][ZZ]);
444 fprintf(out,"ROTMAT: %12.8f %12.8f %12.8f\n",rotmat[YY][XX],rotmat[YY][YY],rotmat[YY][ZZ]);
445 fprintf(out,"ROTMAT: %12.8f %12.8f %12.8f\n",rotmat[ZZ][XX],rotmat[ZZ][YY],rotmat[ZZ][ZZ]);
450 static void dump_rvec(FILE *out, int dim, rvec *x)
455 for (i=0; i<dim; i++)
456 fprintf(out,"%4d %f %f %f\n",i,x[i][XX],x[i][YY],x[i][ZZ]);
461 static void dump_mat(FILE* out, int dim, double** mat)
466 fprintf(out,"MATRIX:\n");
470 fprintf(out,"%f ",mat[i][j]);
482 static void do_edfit(int natoms,rvec *xp,rvec *x,matrix R,t_edpar *edi)
484 /* this is a copy of do_fit with some modifications */
491 struct t_do_edfit *loc;
494 if(edi->buf->do_edfit != NULL)
499 snew(edi->buf->do_edfit,1);
501 loc = edi->buf->do_edfit;
505 snew(loc->omega,2*DIM);
507 for(i=0; i<2*DIM; i++)
509 snew(loc->omega[i],2*DIM);
510 snew(loc->om[i],2*DIM);
524 /* calculate the matrix U */
526 for(n=0;(n<natoms);n++)
528 for(c=0; (c<DIM); c++)
531 for(r=0; (r<DIM); r++)
539 /* construct loc->omega */
540 /* loc->omega is symmetric -> loc->omega==loc->omega' */
545 loc->omega[r][c]=u[r-3][c];
546 loc->omega[c][r]=u[r-3][c];
554 /* determine h and k */
558 dump_mat(stderr,2*DIM,loc->omega);
560 fprintf(stderr,"d[%d] = %f\n",i,d[i]);
563 jacobi(loc->omega,6,d,loc->om,&irot);
566 fprintf(stderr,"IROT=0\n");
568 index=0; /* For the compiler only */
582 vh[j][i]=M_SQRT2*loc->om[i][index];
583 vk[j][i]=M_SQRT2*loc->om[i+DIM][index];
590 R[c][r]=vk[0][r]*vh[0][c]+
596 R[c][r]=vk[0][r]*vh[0][c]+
602 static void rmfit(int nat, rvec *xcoll, rvec transvec, matrix rotmat)
609 * The inverse rotation is described by the transposed rotation matrix */
610 transpose(rotmat,tmat);
611 rotate_x(xcoll, nat, tmat);
613 /* Remove translation */
614 vec[XX]=-transvec[XX];
615 vec[YY]=-transvec[YY];
616 vec[ZZ]=-transvec[ZZ];
617 translate_x(xcoll, nat, vec);
621 /**********************************************************************************
622 ******************** FLOODING ****************************************************
623 **********************************************************************************
625 The flooding ability was added later to edsam. Many of the edsam functionality could be reused for that purpose.
626 The flooding covariance matrix, i.e. the selected eigenvectors and their corresponding eigenvalues are
627 read as 7th Component Group. The eigenvalues are coded into the stepsize parameter (as used by -linfix or -linacc).
629 do_md clls right in the beginning the function init_edsam, which reads the edi file, saves all the necessary information in
630 the edi structure and calls init_flood, to initialise some extra fields in the edi->flood structure.
632 since the flooding acts on forces do_flood is called from the function force() (force.c), while the other
633 edsam functionality is hooked into md via the update() (update.c) function acting as constraint on positions.
635 do_flood makes a copy of the positions,
636 fits them, projects them computes flooding_energy, and flooding forces. The forces are computed in the
637 space of the eigenvectors and are then blown up to the full cartesian space and rotated back to remove the
638 fit. Then do_flood adds these forces to the forcefield-forces
639 (given as parameter) and updates the adaptive flooding parameters Efl and deltaF.
641 To center the flooding potential at a different location one can use the -ori option in make_edi. The ori
642 structure is projected to the system of eigenvectors and then this position in the subspace is used as
643 center of the flooding potential. If the option is not used, the center will be zero in the subspace,
644 i.e. the average structure as given in the make_edi file.
646 To use the flooding potential as restraint, make_edi has the option -restrain, which leads to inverted
647 signs of alpha2 and Efl, such that the sign in the exponential of Vfl is not inverted but the sign of
648 Vfl is inverted. Vfl = Efl * exp (- .../Efl/alpha2*x^2...) With tau>0 the negative Efl will grow slowly
649 so that the restraint is switched off slowly. When Efl==0 and inverted flooding is ON is reached no
650 further adaption is applied, Efl will stay constant at zero.
652 To use restraints with harmonic potentials switch -restrain and -harmonic. Then the eigenvalues are
653 used as spring constants for the harmonic potential.
654 Note that eq3 in the flooding paper (J. Comp. Chem. 2006, 27, 1693-1702) defines the parameter lambda \
655 as the inverse of the spring constant, whereas the implementation uses lambda as the spring constant.
657 To use more than one flooding matrix just concatenate several .edi files (cat flood1.edi flood2.edi > flood_all.edi)
658 the routine read_edi_file reads all of theses flooding files.
659 The structure t_edi is now organized as a list of t_edis and the function do_flood cycles through the list
660 calling the do_single_flood() routine for every single entry. Since every state variables have been kept in one
661 edi there is no interdependence whatsoever. The forces are added together.
663 To write energies into the .edr file, call the function
664 get_flood_enx_names(char**, int *nnames) to get the Header (Vfl1 Vfl2... Vfln)
666 get_flood_energies(real Vfl[],int nnames);
669 - one could program the whole thing such that Efl, Vfl and deltaF is written to the .edr file. -- i dont know how to do that, yet.
671 Maybe one should give a range of atoms for which to remove motion, so that motion is removed with
672 two edsam files from two peptide chains
675 static void write_edo_flood(t_edpar *edi, FILE *fp, gmx_large_int_t step)
679 gmx_bool bOutputRef=FALSE;
682 fprintf(fp,"%d.th FL: %s %12.5e %12.5e %12.5e\n",
683 edi->flood.flood_id, gmx_step_str(step,buf),
684 edi->flood.Efl, edi->flood.Vfl, edi->flood.deltaF);
687 /* Check whether any of the references changes with time (this can happen
688 * in case flooding is used as harmonic restraint). If so, output all the
689 * current reference projections. */
690 if (edi->flood.bHarmonic)
692 for (i = 0; i < edi->flood.vecs.neig; i++)
694 if (edi->flood.vecs.refprojslope[i] != 0.0)
699 fprintf(fp, "Ref. projs.: ");
700 for (i = 0; i < edi->flood.vecs.neig; i++)
702 fprintf(fp, "%12.5e ", edi->flood.vecs.refproj[i]);
707 fprintf(fp,"FL_FORCES: ");
709 for (i=0; i<edi->flood.vecs.neig; i++)
710 fprintf(fp," %12.5e",edi->flood.vecs.fproj[i]);
716 /* From flood.xproj compute the Vfl(x) at this point */
717 static real flood_energy(t_edpar *edi, gmx_large_int_t step)
719 /* compute flooding energy Vfl
720 Vfl = Efl * exp( - \frac {kT} {2Efl alpha^2} * sum_i { \lambda_i c_i^2 } )
721 \lambda_i is the reciprocal eigenvalue 1/\sigma_i
722 it is already computed by make_edi and stored in stpsz[i]
724 Vfl = - Efl * 1/2(sum _i {\frac 1{\lambda_i} c_i^2})
731 /* Each time this routine is called (i.e. each time step), we add a small
732 * value to the reference projection. This way a harmonic restraint towards
733 * a moving reference is realized. If no value for the additive constant
734 * is provided in the edi file, the reference will not change. */
735 if (edi->flood.bHarmonic)
737 for (i=0; i<edi->flood.vecs.neig; i++)
739 edi->flood.vecs.refproj[i] = edi->flood.vecs.refproj0[i] + step * edi->flood.vecs.refprojslope[i];
744 /* Compute sum which will be the exponent of the exponential */
745 for (i=0; i<edi->flood.vecs.neig; i++)
747 /* stpsz stores the reciprocal eigenvalue 1/sigma_i */
748 sum += edi->flood.vecs.stpsz[i]*(edi->flood.vecs.xproj[i]-edi->flood.vecs.refproj[i])*(edi->flood.vecs.xproj[i]-edi->flood.vecs.refproj[i]);
751 /* Compute the Gauss function*/
752 if (edi->flood.bHarmonic)
754 Vfl = -0.5*edi->flood.Efl*sum; /* minus sign because Efl is negative, if restrain is on. */
758 Vfl = edi->flood.Efl!=0 ? edi->flood.Efl*exp(-edi->flood.kT/2/edi->flood.Efl/edi->flood.alpha2*sum) :0;
765 /* From the position and from Vfl compute forces in subspace -> store in edi->vec.flood.fproj */
766 static void flood_forces(t_edpar *edi)
768 /* compute the forces in the subspace of the flooding eigenvectors
769 * by the formula F_i= V_{fl}(c) * ( \frac {kT} {E_{fl}} \lambda_i c_i */
772 real energy=edi->flood.Vfl;
775 if (edi->flood.bHarmonic)
776 for (i=0; i<edi->flood.vecs.neig; i++)
778 edi->flood.vecs.fproj[i] = edi->flood.Efl* edi->flood.vecs.stpsz[i]*(edi->flood.vecs.xproj[i]-edi->flood.vecs.refproj[i]);
781 for (i=0; i<edi->flood.vecs.neig; i++)
783 /* if Efl is zero the forces are zero if not use the formula */
784 edi->flood.vecs.fproj[i] = edi->flood.Efl!=0 ? edi->flood.kT/edi->flood.Efl/edi->flood.alpha2*energy*edi->flood.vecs.stpsz[i]*(edi->flood.vecs.xproj[i]-edi->flood.vecs.refproj[i]) : 0;
789 /* Raise forces from subspace into cartesian space */
790 static void flood_blowup(t_edpar *edi, rvec *forces_cart)
792 /* this function lifts the forces from the subspace to the cartesian space
793 all the values not contained in the subspace are assumed to be zero and then
794 a coordinate transformation from eigenvector to cartesian vectors is performed
795 The nonexistent values don't have to be set to zero explicitly, they would occur
796 as zero valued summands, hence we just stop to compute this part of the sum.
