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.
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 * GROwing Monsters And Cloning Shrimps
44 #include "types/commrec.h"
53 #include "gromacs/fileio/confio.h"
55 #include "gmx_random.h"
56 #include "gromacs/fileio/futil.h"
68 #include "gmx_wallcycle.h"
69 #include "gmx_omp_nthreads.h"
72 /*For debugging, start at v(-dt/2) for velolcity verlet -- uncomment next line */
73 /*#define STARTFROMDT2*/
93 /* The random state for ngaussrand threads.
94 * Normal thermostats need just 1 random number generator,
95 * but SD and BD with OpenMP parallelization need 1 for each thread.
103 gmx_sd_sigma_t *sdsig;
106 /* andersen temperature control stuff */
107 gmx_bool *randomize_group;
111 typedef struct gmx_update
114 /* xprime for constraint algorithms */
118 /* variable size arrays for andersen */
121 gmx_bool randatom_list_init;
123 /* Variables for the deform algorithm */
124 gmx_large_int_t deformref_step;
125 matrix deformref_box;
129 static void do_update_md(int start, int nrend, double dt,
130 t_grp_tcstat *tcstat,
132 gmx_bool bNEMD, t_grp_acc *gstat, rvec accel[],
135 unsigned short ptype[], unsigned short cFREEZE[],
136 unsigned short cACC[], unsigned short cTC[],
137 rvec x[], rvec xprime[], rvec v[],
139 gmx_bool bNH, gmx_bool bPR)
142 int gf = 0, ga = 0, gt = 0;
144 real vn, vv, va, vb, vnrel;
150 /* Update with coupling to extended ensembles, used for
151 * Nose-Hoover and Parrinello-Rahman coupling
152 * Nose-Hoover uses the reversible leap-frog integrator from
153 * Holian et al. Phys Rev E 52(3) : 2338, 1995
155 for (n = start; n < nrend; n++)
170 lg = tcstat[gt].lambda;
175 rvec_sub(v[n], gstat[ga].u, vrel);
177 for (d = 0; d < DIM; d++)
179 if ((ptype[n] != eptVSite) && (ptype[n] != eptShell) && !nFreeze[gf][d])
181 vnrel = (lg*vrel[d] + dt*(imass*f[n][d] - 0.5*vxi*vrel[d]
182 - iprod(M[d], vrel)))/(1 + 0.5*vxi*dt);
183 /* do not scale the mean velocities u */
184 vn = gstat[ga].u[d] + accel[ga][d]*dt + vnrel;
186 xprime[n][d] = x[n][d]+vn*dt;
191 xprime[n][d] = x[n][d];
196 else if (cFREEZE != NULL ||
197 nFreeze[0][XX] || nFreeze[0][YY] || nFreeze[0][ZZ] ||
200 /* Update with Berendsen/v-rescale coupling and freeze or NEMD */
201 for (n = start; n < nrend; n++)
203 w_dt = invmass[n]*dt;
216 lg = tcstat[gt].lambda;
218 for (d = 0; d < DIM; d++)
221 if ((ptype[n] != eptVSite) && (ptype[n] != eptShell) && !nFreeze[gf][d])
223 vv = lg*vn + f[n][d]*w_dt;
225 /* do not scale the mean velocities u */
227 va = vv + accel[ga][d]*dt;
228 vb = va + (1.0-lg)*u;
230 xprime[n][d] = x[n][d]+vb*dt;
235 xprime[n][d] = x[n][d];
242 /* Plain update with Berendsen/v-rescale coupling */
243 for (n = start; n < nrend; n++)
245 if ((ptype[n] != eptVSite) && (ptype[n] != eptShell))
247 w_dt = invmass[n]*dt;
252 lg = tcstat[gt].lambda;
254 for (d = 0; d < DIM; d++)
256 vn = lg*v[n][d] + f[n][d]*w_dt;
258 xprime[n][d] = x[n][d] + vn*dt;
263 for (d = 0; d < DIM; d++)
266 xprime[n][d] = x[n][d];
273 static void do_update_vv_vel(int start, int nrend, double dt,
274 rvec accel[], ivec nFreeze[], real invmass[],
275 unsigned short ptype[], unsigned short cFREEZE[],
276 unsigned short cACC[], rvec v[], rvec f[],