798 for every atom we add all the contributions to this atom from all the different eigenvectors.
800 NOTE: one could add directly to the forcefield forces, would mean we wouldn't have to clear the
801 field forces_cart prior the computation, but we compute the forces separately
802 to have them accessible for diagnostics
809 forces_sub = edi->flood.vecs.fproj;
812 /* Calculate the cartesian forces for the local atoms */
814 /* Clear forces first */
815 for (j=0; j<edi->sav.nr_loc; j++)
816 clear_rvec(forces_cart[j]);
818 /* Now compute atomwise */
819 for (j=0; j<edi->sav.nr_loc; j++)
821 /* Compute forces_cart[edi->sav.anrs[j]] */
822 for (eig=0; eig<edi->flood.vecs.neig; eig++)
824 /* Force vector is force * eigenvector (compute only atom j) */
825 svmul(forces_sub[eig],edi->flood.vecs.vec[eig][edi->sav.c_ind[j]],dum);
826 /* Add this vector to the cartesian forces */
827 rvec_inc(forces_cart[j],dum);
833 /* Update the values of Efl, deltaF depending on tau and Vfl */
834 static void update_adaption(t_edpar *edi)
836 /* this function updates the parameter Efl and deltaF according to the rules given in
837 * 'predicting unimolecular chemical reactions: chemical flooding' M Mueller et al,
840 if ((edi->flood.tau < 0 ? -edi->flood.tau : edi->flood.tau ) > 0.00000001)
842 edi->flood.Efl = edi->flood.Efl+edi->flood.dt/edi->flood.tau*(edi->flood.deltaF0-edi->flood.deltaF);
843 /* check if restrain (inverted flooding) -> don't let EFL become positive */
844 if (edi->flood.alpha2<0 && edi->flood.Efl>-0.00000001)
847 edi->flood.deltaF = (1-edi->flood.dt/edi->flood.tau)*edi->flood.deltaF+edi->flood.dt/edi->flood.tau*edi->flood.Vfl;
852 static void do_single_flood(
857 gmx_large_int_t step,
860 gmx_bool bNS) /* Are we in a neighbor searching step? */
863 matrix rotmat; /* rotation matrix */
864 matrix tmat; /* inverse rotation */
865 rvec transvec; /* translation vector */
866 struct t_do_edsam *buf;
869 buf=edi->buf->do_edsam;
872 /* Broadcast the positions of the AVERAGE structure such that they are known on
873 * every processor. Each node contributes its local positions x and stores them in
874 * the collective ED array buf->xcoll */
875 communicate_group_positions(cr, buf->xcoll, buf->shifts_xcoll, buf->extra_shifts_xcoll, bNS, x,
876 edi->sav.nr, edi->sav.nr_loc, edi->sav.anrs_loc, edi->sav.c_ind, edi->sav.x_old, box);
878 /* Only assembly REFERENCE positions if their indices differ from the average ones */
880 communicate_group_positions(cr, buf->xc_ref, buf->shifts_xc_ref, buf->extra_shifts_xc_ref, bNS, x,
881 edi->sref.nr, edi->sref.nr_loc, edi->sref.anrs_loc, edi->sref.c_ind, edi->sref.x_old, box);
883 /* If bUpdateShifts was TRUE, the shifts have just been updated in get_positions.
884 * We do not need to update the shifts until the next NS step */
885 buf->bUpdateShifts = FALSE;
887 /* Now all nodes have all of the ED/flooding positions in edi->sav->xcoll,
888 * as well as the indices in edi->sav.anrs */
890 /* Fit the reference indices to the reference structure */
892 fit_to_reference(buf->xcoll , transvec, rotmat, edi);
894 fit_to_reference(buf->xc_ref, transvec, rotmat, edi);
896 /* Now apply the translation and rotation to the ED structure */
897 translate_and_rotate(buf->xcoll, edi->sav.nr, transvec, rotmat);
899 /* Project fitted structure onto supbspace -> store in edi->flood.vecs.xproj */
900 project_to_eigvectors(buf->xcoll,&edi->flood.vecs,edi);
902 if (FALSE == edi->flood.bConstForce)
904 /* Compute Vfl(x) from flood.xproj */
905 edi->flood.Vfl = flood_energy(edi, step);
907 update_adaption(edi);
909 /* Compute the flooding forces */
913 /* Translate them into cartesian positions */
914 flood_blowup(edi, edi->flood.forces_cartesian);
916 /* Rotate forces back so that they correspond to the given structure and not to the fitted one */
917 /* Each node rotates back its local forces */
918 transpose(rotmat,tmat);
919 rotate_x(edi->flood.forces_cartesian, edi->sav.nr_loc, tmat);
921 /* Finally add forces to the main force variable */
922 for (i=0; i<edi->sav.nr_loc; i++)
923 rvec_inc(force[edi->sav.anrs_loc[i]],edi->flood.forces_cartesian[i]);
925 /* Output is written by the master process */
926 if (do_per_step(step,edi->outfrq) && MASTER(cr))
927 write_edo_flood(edi,edo,step);
931 /* Main flooding routine, called from do_force */
932 extern void do_flood(
933 FILE *log, /* md.log file */
934 t_commrec *cr, /* Communication record */
935 rvec x[], /* Positions on the local processor */
936 rvec force[], /* forcefield forces, to these the flooding forces are added */
937 gmx_edsam_t ed, /* ed data structure contains all ED and flooding datasets */
938 matrix box, /* the box */
939 gmx_large_int_t step, /* The relative time step since ir->init_step is already subtracted */
940 gmx_bool bNS) /* Are we in a neighbor searching step? */
945 if (ed->eEDtype != eEDflood)
951 /* Call flooding for one matrix */
952 if (edi->flood.vecs.neig)
953 do_single_flood(ed->edo,x,force,edi,step,box,cr,bNS);
959 /* Called by init_edi, configure some flooding related variables and structures,
960 * print headers to output files */
961 static void init_flood(t_edpar *edi, gmx_edsam_t ed, real dt, t_commrec *cr)
966 edi->flood.Efl = edi->flood.constEfl;
970 if (edi->flood.vecs.neig)
972 /* If in any of the datasets we find a flooding vector, flooding is turned on */
973 ed->eEDtype = eEDflood;
975 fprintf(stderr,"ED: Flooding of matrix %d is switched on.\n", edi->flood.flood_id);
977 if (edi->flood.bConstForce)
979 /* We have used stpsz as a vehicle to carry the fproj values for constant
980 * force flooding. Now we copy that to flood.vecs.fproj. Note that
981 * in const force flooding, fproj is never changed. */
982 for (i=0; i<edi->flood.vecs.neig; i++)
984 edi->flood.vecs.fproj[i] = edi->flood.vecs.stpsz[i];
986 fprintf(stderr, "ED: applying on eigenvector %d a constant force of %g\n",
987 edi->flood.vecs.ieig[i], edi->flood.vecs.fproj[i]);
990 fprintf(ed->edo,"FL_HEADER: Flooding of matrix %d is switched on! The flooding output will have the following format:\n",
991 edi->flood.flood_id);
992 fprintf(ed->edo,"FL_HEADER: Step Efl Vfl deltaF\n");
998 /*********** Energy book keeping ******/
999 static void get_flood_enx_names(t_edpar *edi, char** names, int *nnames) /* get header of energies */
1008 srenew(names,count);
1009 sprintf(buf,"Vfl_%d",count);
1010 names[count-1]=strdup(buf);
1011 actual=actual->next_edi;
1018 static void get_flood_energies(t_edpar *edi, real Vfl[],int nnames)
1020 /*fl has to be big enough to capture nnames-many entries*/
1029 Vfl[count-1]=actual->flood.Vfl;
1030 actual=actual->next_edi;
1033 if (nnames!=count-1)
1034 gmx_fatal(FARGS,"Number of energies is not consistent with t_edi structure");
1036 /************* END of FLOODING IMPLEMENTATION ****************************/
1040 gmx_edsam_t ed_open(int nfile,const t_filenm fnm[],unsigned long Flags,t_commrec *cr)
1045 /* Allocate space for the ED data structure */
1048 /* We want to perform ED (this switch might later be upgraded to eEDflood) */
1049 ed->eEDtype = eEDedsam;
1053 /* Open .edi input file: */
1054 ed->edinam=ftp2fn(efEDI,nfile,fnm);
1055 /* The master opens the .edo output file */
1056 fprintf(stderr,"ED sampling will be performed!\n");
1057 ed->edonam = ftp2fn(efEDO,nfile,fnm);
1058 ed->edo = gmx_fio_fopen(ed->edonam,(Flags & MD_APPENDFILES)? "a+" : "w+");
1059 ed->bStartFromCpt = Flags & MD_STARTFROMCPT;
1065 /* Broadcasts the structure data */
1066 static void bc_ed_positions(t_commrec *cr, struct gmx_edx *s, int stype)
1068 snew_bc(cr, s->anrs, s->nr ); /* Index numbers */
1069 snew_bc(cr, s->x , s->nr ); /* Positions */
1070 nblock_bc(cr, s->nr, s->anrs );
1071 nblock_bc(cr, s->nr, s->x );
1073 /* For the average & reference structures we need an array for the collective indices,
1074 * and we need to broadcast the masses as well */
1075 if (stype == eedAV || stype == eedREF)
1077 /* We need these additional variables in the parallel case: */
1078 snew(s->c_ind , s->nr ); /* Collective indices */
1079 /* Local atom indices get assigned in dd_make_local_group_indices.
1080 * There, also memory is allocated */
1081 s->nalloc_loc = 0; /* allocation size of s->anrs_loc */
1082 snew_bc(cr, s->x_old, s->nr); /* To be able to always make the ED molecule whole, ... */
1083 nblock_bc(cr, s->nr, s->x_old); /* ... keep track of shift changes with the help of old coords */
1086 /* broadcast masses for the reference structure (for mass-weighted fitting) */
1087 if (stype == eedREF)
1089 snew_bc(cr, s->m, s->nr);
1090 nblock_bc(cr, s->nr, s->m);
1093 /* For the average structure we might need the masses for mass-weighting */
1096 snew_bc(cr, s->sqrtm, s->nr);
1097 nblock_bc(cr, s->nr, s->sqrtm);
1098 snew_bc(cr, s->m, s->nr);
1099 nblock_bc(cr, s->nr, s->m);
1104 /* Broadcasts the eigenvector data */
1105 static void bc_ed_vecs(t_commrec *cr, t_eigvec *ev, int length, gmx_bool bHarmonic)
1109 snew_bc(cr, ev->ieig , ev->neig); /* index numbers of eigenvector */
1110 snew_bc(cr, ev->stpsz , ev->neig); /* stepsizes per eigenvector */
1111 snew_bc(cr, ev->xproj , ev->neig); /* instantaneous x projection */
1112 snew_bc(cr, ev->fproj , ev->neig); /* instantaneous f projection */
1113 snew_bc(cr, ev->refproj, ev->neig); /* starting or target projection */
1115 nblock_bc(cr, ev->neig, ev->ieig );
1116 nblock_bc(cr, ev->neig, ev->stpsz );
1117 nblock_bc(cr, ev->neig, ev->xproj );
1118 nblock_bc(cr, ev->neig, ev->fproj );
1119 nblock_bc(cr, ev->neig, ev->refproj);
1121 snew_bc(cr, ev->vec, ev->neig); /* Eigenvector components */
1122 for (i=0; i<ev->neig; i++)
1124 snew_bc(cr, ev->vec[i], length);
1125 nblock_bc(cr, length, ev->vec[i]);
1128 /* For harmonic restraints the reference projections can change with time */
1131 snew_bc(cr, ev->refproj0 , ev->neig);
1132 snew_bc(cr, ev->refprojslope, ev->neig);
1133 nblock_bc(cr, ev->neig, ev->refproj0 );
1134 nblock_bc(cr, ev->neig, ev->refprojslope);
1139 /* Broadcasts the ED / flooding data to other nodes
1140 * and allocates memory where needed */
1141 static void broadcast_ed_data(t_commrec *cr, gmx_edsam_t ed, int numedis)
1147 /* Master lets the other nodes know if its ED only or also flooding */
1148 gmx_bcast(sizeof(ed->eEDtype), &(ed->eEDtype), cr);
1150 snew_bc(cr, ed->edpar,1);
1151 /* Now transfer the ED data set(s) */
1153 for (nr=0; nr<numedis; nr++)
1155 /* Broadcast a single ED data set */
1158 /* Broadcast positions */
1159 bc_ed_positions(cr, &(edi->sref), eedREF); /* reference positions (don't broadcast masses) */
1160 bc_ed_positions(cr, &(edi->sav ), eedAV ); /* average positions (do broadcast masses as well) */
1161 bc_ed_positions(cr, &(edi->star), eedTAR); /* target positions */
1162 bc_ed_positions(cr, &(edi->sori), eedORI); /* origin positions */
1164 /* Broadcast eigenvectors */
1165 bc_ed_vecs(cr, &edi->vecs.mon , edi->sav.nr, FALSE);
1166 bc_ed_vecs(cr, &edi->vecs.linfix, edi->sav.nr, FALSE);
1167 bc_ed_vecs(cr, &edi->vecs.linacc, edi->sav.nr, FALSE);
1168 bc_ed_vecs(cr, &edi->vecs.radfix, edi->sav.nr, FALSE);
1169 bc_ed_vecs(cr, &edi->vecs.radacc, edi->sav.nr, FALSE);
1170 bc_ed_vecs(cr, &edi->vecs.radcon, edi->sav.nr, FALSE);
1171 /* Broadcast flooding eigenvectors and, if needed, values for the moving reference */
1172 bc_ed_vecs(cr, &edi->flood.vecs, edi->sav.nr, edi->flood.bHarmonic);
1174 /* Set the pointer to the next ED dataset */
1177 snew_bc(cr, edi->next_edi, 1);
1178 edi = edi->next_edi;
1184 /* init-routine called for every *.edi-cycle, initialises t_edpar structure */
1185 static void init_edi(gmx_mtop_t *mtop,t_inputrec *ir,
1186 t_commrec *cr,gmx_edsam_t ed,t_edpar *edi)
1189 real totalmass = 0.0;
1193 /* NOTE Init_edi is executed on the master process only
1194 * The initialized data sets are then transmitted to the
1195 * other nodes in broadcast_ed_data */
1197 edi->bNeedDoEdsam = edi->vecs.mon.neig
1198 || edi->vecs.linfix.neig
1199 || edi->vecs.linacc.neig
1200 || edi->vecs.radfix.neig
1201 || edi->vecs.radacc.neig
1202 || edi->vecs.radcon.neig;
1204 /* evaluate masses (reference structure) */
1205 snew(edi->sref.m, edi->sref.nr);
1206 for (i = 0; i < edi->sref.nr; i++)
1210 gmx_mtop_atomnr_to_atom(mtop,edi->sref.anrs[i],&atom);
1211 edi->sref.m[i] = atom->m;
1215 edi->sref.m[i] = 1.0;
1218 /* Check that every m > 0. Bad things will happen otherwise. */
1219 if (edi->sref.m[i] <= 0.0)
1221 gmx_fatal(FARGS, "Reference structure atom %d (sam.edi index %d) has a mass of %g.\n"
1222 "For a mass-weighted fit, all reference structure atoms need to have a mass >0.\n"
1223 "Either make the covariance analysis non-mass-weighted, or exclude massless\n"
1224 "atoms from the reference structure by creating a proper index group.\n",
1225 i, edi->sref.anrs[i]+1, edi->sref.m[i]);
1228 totalmass += edi->sref.m[i];
1230 edi->sref.mtot = totalmass;
1232 /* Masses m and sqrt(m) for the average structure. Note that m
1233 * is needed if forces have to be evaluated in do_edsam */
1234 snew(edi->sav.sqrtm, edi->sav.nr );
1235 snew(edi->sav.m , edi->sav.nr );
1236 for (i = 0; i < edi->sav.nr; i++)
1238 gmx_mtop_atomnr_to_atom(mtop,edi->sav.anrs[i],&atom);
1239 edi->sav.m[i] = atom->m;
1242 edi->sav.sqrtm[i] = sqrt(atom->m);
1246 edi->sav.sqrtm[i] = 1.0;
1249 /* Check that every m > 0. Bad things will happen otherwise. */
1250 if (edi->sav.sqrtm[i] <= 0.0)
1252 gmx_fatal(FARGS, "Average structure atom %d (sam.edi index %d) has a mass of %g.\n"
1253 "For ED with mass-weighting, all average structure atoms need to have a mass >0.\n"
1254 "Either make the covariance analysis non-mass-weighted, or exclude massless\n"
1255 "atoms from the average structure by creating a proper index group.\n",
1256 i, edi->sav.anrs[i]+1, atom->m);
1260 /* put reference structure in origin */
1261 get_center(edi->sref.x, edi->sref.m, edi->sref.nr, com);
1265 translate_x(edi->sref.x, edi->sref.nr, com);
1267 /* Init ED buffer */
1272 static void check(const char *line, const char *label)
1274 if (!strstr(line,label))
1275 gmx_fatal(FARGS,"Could not find input parameter %s at expected position in edsam input-file (.edi)\nline read instead is %s",label,line);
1279 static int read_checked_edint(FILE *file,const char *label)
1281 char line[STRLEN+1];
1285 fgets2 (line,STRLEN,file);
1287 fgets2 (line,STRLEN,file);
1288 sscanf (line,"%d",&idum);
1293 static int read_edint(FILE *file,gmx_bool *bEOF)
1295 char line[STRLEN+1];
1300 eof=fgets2 (line,STRLEN,file);
1306 eof=fgets2 (line,STRLEN,file);
1312 sscanf (line,"%d",&idum);
1318 static real read_checked_edreal(FILE *file,const char *label)
1320 char line[STRLEN+1];
1324 fgets2 (line,STRLEN,file);
1326 fgets2 (line,STRLEN,file);
1327 sscanf (line,"%lf",&rdum);
1328 return (real) rdum; /* always read as double and convert to single */
1332 static void read_edx(FILE *file,int number,int *anrs,rvec *x)
1335 char line[STRLEN+1];
1339 for(i=0; i<number; i++)
1341 fgets2 (line,STRLEN,file);
1342 sscanf (line,"%d%lf%lf%lf",&anrs[i],&d[0],&d[1],&d[2]);
1343 anrs[i]--; /* we are reading FORTRAN indices */
1345 x[i][j]=d[j]; /* always read as double and convert to single */
1350 static void scan_edvec(FILE *in,int nr,rvec *vec)
1352 char line[STRLEN+1];
1357 for(i=0; (i < nr); i++)
1359 fgets2 (line,STRLEN,in);
1360 sscanf (line,"%le%le%le",&x,&y,&z);
1368 static void read_edvec(FILE *in,int nr,t_eigvec *tvec,gmx_bool bReadRefproj, gmx_bool *bHaveReference)
1371 double rdum,refproj_dum=0.0,refprojslope_dum=0.0;
1372 char line[STRLEN+1];
1375 tvec->neig=read_checked_edint(in,"NUMBER OF EIGENVECTORS");
1378 snew(tvec->ieig ,tvec->neig);
1379 snew(tvec->stpsz ,tvec->neig);
1380 snew(tvec->vec ,tvec->neig);
1381 snew(tvec->xproj ,tvec->neig);
1382 snew(tvec->fproj ,tvec->neig);
1383 snew(tvec->refproj,tvec->neig);
1386 snew(tvec->refproj0 ,tvec->neig);
1387 snew(tvec->refprojslope,tvec->neig);
1390 for(i=0; (i < tvec->neig); i++)
1392 fgets2 (line,STRLEN,in);
1393 if (bReadRefproj) /* ONLY when using flooding as harmonic restraint */