277 gmx_bool bExtended, real veta, real alpha)
282 real u, vn, vv, va, vb, vnrel;
288 g = 0.25*dt*veta*alpha;
290 mv2 = series_sinhx(g);
297 for (n = start; n < nrend; n++)
299 w_dt = invmass[n]*dt;
309 for (d = 0; d < DIM; d++)
311 if ((ptype[n] != eptVSite) && (ptype[n] != eptShell) && !nFreeze[gf][d])
313 v[n][d] = mv1*(mv1*v[n][d] + 0.5*(w_dt*mv2*f[n][d]))+0.5*accel[ga][d]*dt;
321 } /* do_update_vv_vel */
323 static void do_update_vv_pos(int start, int nrend, double dt,
325 unsigned short ptype[], unsigned short cFREEZE[],
326 rvec x[], rvec xprime[], rvec v[],
327 gmx_bool bExtended, real veta)
334 /* Would it make more sense if Parrinello-Rahman was put here? */
339 mr2 = series_sinhx(g);
347 for (n = start; n < nrend; n++)
355 for (d = 0; d < DIM; d++)
357 if ((ptype[n] != eptVSite) && (ptype[n] != eptShell) && !nFreeze[gf][d])
359 xprime[n][d] = mr1*(mr1*x[n][d]+mr2*dt*v[n][d]);
363 xprime[n][d] = x[n][d];
367 } /* do_update_vv_pos */
369 static void do_update_visc(int start, int nrend, double dt,
370 t_grp_tcstat *tcstat,
373 unsigned short ptype[], unsigned short cTC[],
374 rvec x[], rvec xprime[], rvec v[],
375 rvec f[], matrix M, matrix box, real
376 cos_accel, real vcos,
377 gmx_bool bNH, gmx_bool bPR)
382 real lg, vxi = 0, vv;
387 fac = 2*M_PI/(box[ZZ][ZZ]);
391 /* Update with coupling to extended ensembles, used for
392 * Nose-Hoover and Parrinello-Rahman coupling
394 for (n = start; n < nrend; n++)
401 lg = tcstat[gt].lambda;
402 cosz = cos(fac*x[n][ZZ]);
404 copy_rvec(v[n], vrel);
412 for (d = 0; d < DIM; d++)
416 if ((ptype[n] != eptVSite) && (ptype[n] != eptShell))
418 vn = (lg*vrel[d] + dt*(imass*f[n][d] - 0.5*vxi*vrel[d]
419 - iprod(M[d], vrel)))/(1 + 0.5*vxi*dt);
422 vn += vc + dt*cosz*cos_accel;
425 xprime[n][d] = x[n][d]+vn*dt;
429 xprime[n][d] = x[n][d];
436 /* Classic version of update, used with berendsen coupling */
437 for (n = start; n < nrend; n++)
439 w_dt = invmass[n]*dt;
444 lg = tcstat[gt].lambda;
445 cosz = cos(fac*x[n][ZZ]);
447 for (d = 0; d < DIM; d++)
451 if ((ptype[n] != eptVSite) && (ptype[n] != eptShell))
456 /* Do not scale the cosine velocity profile */
457 vv = vc + lg*(vn - vc + f[n][d]*w_dt);
458 /* Add the cosine accelaration profile */
459 vv += dt*cosz*cos_accel;
463 vv = lg*(vn + f[n][d]*w_dt);
466 xprime[n][d] = x[n][d]+vv*dt;
471 xprime[n][d] = x[n][d];
478 /* Allocates and initializes sd->gaussrand[i] for i=1, i<sd->ngaussrand,
479 * Using seeds generated from sd->gaussrand[0].
481 static void init_multiple_gaussrand(gmx_stochd_t *sd)
486 ngr = sd->ngaussrand;
489 for (i = 1; i < ngr; i++)
491 seed[i] = gmx_rng_uniform_uint32(sd->gaussrand[0]);
494 if (ngr != gmx_omp_nthreads_get(emntUpdate))
496 gmx_incons("The number of Gaussian number generators should be equal to gmx_omp_nthreads_get(emntUpdate)");
499 #pragma omp parallel num_threads(gmx_omp_nthreads_get(emntUpdate))
503 th = gmx_omp_get_thread_num();
506 /* Initialize on each thread to get memory allocated thread-local */
507 sd->gaussrand[th] = gmx_rng_init(seed[th]);
514 static gmx_stochd_t *init_stochd(t_inputrec *ir, int nthreads)
523 /* Initiate random number generator for langevin type dynamics,
524 * for BD, SD or velocity rescaling temperature coupling.