1395 nscan = sscanf(line,"%d%lf%lf%lf",&idum,&rdum,&refproj_dum,&refprojslope_dum);
1396 /* Zero out values which were not scanned */
1400 /* Every 4 values read, including reference position */
1401 *bHaveReference = TRUE;
1404 /* A reference position is provided */
1405 *bHaveReference = TRUE;
1406 /* No value for slope, set to 0 */
1407 refprojslope_dum = 0.0;
1410 /* No values for reference projection and slope, set to 0 */
1412 refprojslope_dum = 0.0;
1415 gmx_fatal(FARGS,"Expected 2 - 4 (not %d) values for flooding vec: <nr> <spring const> <refproj> <refproj-slope>\n", nscan);
1418 tvec->refproj[i]=refproj_dum;
1419 tvec->refproj0[i]=refproj_dum;
1420 tvec->refprojslope[i]=refprojslope_dum;
1422 else /* Normal flooding */
1424 nscan = sscanf(line,"%d%lf",&idum,&rdum);
1426 gmx_fatal(FARGS,"Expected 2 values for flooding vec: <nr> <stpsz>\n");
1429 tvec->stpsz[i]=rdum;
1430 } /* end of loop over eigenvectors */
1432 for(i=0; (i < tvec->neig); i++)
1434 snew(tvec->vec[i],nr);
1435 scan_edvec(in,nr,tvec->vec[i]);
1441 /* calls read_edvec for the vector groups, only for flooding there is an extra call */
1442 static void read_edvecs(FILE *in,int nr,t_edvecs *vecs)
1444 gmx_bool bHaveReference = FALSE;
1447 read_edvec(in, nr, &vecs->mon , FALSE, &bHaveReference);
1448 read_edvec(in, nr, &vecs->linfix, FALSE, &bHaveReference);
1449 read_edvec(in, nr, &vecs->linacc, FALSE, &bHaveReference);
1450 read_edvec(in, nr, &vecs->radfix, FALSE, &bHaveReference);
1451 read_edvec(in, nr, &vecs->radacc, FALSE, &bHaveReference);
1452 read_edvec(in, nr, &vecs->radcon, FALSE, &bHaveReference);
1456 /* Check if the same atom indices are used for reference and average positions */
1457 static gmx_bool check_if_same(struct gmx_edx sref, struct gmx_edx sav)
1462 /* If the number of atoms differs between the two structures,
1463 * they cannot be identical */
1464 if (sref.nr != sav.nr)
1467 /* Now that we know that both stuctures have the same number of atoms,
1468 * check if also the indices are identical */
1469 for (i=0; i < sav.nr; i++)
1471 if (sref.anrs[i] != sav.anrs[i])
1474 fprintf(stderr, "ED: Note: Reference and average structure are composed of the same atom indices.\n");
1480 static int read_edi(FILE* in, gmx_edsam_t ed,t_edpar *edi,int nr_mdatoms, int edi_nr, t_commrec *cr)
1483 const int magic=670;
1486 /* Was a specific reference point for the flooding/umbrella potential provided in the edi file? */
1487 gmx_bool bHaveReference = FALSE;
1490 /* the edi file is not free format, so expect problems if the input is corrupt. */
1492 /* check the magic number */
1493 readmagic=read_edint(in,&bEOF);
1494 /* Check whether we have reached the end of the input file */
1498 if (readmagic != magic)
1500 if (readmagic==666 || readmagic==667 || readmagic==668)
1501 gmx_fatal(FARGS,"Wrong magic number: Use newest version of make_edi to produce edi file");
1502 else if (readmagic != 669)
1503 gmx_fatal(FARGS,"Wrong magic number %d in %s",readmagic,ed->edinam);
1506 /* check the number of atoms */
1507 edi->nini=read_edint(in,&bEOF);
1508 if (edi->nini != nr_mdatoms)
1509 gmx_fatal(FARGS,"Nr of atoms in %s (%d) does not match nr of md atoms (%d)",
1510 ed->edinam,edi->nini,nr_mdatoms);
1512 /* Done checking. For the rest we blindly trust the input */
1513 edi->fitmas = read_checked_edint(in,"FITMAS");
1514 edi->pcamas = read_checked_edint(in,"ANALYSIS_MAS");
1515 edi->outfrq = read_checked_edint(in,"OUTFRQ");
1516 edi->maxedsteps = read_checked_edint(in,"MAXLEN");
1517 edi->slope = read_checked_edreal(in,"SLOPECRIT");
1519 edi->presteps = read_checked_edint(in,"PRESTEPS");
1520 edi->flood.deltaF0 = read_checked_edreal(in,"DELTA_F0");
1521 edi->flood.deltaF = read_checked_edreal(in,"INIT_DELTA_F");
1522 edi->flood.tau = read_checked_edreal(in,"TAU");
1523 edi->flood.constEfl = read_checked_edreal(in,"EFL_NULL");
1524 edi->flood.alpha2 = read_checked_edreal(in,"ALPHA2");
1525 edi->flood.kT = read_checked_edreal(in,"KT");
1526 edi->flood.bHarmonic = read_checked_edint(in,"HARMONIC");
1527 if (readmagic > 669)
1528 edi->flood.bConstForce = read_checked_edint(in,"CONST_FORCE_FLOODING");
1530 edi->flood.bConstForce = FALSE;
1531 edi->flood.flood_id = edi_nr;
1532 edi->sref.nr = read_checked_edint(in,"NREF");
1534 /* allocate space for reference positions and read them */
1535 snew(edi->sref.anrs,edi->sref.nr);
1536 snew(edi->sref.x ,edi->sref.nr);
1537 snew(edi->sref.x_old,edi->sref.nr);
1538 edi->sref.sqrtm =NULL;
1539 read_edx(in,edi->sref.nr,edi->sref.anrs,edi->sref.x);
1541 /* average positions. they define which atoms will be used for ED sampling */
1542 edi->sav.nr=read_checked_edint(in,"NAV");
1543 snew(edi->sav.anrs,edi->sav.nr);
1544 snew(edi->sav.x ,edi->sav.nr);
1545 snew(edi->sav.x_old,edi->sav.nr);
1546 read_edx(in,edi->sav.nr,edi->sav.anrs,edi->sav.x);
1548 /* Check if the same atom indices are used for reference and average positions */
1549 edi->bRefEqAv = check_if_same(edi->sref, edi->sav);
1552 read_edvecs(in,edi->sav.nr,&edi->vecs);
1553 read_edvec(in,edi->sav.nr,&edi->flood.vecs,edi->flood.bHarmonic, &bHaveReference);
1555 /* target positions */
1556 edi->star.nr=read_edint(in,&bEOF);
1557 if (edi->star.nr > 0)
1559 snew(edi->star.anrs,edi->star.nr);
1560 snew(edi->star.x ,edi->star.nr);
1561 edi->star.sqrtm =NULL;
1562 read_edx(in,edi->star.nr,edi->star.anrs,edi->star.x);
1565 /* positions defining origin of expansion circle */
1566 edi->sori.nr=read_edint(in,&bEOF);
1567 if (edi->sori.nr > 0)
1571 /* Both an -ori structure and a at least one manual reference point have been
1572 * specified. That's ambiguous and probably not intentional. */
1573 gmx_fatal(FARGS, "ED: An origin structure has been provided and a at least one (moving) reference\n"
1574 " point was manually specified in the edi file. That is ambiguous. Aborting.\n");
1576 snew(edi->sori.anrs,edi->sori.nr);
1577 snew(edi->sori.x ,edi->sori.nr);
1578 edi->sori.sqrtm =NULL;
1579 read_edx(in,edi->sori.nr,edi->sori.anrs,edi->sori.x);
1588 /* Read in the edi input file. Note that it may contain several ED data sets which were
1589 * achieved by concatenating multiple edi files. The standard case would be a single ED
1590 * data set, though. */
1591 static void read_edi_file(gmx_edsam_t ed, t_edpar *edi, int nr_mdatoms, t_commrec *cr)
1594 t_edpar *curr_edi,*last_edi;
1599 /* This routine is executed on the master only */
1601 /* Open the .edi parameter input file */
1602 in = gmx_fio_fopen(ed->edinam,"r");
1603 fprintf(stderr, "ED: Reading edi file %s\n", ed->edinam);
1605 /* Now read a sequence of ED input parameter sets from the edi file */
1608 while( read_edi(in, ed, curr_edi, nr_mdatoms, edi_nr, cr) )
1611 /* Make shure that the number of atoms in each dataset is the same as in the tpr file */
1612 if (edi->nini != nr_mdatoms)
1613 gmx_fatal(FARGS,"edi file %s (dataset #%d) was made for %d atoms, but the simulation contains %d atoms.",
1614 ed->edinam, edi_nr, edi->nini, nr_mdatoms);
1615 /* Since we arrived within this while loop we know that there is still another data set to be read in */
1616 /* We need to allocate space for the data: */
1618 /* Point the 'next_edi' entry to the next edi: */
1619 curr_edi->next_edi=edi_read;
1620 /* Keep the curr_edi pointer for the case that the next dataset is empty: */
1621 last_edi = curr_edi;
1622 /* Let's prepare to read in the next edi data set: */
1623 curr_edi = edi_read;
1626 gmx_fatal(FARGS, "No complete ED data set found in edi file %s.", ed->edinam);
1628 /* Terminate the edi dataset list with a NULL pointer: */
1629 last_edi->next_edi = NULL;
1631 fprintf(stderr, "ED: Found %d ED dataset%s.\n", edi_nr, edi_nr>1? "s" : "");
1633 /* Close the .edi file again */
1638 struct t_fit_to_ref {
1639 rvec *xcopy; /* Working copy of the positions in fit_to_reference */
1642 /* Fit the current positions to the reference positions
1643 * Do not actually do the fit, just return rotation and translation.
1644 * Note that the COM of the reference structure was already put into
1645 * the origin by init_edi. */
1646 static void fit_to_reference(rvec *xcoll, /* The positions to be fitted */
1647 rvec transvec, /* The translation vector */
1648 matrix rotmat, /* The rotation matrix */
1649 t_edpar *edi) /* Just needed for do_edfit */
1651 rvec com; /* center of mass */
1653 struct t_fit_to_ref *loc;
1656 /* Allocate memory the first time this routine is called for each edi dataset */
1657 if (NULL == edi->buf->fit_to_ref)
1659 snew(edi->buf->fit_to_ref, 1);
1660 snew(edi->buf->fit_to_ref->xcopy, edi->sref.nr);
1662 loc = edi->buf->fit_to_ref;
1664 /* We do not touch the original positions but work on a copy. */
1665 for (i=0; i<edi->sref.nr; i++)
1666 copy_rvec(xcoll[i], loc->xcopy[i]);
1668 /* Calculate the center of mass */
1669 get_center(loc->xcopy, edi->sref.m, edi->sref.nr, com);
1671 transvec[XX] = -com[XX];
1672 transvec[YY] = -com[YY];
1673 transvec[ZZ] = -com[ZZ];
1675 /* Subtract the center of mass from the copy */
1676 translate_x(loc->xcopy, edi->sref.nr, transvec);
1678 /* Determine the rotation matrix */
1679 do_edfit(edi->sref.nr, edi->sref.x, loc->xcopy, rotmat, edi);
1683 static void translate_and_rotate(rvec *x, /* The positions to be translated and rotated */
1684 int nat, /* How many positions are there? */
1685 rvec transvec, /* The translation vector */
1686 matrix rotmat) /* The rotation matrix */
1689 translate_x(x, nat, transvec);
1692 rotate_x(x, nat, rotmat);
1696 /* Gets the rms deviation of the positions to the structure s */
1697 /* fit_to_structure has to be called before calling this routine! */
1698 static real rmsd_from_structure(rvec *x, /* The positions under consideration */
1699 struct gmx_edx *s) /* The structure from which the rmsd shall be computed */
1705 for (i=0; i < s->nr; i++)
1706 rmsd += distance2(s->x[i], x[i]);
1708 rmsd /= (real) s->nr;
1715 void dd_make_local_ed_indices(gmx_domdec_t *dd, struct gmx_edsam *ed)
1720 if (ed->eEDtype != eEDnone)
1722 /* Loop over ED datasets (usually there is just one dataset, though) */
1726 /* Local atoms of the reference structure (for fitting), need only be assembled
1727 * if their indices differ from the average ones */
1729 dd_make_local_group_indices(dd->ga2la, edi->sref.nr, edi->sref.anrs,
1730 &edi->sref.nr_loc, &edi->sref.anrs_loc, &edi->sref.nalloc_loc, edi->sref.c_ind);
1732 /* Local atoms of the average structure (on these ED will be performed) */
1733 dd_make_local_group_indices(dd->ga2la, edi->sav.nr, edi->sav.anrs,
1734 &edi->sav.nr_loc, &edi->sav.anrs_loc, &edi->sav.nalloc_loc, edi->sav.c_ind);
1736 /* Indicate that the ED shift vectors for this structure need to be updated
1737 * at the next call to communicate_group_positions, since obviously we are in a NS step */
1738 edi->buf->do_edsam->bUpdateShifts = TRUE;
1740 /* Set the pointer to the next ED dataset (if any) */
1747 static inline void ed_unshift_single_coord(matrix box, const rvec x, const ivec is, rvec xu)
1758 xu[XX] = x[XX]-tx*box[XX][XX]-ty*box[YY][XX]-tz*box[ZZ][XX];
1759 xu[YY] = x[YY]-ty*box[YY][YY]-tz*box[ZZ][YY];
1760 xu[ZZ] = x[ZZ]-tz*box[ZZ][ZZ];
1763 xu[XX] = x[XX]-tx*box[XX][XX];
1764 xu[YY] = x[YY]-ty*box[YY][YY];
1765 xu[ZZ] = x[ZZ]-tz*box[ZZ][ZZ];
1770 static void do_linfix(rvec *xcoll, t_edpar *edi, int step, t_commrec *cr)
1777 /* loop over linfix vectors */
1778 for (i=0; i<edi->vecs.linfix.neig; i++)
1780 /* calculate the projection */
1781 proj = projectx(edi, xcoll, edi->vecs.linfix.vec[i]);
1783 /* calculate the correction */
1784 add = edi->vecs.linfix.refproj[i] + step*edi->vecs.linfix.stpsz[i] - proj;
1786 /* apply the correction */
1787 add /= edi->sav.sqrtm[i];
1788 for (j=0; j<edi->sav.nr; j++)
1790 svmul(add, edi->vecs.linfix.vec[i][j], vec_dum);
1791 rvec_inc(xcoll[j], vec_dum);
1797 static void do_linacc(rvec *xcoll, t_edpar *edi, t_commrec *cr)
1804 /* loop over linacc vectors */
1805 for (i=0; i<edi->vecs.linacc.neig; i++)
1807 /* calculate the projection */
1808 proj=projectx(edi, xcoll, edi->vecs.linacc.vec[i]);
1810 /* calculate the correction */
1812 if (edi->vecs.linacc.stpsz[i] > 0.0)
1814 if ((proj-edi->vecs.linacc.refproj[i]) < 0.0)
1815 add = edi->vecs.linacc.refproj[i] - proj;
1817 if (edi->vecs.linacc.stpsz[i] < 0.0)
1819 if ((proj-edi->vecs.linacc.refproj[i]) > 0.0)
1820 add = edi->vecs.linacc.refproj[i] - proj;
1823 /* apply the correction */
1824 add /= edi->sav.sqrtm[i];
1825 for (j=0; j<edi->sav.nr; j++)
1827 svmul(add, edi->vecs.linacc.vec[i][j], vec_dum);
1828 rvec_inc(xcoll[j], vec_dum);
1831 /* new positions will act as reference */
1832 edi->vecs.linacc.refproj[i] = proj + add;
1837 static void do_radfix(rvec *xcoll, t_edpar *edi, int step, t_commrec *cr)
1840 real *proj, rad=0.0, ratio;
1844 if (edi->vecs.radfix.neig == 0)
1847 snew(proj, edi->vecs.radfix.neig);
1849 /* loop over radfix vectors */
1850 for (i=0; i<edi->vecs.radfix.neig; i++)
1852 /* calculate the projections, radius */
1853 proj[i] = projectx(edi, xcoll, edi->vecs.radfix.vec[i]);
1854 rad += pow(proj[i] - edi->vecs.radfix.refproj[i], 2);
1858 ratio = (edi->vecs.radfix.stpsz[0]+edi->vecs.radfix.radius)/rad - 1.0;
1859 edi->vecs.radfix.radius += edi->vecs.radfix.stpsz[0];
1861 /* loop over radfix vectors */
1862 for (i=0; i<edi->vecs.radfix.neig; i++)
1864 proj[i] -= edi->vecs.radfix.refproj[i];
1866 /* apply the correction */
1867 proj[i] /= edi->sav.sqrtm[i];
1869 for (j=0; j<edi->sav.nr; j++) {
1870 svmul(proj[i], edi->vecs.radfix.vec[i][j], vec_dum);
1871 rvec_inc(xcoll[j], vec_dum);
1879 static void do_radacc(rvec *xcoll, t_edpar *edi, t_commrec *cr)
1882 real *proj, rad=0.0, ratio=0.0;
1886 if (edi->vecs.radacc.neig == 0)
1889 snew(proj,edi->vecs.radacc.neig);
1891 /* loop over radacc vectors */
1892 for (i=0; i<edi->vecs.radacc.neig; i++)
1894 /* calculate the projections, radius */
1895 proj[i] = projectx(edi, xcoll, edi->vecs.radacc.vec[i]);
1896 rad += pow(proj[i] - edi->vecs.radacc.refproj[i], 2);
1900 /* only correct when radius decreased */
1901 if (rad < edi->vecs.radacc.radius)
1903 ratio = edi->vecs.radacc.radius/rad - 1.0;
1904 rad = edi->vecs.radacc.radius;
1907 edi->vecs.radacc.radius = rad;
1909 /* loop over radacc vectors */
1910 for (i=0; i<edi->vecs.radacc.neig; i++)
1912 proj[i] -= edi->vecs.radacc.refproj[i];
1914 /* apply the correction */
1915 proj[i] /= edi->sav.sqrtm[i];
1917 for (j=0; j<edi->sav.nr; j++)
1919 svmul(proj[i], edi->vecs.radacc.vec[i][j], vec_dum);
1920 rvec_inc(xcoll[j], vec_dum);
1927 struct t_do_radcon {
1931 static void do_radcon(rvec *xcoll, t_edpar *edi, t_commrec *cr)