526 if (ir->eI == eiBD || EI_SD(ir->eI))
528 sd->ngaussrand = nthreads;
534 snew(sd->gaussrand, sd->ngaussrand);
536 /* Initialize the first random generator */
537 sd->gaussrand[0] = gmx_rng_init(ir->ld_seed);
539 if (sd->ngaussrand > 1)
541 /* Initialize the rest of the random number generators,
542 * using the first one to generate seeds.
544 init_multiple_gaussrand(sd);
547 ngtc = ir->opts.ngtc;
551 snew(sd->bd_rf, ngtc);
553 else if (EI_SD(ir->eI))
556 snew(sd->sdsig, ngtc);
559 for (n = 0; n < ngtc; n++)
561 if (ir->opts.tau_t[n] > 0)
563 sdc[n].gdt = ir->delta_t/ir->opts.tau_t[n];
564 sdc[n].eph = exp(sdc[n].gdt/2);
565 sdc[n].emh = exp(-sdc[n].gdt/2);
566 sdc[n].em = exp(-sdc[n].gdt);
570 /* No friction and noise on this group */
576 if (sdc[n].gdt >= 0.05)
578 sdc[n].b = sdc[n].gdt*(sdc[n].eph*sdc[n].eph - 1)
579 - 4*(sdc[n].eph - 1)*(sdc[n].eph - 1);
580 sdc[n].c = sdc[n].gdt - 3 + 4*sdc[n].emh - sdc[n].em;
581 sdc[n].d = 2 - sdc[n].eph - sdc[n].emh;
586 /* Seventh order expansions for small y */
587 sdc[n].b = y*y*y*y*(1/3.0+y*(1/3.0+y*(17/90.0+y*7/9.0)));
588 sdc[n].c = y*y*y*(2/3.0+y*(-1/2.0+y*(7/30.0+y*(-1/12.0+y*31/1260.0))));
589 sdc[n].d = y*y*(-1+y*y*(-1/12.0-y*y/360.0));
593 fprintf(debug, "SD const tc-grp %d: b %g c %g d %g\n",
594 n, sdc[n].b, sdc[n].c, sdc[n].d);
598 else if (ETC_ANDERSEN(ir->etc))
607 snew(sd->randomize_group, ngtc);
608 snew(sd->boltzfac, ngtc);
610 /* for now, assume that all groups, if randomized, are randomized at the same rate, i.e. tau_t is the same. */
611 /* since constraint groups don't necessarily match up with temperature groups! This is checked in readir.c */
613 for (n = 0; n < ngtc; n++)
615 reft = max(0.0, opts->ref_t[n]);
616 if ((opts->tau_t[n] > 0) && (reft > 0)) /* tau_t or ref_t = 0 means that no randomization is done */
618 sd->randomize_group[n] = TRUE;
619 sd->boltzfac[n] = BOLTZ*opts->ref_t[n];
623 sd->randomize_group[n] = FALSE;
630 void get_stochd_state(gmx_update_t upd, t_state *state)
632 /* Note that we only get the state of the first random generator,
633 * even if there are multiple. This avoids repetition.
635 gmx_rng_get_state(upd->sd->gaussrand[0], state->ld_rng, state->ld_rngi);
638 void set_stochd_state(gmx_update_t upd, t_state *state)
645 gmx_rng_set_state(sd->gaussrand[0], state->ld_rng, state->ld_rngi[0]);
647 if (sd->ngaussrand > 1)
649 /* We only end up here with SD or BD with OpenMP.
650 * Destroy and reinitialize the rest of the random number generators,
651 * using seeds generated from the first one.
652 * Although this doesn't recover the previous state,
653 * it at least avoids repetition, which is most important.
654 * Exaclty restoring states with all MPI+OpenMP setups is difficult
655 * and as the integrator is random to start with, doesn't gain us much.