1934 real rad=0.0, ratio=0.0;
1935 struct t_do_radcon *loc;
1940 if(edi->buf->do_radcon != NULL)
1943 loc = edi->buf->do_radcon;
1948 snew(edi->buf->do_radcon, 1);
1950 loc = edi->buf->do_radcon;
1952 if (edi->vecs.radcon.neig == 0)
1956 snew(loc->proj, edi->vecs.radcon.neig);
1958 /* loop over radcon vectors */
1959 for (i=0; i<edi->vecs.radcon.neig; i++)
1961 /* calculate the projections, radius */
1962 loc->proj[i] = projectx(edi, xcoll, edi->vecs.radcon.vec[i]);
1963 rad += pow(loc->proj[i] - edi->vecs.radcon.refproj[i], 2);
1966 /* only correct when radius increased */
1967 if (rad > edi->vecs.radcon.radius)
1969 ratio = edi->vecs.radcon.radius/rad - 1.0;
1971 /* loop over radcon vectors */
1972 for (i=0; i<edi->vecs.radcon.neig; i++)
1974 /* apply the correction */
1975 loc->proj[i] -= edi->vecs.radcon.refproj[i];
1976 loc->proj[i] /= edi->sav.sqrtm[i];
1977 loc->proj[i] *= ratio;
1979 for (j=0; j<edi->sav.nr; j++)
1981 svmul(loc->proj[i], edi->vecs.radcon.vec[i][j], vec_dum);
1982 rvec_inc(xcoll[j], vec_dum);
1987 edi->vecs.radcon.radius = rad;
1989 if (rad != edi->vecs.radcon.radius)
1992 for (i=0; i<edi->vecs.radcon.neig; i++)
1994 /* calculate the projections, radius */
1995 loc->proj[i] = projectx(edi, xcoll, edi->vecs.radcon.vec[i]);
1996 rad += pow(loc->proj[i] - edi->vecs.radcon.refproj[i], 2);
2003 static void ed_apply_constraints(rvec *xcoll, t_edpar *edi, gmx_large_int_t step, t_commrec *cr)
2008 /* subtract the average positions */
2009 for (i=0; i<edi->sav.nr; i++)
2010 rvec_dec(xcoll[i], edi->sav.x[i]);
2012 /* apply the constraints */
2014 do_linfix(xcoll, edi, step, cr);
2015 do_linacc(xcoll, edi, cr);
2017 do_radfix(xcoll, edi, step, cr);
2018 do_radacc(xcoll, edi, cr);
2019 do_radcon(xcoll, edi, cr);
2021 /* add back the average positions */
2022 for (i=0; i<edi->sav.nr; i++)
2023 rvec_inc(xcoll[i], edi->sav.x[i]);
2027 /* Write out the projections onto the eigenvectors */
2028 static void write_edo(int nr_edi, t_edpar *edi, gmx_edsam_t ed, gmx_large_int_t step,real rmsd)
2034 if (edi->bNeedDoEdsam)
2037 fprintf(ed->edo, "Initial projections:\n");
2040 fprintf(ed->edo,"Step %s, ED #%d ", gmx_step_str(step, buf), nr_edi);
2041 fprintf(ed->edo," RMSD %f nm\n",rmsd);
2044 if (edi->vecs.mon.neig)
2046 fprintf(ed->edo," Monitor eigenvectors");
2047 for (i=0; i<edi->vecs.mon.neig; i++)
2048 fprintf(ed->edo," %d: %12.5e ",edi->vecs.mon.ieig[i],edi->vecs.mon.xproj[i]);
2049 fprintf(ed->edo,"\n");
2051 if (edi->vecs.linfix.neig)
2053 fprintf(ed->edo," Linfix eigenvectors");
2054 for (i=0; i<edi->vecs.linfix.neig; i++)
2055 fprintf(ed->edo," %d: %12.5e ",edi->vecs.linfix.ieig[i],edi->vecs.linfix.xproj[i]);
2056 fprintf(ed->edo,"\n");
2058 if (edi->vecs.linacc.neig)
2060 fprintf(ed->edo," Linacc eigenvectors");
2061 for (i=0; i<edi->vecs.linacc.neig; i++)
2062 fprintf(ed->edo," %d: %12.5e ",edi->vecs.linacc.ieig[i],edi->vecs.linacc.xproj[i]);
2063 fprintf(ed->edo,"\n");
2065 if (edi->vecs.radfix.neig)
2067 fprintf(ed->edo," Radfix eigenvectors");
2068 for (i=0; i<edi->vecs.radfix.neig; i++)
2069 fprintf(ed->edo," %d: %12.5e ",edi->vecs.radfix.ieig[i],edi->vecs.radfix.xproj[i]);
2070 fprintf(ed->edo,"\n");
2071 fprintf(ed->edo," fixed increment radius = %f\n", calc_radius(&edi->vecs.radfix));
2073 if (edi->vecs.radacc.neig)
2075 fprintf(ed->edo," Radacc eigenvectors");
2076 for (i=0; i<edi->vecs.radacc.neig; i++)
2077 fprintf(ed->edo," %d: %12.5e ",edi->vecs.radacc.ieig[i],edi->vecs.radacc.xproj[i]);
2078 fprintf(ed->edo,"\n");
2079 fprintf(ed->edo," acceptance radius = %f\n", calc_radius(&edi->vecs.radacc));
2081 if (edi->vecs.radcon.neig)
2083 fprintf(ed->edo," Radcon eigenvectors");
2084 for (i=0; i<edi->vecs.radcon.neig; i++)
2085 fprintf(ed->edo," %d: %12.5e ",edi->vecs.radcon.ieig[i],edi->vecs.radcon.xproj[i]);
2086 fprintf(ed->edo,"\n");
2087 fprintf(ed->edo," contracting radius = %f\n", calc_radius(&edi->vecs.radcon));
2092 /* Returns if any constraints are switched on */
2093 static int ed_constraints(gmx_bool edtype, t_edpar *edi)
2095 if (edtype == eEDedsam || edtype == eEDflood)
2097 return (edi->vecs.linfix.neig || edi->vecs.linacc.neig ||
2098 edi->vecs.radfix.neig || edi->vecs.radacc.neig ||
2099 edi->vecs.radcon.neig);
2105 /* Copies reference projection 'refproj' to fixed 'refproj0' variable for flooding/
2106 * umbrella sampling simulations. */
2107 static void copyEvecReference(t_eigvec* floodvecs)
2112 if (NULL==floodvecs->refproj0)
2113 snew(floodvecs->refproj0, floodvecs->neig);
2115 for (i=0; i<floodvecs->neig; i++)
2117 floodvecs->refproj0[i] = floodvecs->refproj[i];
2122 void init_edsam(gmx_mtop_t *mtop, /* global topology */
2123 t_inputrec *ir, /* input record */
2124 t_commrec *cr, /* communication record */
2125 gmx_edsam_t ed, /* contains all ED data */
2126 rvec x[], /* positions of the whole MD system */
2127 matrix box) /* the box */
2129 t_edpar *edi = NULL; /* points to a single edi data set */
2130 int numedis=0; /* keep track of the number of ED data sets in edi file */
2131 int i,nr_edi,avindex;
2132 rvec *x_pbc = NULL; /* positions of the whole MD system with pbc removed */
2133 rvec *xfit = NULL; /* the positions which will be fitted to the reference structure */
2134 rvec *xstart = NULL; /* the positions which are subject to ED sampling */
2135 rvec fit_transvec; /* translation ... */
2136 matrix fit_rotmat; /* ... and rotation from fit to reference structure */
2139 if (!DOMAINDECOMP(cr) && PAR(cr) && MASTER(cr))
2140 gmx_fatal(FARGS, "Please switch on domain decomposition to use essential dynamics in parallel.");
2143 fprintf(stderr, "ED: Initializing essential dynamics constraints.\n");
2145 /* Needed for initializing radacc radius in do_edsam */
2148 /* The input file is read by the master and the edi structures are
2149 * initialized here. Input is stored in ed->edpar. Then the edi
2150 * structures are transferred to the other nodes */
2154 /* Read the whole edi file at once: */
2155 read_edi_file(ed,ed->edpar,mtop->natoms,cr);
2157 /* Initialization for every ED/flooding dataset. Flooding uses one edi dataset per
2158 * flooding vector, Essential dynamics can be applied to more than one structure
2159 * as well, but will be done in the order given in the edi file, so
2160 * expect different results for different order of edi file concatenation! */
2164 init_edi(mtop,ir,cr,ed,edi);
2166 /* Init flooding parameters if needed */
2167 init_flood(edi,ed,ir->delta_t,cr);
2174 /* The master does the work here. The other nodes get the positions
2175 * not before dd_partition_system which is called after init_edsam */
2178 /* Remove pbc, make molecule whole.