657 for (i = 1; i < sd->ngaussrand; i++)
659 gmx_rng_destroy(sd->gaussrand[i]);
662 init_multiple_gaussrand(sd);
666 gmx_update_t init_update(t_inputrec *ir)
672 if (ir->eI == eiBD || EI_SD(ir->eI) || ir->etc == etcVRESCALE || ETC_ANDERSEN(ir->etc))
674 upd->sd = init_stochd(ir, gmx_omp_nthreads_get(emntUpdate));
679 upd->randatom = NULL;
680 upd->randatom_list = NULL;
681 upd->randatom_list_init = FALSE; /* we have not yet cleared the data structure at this point */
686 static void do_update_sd1(gmx_stochd_t *sd,
688 int start, int nrend, double dt,
689 rvec accel[], ivec nFreeze[],
690 real invmass[], unsigned short ptype[],
691 unsigned short cFREEZE[], unsigned short cACC[],
692 unsigned short cTC[],
693 rvec x[], rvec xprime[], rvec v[], rvec f[],
694 int ngtc, real tau_t[], real ref_t[])
699 int gf = 0, ga = 0, gt = 0;
706 for (n = 0; n < ngtc; n++)
709 /* The mass is encounted for later, since this differs per atom */
710 sig[n].V = sqrt(kT*(1 - sdc[n].em*sdc[n].em));
713 for (n = start; n < nrend; n++)
715 ism = sqrt(invmass[n]);
729 for (d = 0; d < DIM; d++)
731 if ((ptype[n] != eptVSite) && (ptype[n] != eptShell) && !nFreeze[gf][d])
733 sd_V = ism*sig[gt].V*gmx_rng_gaussian_table(gaussrand);
735 v[n][d] = v[n][d]*sdc[gt].em
736 + (invmass[n]*f[n][d] + accel[ga][d])*tau_t[gt]*(1 - sdc[gt].em)
739 xprime[n][d] = x[n][d] + v[n][d]*dt;
744 xprime[n][d] = x[n][d];
750 static void check_sd2_work_data_allocation(gmx_stochd_t *sd, int nrend)
752 if (nrend > sd->sd_V_nalloc)
754 sd->sd_V_nalloc = over_alloc_dd(nrend);
755 srenew(sd->sd_V, sd->sd_V_nalloc);
759 static void do_update_sd2_Tconsts(gmx_stochd_t *sd,
764 /* This is separated from the update below, because it is single threaded */
773 for (gt = 0; gt < ngtc; gt++)
775 kT = BOLTZ*ref_t[gt];
776 /* The mass is encounted for later, since this differs per atom */
777 sig[gt].V = sqrt(kT*(1-sdc[gt].em));
778 sig[gt].X = sqrt(kT*sqr(tau_t[gt])*sdc[gt].c);
779 sig[gt].Yv = sqrt(kT*sdc[gt].b/sdc[gt].c);
780 sig[gt].Yx = sqrt(kT*sqr(tau_t[gt])*sdc[gt].b/(1-sdc[gt].em));
784 static void do_update_sd2(gmx_stochd_t *sd,
787 int start, int nrend,
788 rvec accel[], ivec nFreeze[],
789 real invmass[], unsigned short ptype[],
790 unsigned short cFREEZE[], unsigned short cACC[],
791 unsigned short cTC[],
792 rvec x[], rvec xprime[], rvec v[], rvec f[],
799 /* The random part of the velocity update, generated in the first
800 * half of the update, needs to be remembered for the second half.
804 int gf = 0, ga = 0, gt = 0;
805 real vn = 0, Vmh, Xmh;
813 for (n = start; n < nrend; n++)
815 ism = sqrt(invmass[n]);
829 for (d = 0; d < DIM; d++)
835 if ((ptype[n] != eptVSite) && (ptype[n] != eptShell) && !nFreeze[gf][d])
841 sd_X[n][d] = ism*sig[gt].X*gmx_rng_gaussian_table(gaussrand);
843 Vmh = sd_X[n][d]*sdc[gt].d/(tau_t[gt]*sdc[gt].c)
844 + ism*sig[gt].Yv*gmx_rng_gaussian_table(gaussrand);
845 sd_V[n][d] = ism*sig[gt].V*gmx_rng_gaussian_table(gaussrand);
847 v[n][d] = vn*sdc[gt].em
848 + (invmass[n]*f[n][d] + accel[ga][d])*tau_t[gt]*(1 - sdc[gt].em)
849 + sd_V[n][d] - sdc[gt].em*Vmh;
851 xprime[n][d] = x[n][d] + v[n][d]*tau_t[gt]*(sdc[gt].eph - sdc[gt].emh);
856 /* Correct the velocities for the constraints.
857 * This operation introduces some inaccuracy,
858 * since the velocity is determined from differences in coordinates.