2179 * When ir->bContinuation=TRUE this has already been done, but ok.
2181 snew(x_pbc,mtop->natoms);
2182 m_rveccopy(mtop->natoms,x,x_pbc);
2183 do_pbc_first_mtop(NULL,ir->ePBC,box,mtop,x_pbc);
2185 /* Reset pointer to first ED data set which contains the actual ED data */
2188 /* Loop over all ED/flooding data sets (usually only one, though) */
2189 for (nr_edi = 1; nr_edi <= numedis; nr_edi++)
2191 /* We use srenew to allocate memory since the size of the buffers
2192 * is likely to change with every ED dataset */
2193 srenew(xfit , edi->sref.nr );
2194 srenew(xstart, edi->sav.nr );
2196 /* Extract the positions of the atoms to which will be fitted */
2197 for (i=0; i < edi->sref.nr; i++)
2199 copy_rvec(x_pbc[edi->sref.anrs[i]], xfit[i]);
2201 /* Save the sref positions such that in the next time step we can make the ED group whole
2202 * in case any of the atoms do not have the correct PBC representation */
2203 copy_rvec(xfit[i], edi->sref.x_old[i]);
2206 /* Extract the positions of the atoms subject to ED sampling */
2207 for (i=0; i < edi->sav.nr; i++)
2209 copy_rvec(x_pbc[edi->sav.anrs[i]], xstart[i]);
2211 /* Save the sav positions such that in the next time step we can make the ED group whole
2212 * in case any of the atoms do not have the correct PBC representation */
2213 copy_rvec(xstart[i], edi->sav.x_old[i]);
2216 /* Make the fit to the REFERENCE structure, get translation and rotation */
2217 fit_to_reference(xfit, fit_transvec, fit_rotmat, edi);
2219 /* Output how well we fit to the reference at the start */
2220 translate_and_rotate(xfit, edi->sref.nr, fit_transvec, fit_rotmat);
2221 fprintf(stderr, "ED: Initial RMSD from reference after fit = %f nm (dataset #%d)\n",
2222 rmsd_from_structure(xfit, &edi->sref), nr_edi);
2224 /* Now apply the translation and rotation to the atoms on which ED sampling will be performed */
2225 translate_and_rotate(xstart, edi->sav.nr, fit_transvec, fit_rotmat);
2227 /* calculate initial projections */
2228 project(xstart, edi);
2230 /* For the target and origin structure both a reference (fit) and an
2231 * average structure can be provided in make_edi. If both structures
2232 * are the same, make_edi only stores one of them in the .edi file.
2233 * If they differ, first the fit and then the average structure is stored
2234 * in star (or sor), thus the number of entries in star/sor is
2235 * (n_fit + n_av) with n_fit the size of the fitting group and n_av
2236 * the size of the average group. */
2238 /* process target structure, if required */
2239 if (edi->star.nr > 0)
2241 fprintf(stderr, "ED: Fitting target structure to reference structure\n");
2243 /* get translation & rotation for fit of target structure to reference structure */
2244 fit_to_reference(edi->star.x, fit_transvec, fit_rotmat, edi);
2246 translate_and_rotate(edi->star.x, edi->star.nr, fit_transvec, fit_rotmat);
2247 if (edi->star.nr == edi->sav.nr)
2251 else /* edi->star.nr = edi->sref.nr + edi->sav.nr */
2253 /* The last sav.nr indices of the target structure correspond to
2254 * the average structure, which must be projected */
2255 avindex = edi->star.nr - edi->sav.nr;
2257 rad_project(edi, &edi->star.x[avindex], &edi->vecs.radcon, cr);
2259 rad_project(edi, xstart, &edi->vecs.radcon, cr);
2261 /* process structure that will serve as origin of expansion circle */
2262 if ( (eEDflood == ed->eEDtype) && (FALSE == edi->flood.bConstForce) )
2263 fprintf(stderr, "ED: Setting center of flooding potential (0 = average structure)\n");
2265 if (edi->sori.nr > 0)
2267 fprintf(stderr, "ED: Fitting origin structure to reference structure\n");
2269 /* fit this structure to reference structure */
2270 fit_to_reference(edi->sori.x, fit_transvec, fit_rotmat, edi);
2272 translate_and_rotate(edi->sori.x, edi->sori.nr, fit_transvec, fit_rotmat);
2273 if (edi->sori.nr == edi->sav.nr)
2277 else /* edi->sori.nr = edi->sref.nr + edi->sav.nr */
2279 /* For the projection, we need the last sav.nr indices of sori */
2280 avindex = edi->sori.nr - edi->sav.nr;
2283 rad_project(edi, &edi->sori.x[avindex], &edi->vecs.radacc, cr);
2284 rad_project(edi, &edi->sori.x[avindex], &edi->vecs.radfix, cr);
2285 if ( (eEDflood == ed->eEDtype) && (FALSE == edi->flood.bConstForce) )
2287 fprintf(stderr, "ED: The ORIGIN structure will define the flooding potential center.\n");
2288 /* Set center of flooding potential to the ORIGIN structure */
2289 rad_project(edi, &edi->sori.x[avindex], &edi->flood.vecs, cr);
2290 /* We already know that no (moving) reference position was provided,
2291 * therefore we can overwrite refproj[0]*/
2292 copyEvecReference(&edi->flood.vecs);
2295 else /* No origin structure given */
2297 rad_project(edi, xstart, &edi->vecs.radacc, cr);
2298 rad_project(edi, xstart, &edi->vecs.radfix, cr);
2299 if ( (eEDflood == ed->eEDtype) && (FALSE == edi->flood.bConstForce) )
2301 if (edi->flood.bHarmonic)
2303 fprintf(stderr, "ED: A (possibly changing) ref. projection will define the flooding potential center.\n");
2304 for (i=0; i<edi->flood.vecs.neig; i++)
2305 edi->flood.vecs.refproj[i] = edi->flood.vecs.refproj0[i];
2309 fprintf(stderr, "ED: The AVERAGE structure will define the flooding potential center.\n");
2310 /* Set center of flooding potential to the center of the covariance matrix,
2311 * i.e. the average structure, i.e. zero in the projected system */
2312 for (i=0; i<edi->flood.vecs.neig; i++)
2313 edi->flood.vecs.refproj[i] = 0.0;
2317 /* For convenience, output the center of the flooding potential for the eigenvectors */
2318 if ( (eEDflood == ed->eEDtype) && (FALSE == edi->flood.bConstForce) )
2320 for (i=0; i<edi->flood.vecs.neig; i++)
2322 fprintf(stdout, "ED: EV %d flooding potential center: %11.4e", i, edi->flood.vecs.refproj[i]);
2323 if (edi->flood.bHarmonic)
2324 fprintf(stdout, " (adding %11.4e/timestep)", edi->flood.vecs.refprojslope[i]);
2325 fprintf(stdout, "\n");
2329 /* set starting projections for linsam */
2330 rad_project(edi, xstart, &edi->vecs.linacc, cr);
2331 rad_project(edi, xstart, &edi->vecs.linfix, cr);
2333 /* Output to file, set the step to -1 so that write_edo knows it was called from init_edsam */
2334 if (ed->edo && !(ed->bStartFromCpt))
2335 write_edo(nr_edi, edi, ed, -1, 0);
2337 /* Prepare for the next edi data set: */
2340 /* Cleaning up on the master node: */
2345 } /* end of MASTER only section */
2349 /* First let everybody know how many ED data sets to expect */
2350 gmx_bcast(sizeof(numedis), &numedis, cr);
2351 /* Broadcast the essential dynamics / flooding data to all nodes */
2352 broadcast_ed_data(cr, ed, numedis);
2356 /* In the single-CPU case, point the local atom numbers pointers to the global
2357 * one, so that we can use the same notation in serial and parallel case: */
2359 /* Loop over all ED data sets (usually only one, though) */
2361 for (nr_edi = 1; nr_edi <= numedis; nr_edi++)
2363 edi->sref.anrs_loc = edi->sref.anrs;
2364 edi->sav.anrs_loc = edi->sav.anrs;
2365 edi->star.anrs_loc = edi->star.anrs;
2366 edi->sori.anrs_loc = edi->sori.anrs;
2367 /* For the same reason as above, make a dummy c_ind array: */
2368 snew(edi->sav.c_ind, edi->sav.nr);
2369 /* Initialize the array */
2370 for (i=0; i<edi->sav.nr; i++)
2371 edi->sav.c_ind[i] = i;
2372 /* In the general case we will need a different-sized array for the reference indices: */
2375 snew(edi->sref.c_ind, edi->sref.nr);
2376 for (i=0; i<edi->sref.nr; i++)
2377 edi->sref.c_ind[i] = i;
2379 /* Point to the very same array in case of other structures: */
2380 edi->star.c_ind = edi->sav.c_ind;
2381 edi->sori.c_ind = edi->sav.c_ind;
2382 /* In the serial case, the local number of atoms is the global one: */
2383 edi->sref.nr_loc = edi->sref.nr;
2384 edi->sav.nr_loc = edi->sav.nr;
2385 edi->star.nr_loc = edi->star.nr;
2386 edi->sori.nr_loc = edi->sori.nr;
2388 /* An on we go to the next edi dataset */
2393 /* Allocate space for ED buffer variables */
2394 /* Again, loop over ED data sets */
2396 for (nr_edi = 1; nr_edi <= numedis; nr_edi++)
2398 /* Allocate space for ED buffer */
2400 snew(edi->buf->do_edsam, 1);
2402 /* Space for collective ED buffer variables */
2404 /* Collective positions of atoms with the average indices */
2405 snew(edi->buf->do_edsam->xcoll , edi->sav.nr);
2406 snew(edi->buf->do_edsam->shifts_xcoll , edi->sav.nr); /* buffer for xcoll shifts */
2407 snew(edi->buf->do_edsam->extra_shifts_xcoll , edi->sav.nr);
2408 /* Collective positions of atoms with the reference indices */
2411 snew(edi->buf->do_edsam->xc_ref , edi->sref.nr);
2412 snew(edi->buf->do_edsam->shifts_xc_ref , edi->sref.nr); /* To store the shifts in */
2413 snew(edi->buf->do_edsam->extra_shifts_xc_ref, edi->sref.nr);
2416 /* Get memory for flooding forces */
2417 snew(edi->flood.forces_cartesian , edi->sav.nr);
2420 /* Dump it all into one file per process */
2421 dump_edi(edi, cr, nr_edi);
2424 /* An on we go to the next edi dataset */
2428 /* Flush the edo file so that the user can check some things
2429 * when the simulation has started */
2435 void do_edsam(t_inputrec *ir,
2436 gmx_large_int_t step,
2439 rvec xs[], /* The local current positions on this processor */
2440 rvec v[], /* The velocities */
2444 int i,edinr,iupdate=500;
2445 matrix rotmat; /* rotation matrix */
2446 rvec transvec; /* translation vector */
2447 rvec dv,dx,x_unsh; /* tmp vectors for velocity, distance, unshifted x coordinate */
2448 real dt_1; /* 1/dt */
2449 struct t_do_edsam *buf;
2451 real rmsdev=-1; /* RMSD from reference structure prior to applying the constraints */
2452 gmx_bool bSuppress=FALSE; /* Write .edo file on master? */
2455 /* Check if ED sampling has to be performed */
2456 if ( ed->eEDtype==eEDnone )
2459 /* Suppress output on first call of do_edsam if
2460 * two-step sd2 integrator is used */
2461 if ( (ir->eI==eiSD2) && (v != NULL) )
2464 dt_1 = 1.0/ir->delta_t;
2466 /* Loop over all ED datasets (usually one) */
2472 if (edi->bNeedDoEdsam)
2475 buf=edi->buf->do_edsam;
2478 /* initialise radacc radius for slope criterion */
2479 buf->oldrad=calc_radius(&edi->vecs.radacc);
2481 /* Copy the positions into buf->xc* arrays and after ED
2482 * feed back corrections to the official positions */
2484 /* Broadcast the ED positions such that every node has all of them
2485 * Every node contributes its local positions xs and stores it in
2486 * the collective buf->xcoll array. Note that for edinr > 1
2487 * xs could already have been modified by an earlier ED */
2489 communicate_group_positions(cr, buf->xcoll, buf->shifts_xcoll, buf->extra_shifts_xcoll, PAR(cr) ? buf->bUpdateShifts : TRUE, xs,
2490 edi->sav.nr, edi->sav.nr_loc, edi->sav.anrs_loc, edi->sav.c_ind, edi->sav.x_old, box);
2493 dump_xcoll(edi, buf, cr, step);
2495 /* Only assembly reference positions if their indices differ from the average ones */
2497 communicate_group_positions(cr, buf->xc_ref, buf->shifts_xc_ref, buf->extra_shifts_xc_ref, PAR(cr) ? buf->bUpdateShifts : TRUE, xs,
2498 edi->sref.nr, edi->sref.nr_loc, edi->sref.anrs_loc, edi->sref.c_ind, edi->sref.x_old, box);
2500 /* If bUpdateShifts was TRUE then the shifts have just been updated in communicate_group_positions.
2501 * We do not need to update the shifts until the next NS step. Note that dd_make_local_ed_indices
2502 * set bUpdateShifts=TRUE in the parallel case. */
2503 buf->bUpdateShifts = FALSE;
2505 /* Now all nodes have all of the ED positions in edi->sav->xcoll,
2506 * as well as the indices in edi->sav.anrs */
2508 /* Fit the reference indices to the reference structure */
2510 fit_to_reference(buf->xcoll , transvec, rotmat, edi);
2512 fit_to_reference(buf->xc_ref, transvec, rotmat, edi);
2514 /* Now apply the translation and rotation to the ED structure */
2515 translate_and_rotate(buf->xcoll, edi->sav.nr, transvec, rotmat);
2517 /* Find out how well we fit to the reference (just for output steps) */
2518 if (do_per_step(step,edi->outfrq) && MASTER(cr))
2522 /* Indices of reference and average structures are identical,
2523 * thus we can calculate the rmsd to SREF using xcoll */
2524 rmsdev = rmsd_from_structure(buf->xcoll,&edi->sref);
2528 /* We have to translate & rotate the reference atoms first */
2529 translate_and_rotate(buf->xc_ref, edi->sref.nr, transvec, rotmat);
2530 rmsdev = rmsd_from_structure(buf->xc_ref,&edi->sref);
2534 /* update radsam references, when required */
2535 if (do_per_step(step,edi->maxedsteps) && step >= edi->presteps)
2537 project(buf->xcoll, edi);
2538 rad_project(edi, buf->xcoll, &edi->vecs.radacc, cr);
2539 rad_project(edi, buf->xcoll, &edi->vecs.radfix, cr);
2543 /* update radacc references, when required */
2544 if (do_per_step(step,iupdate) && step >= edi->presteps)
2546 edi->vecs.radacc.radius = calc_radius(&edi->vecs.radacc);
2547 if (edi->vecs.radacc.radius - buf->oldrad < edi->slope)
2549 project(buf->xcoll, edi);
2550 rad_project(edi, buf->xcoll, &edi->vecs.radacc, cr);
2553 buf->oldrad = edi->vecs.radacc.radius;
2556 /* apply the constraints */
2557 if (step >= edi->presteps && ed_constraints(ed->eEDtype, edi))
2559 /* ED constraints should be applied already in the first MD step
2560 * (which is step 0), therefore we pass step+1 to the routine */
2561 ed_apply_constraints(buf->xcoll, edi, step+1 - ir->init_step, cr);
2564 /* write to edo, when required */
2565 if (do_per_step(step,edi->outfrq))
2567 project(buf->xcoll, edi);
2568 if (MASTER(cr) && !bSuppress)
2569 write_edo(edinr, edi, ed, step, rmsdev);
2572 /* Copy back the positions unless monitoring only */
2573 if (ed_constraints(ed->eEDtype, edi))
2575 /* remove fitting */
2576 rmfit(edi->sav.nr, buf->xcoll, transvec, rotmat);
2578 /* Copy the ED corrected positions into the coordinate array */
2579 /* Each node copies its local part. In the serial case, nat_loc is the
2580 * total number of ED atoms */
2581 for (i=0; i<edi->sav.nr_loc; i++)
2583 /* Unshift local ED coordinate and store in x_unsh */
2584 ed_unshift_single_coord(box, buf->xcoll[edi->sav.c_ind[i]],
2585 buf->shifts_xcoll[edi->sav.c_ind[i]], x_unsh);
2587 /* dx is the ED correction to the positions: */
2588 rvec_sub(x_unsh, xs[edi->sav.anrs_loc[i]], dx);
2592 /* dv is the ED correction to the velocity: */
2593 svmul(dt_1, dx, dv);
2594 /* apply the velocity correction: */
2595 rvec_inc(v[edi->sav.anrs_loc[i]], dv);
2597 /* Finally apply the position correction due to ED: */
2598 copy_rvec(x_unsh, xs[edi->sav.anrs_loc[i]]);
2601 } /* END of if (edi->bNeedDoEdsam) */
2603 /* Prepare for the next ED dataset */
2604 edi = edi->next_edi;
2606 } /* END of loop over ED datasets */