861 (xprime[n][d] - x[n][d])/(tau_t[gt]*(sdc[gt].eph - sdc[gt].emh));
863 Xmh = sd_V[n][d]*tau_t[gt]*sdc[gt].d/(sdc[gt].em-1)
864 + ism*sig[gt].Yx*gmx_rng_gaussian_table(gaussrand);
865 sd_X[n][d] = ism*sig[gt].X*gmx_rng_gaussian_table(gaussrand);
867 xprime[n][d] += sd_X[n][d] - Xmh;
876 xprime[n][d] = x[n][d];
883 static void do_update_bd_Tconsts(double dt, real friction_coefficient,
884 int ngtc, const real ref_t[],
887 /* This is separated from the update below, because it is single threaded */
890 if (friction_coefficient != 0)
892 for (gt = 0; gt < ngtc; gt++)
894 rf[gt] = sqrt(2.0*BOLTZ*ref_t[gt]/(friction_coefficient*dt));
899 for (gt = 0; gt < ngtc; gt++)
901 rf[gt] = sqrt(2.0*BOLTZ*ref_t[gt]);
906 static void do_update_bd(int start, int nrend, double dt,
908 real invmass[], unsigned short ptype[],
909 unsigned short cFREEZE[], unsigned short cTC[],
910 rvec x[], rvec xprime[], rvec v[],
911 rvec f[], real friction_coefficient,
912 real *rf, gmx_rng_t gaussrand)
914 /* note -- these appear to be full step velocities . . . */
920 if (friction_coefficient != 0)
922 invfr = 1.0/friction_coefficient;
925 for (n = start; (n < nrend); n++)
935 for (d = 0; (d < DIM); d++)
937 if ((ptype[n] != eptVSite) && (ptype[n] != eptShell) && !nFreeze[gf][d])
939 if (friction_coefficient != 0)
941 vn = invfr*f[n][d] + rf[gt]*gmx_rng_gaussian_table(gaussrand);
945 /* NOTE: invmass = 2/(mass*friction_constant*dt) */
946 vn = 0.5*invmass[n]*f[n][d]*dt
947 + sqrt(0.5*invmass[n])*rf[gt]*gmx_rng_gaussian_table(gaussrand);
951 xprime[n][d] = x[n][d]+vn*dt;
956 xprime[n][d] = x[n][d];
962 static void dump_it_all(FILE gmx_unused *fp, const char gmx_unused *title,
963 int gmx_unused natoms, rvec gmx_unused x[], rvec gmx_unused xp[],
964 rvec gmx_unused v[], rvec gmx_unused f[])
969 fprintf(fp, "%s\n", title);
970 pr_rvecs(fp, 0, "x", x, natoms);
971 pr_rvecs(fp, 0, "xp", xp, natoms);
972 pr_rvecs(fp, 0, "v", v, natoms);
973 pr_rvecs(fp, 0, "f", f, natoms);
978 static void calc_ke_part_normal(rvec v[], t_grpopts *opts, t_mdatoms *md,
979 gmx_ekindata_t *ekind, t_nrnb *nrnb, gmx_bool bEkinAveVel,
980 gmx_bool bSaveEkinOld)
983 t_grp_tcstat *tcstat = ekind->tcstat;
984 t_grp_acc *grpstat = ekind->grpstat;
987 /* three main: VV with AveVel, vv with AveEkin, leap with AveEkin. Leap with AveVel is also
988 an option, but not supported now. Additionally, if we are doing iterations.
989 bEkinAveVel: If TRUE, we sum into ekin, if FALSE, into ekinh.
990 bSavEkinOld: If TRUE (in the case of iteration = bIterate is TRUE), we don't copy over the ekinh_old.
991 If FALSE, we overrwrite it.
994 /* group velocities are calculated in update_ekindata and
995 * accumulated in acumulate_groups.
996 * Now the partial global and groups ekin.
998 for (g = 0; (g < opts->ngtc); g++)
1003 copy_mat(tcstat[g].ekinh, tcstat[g].ekinh_old);
1007 clear_mat(tcstat[g].ekinf);
1011 clear_mat(tcstat[g].ekinh);
1015 tcstat[g].ekinscalef_nhc = 1.0; /* need to clear this -- logic is complicated! */
1018 ekind->dekindl_old = ekind->dekindl;
1020 nthread = gmx_omp_nthreads_get(emntUpdate);
1022 #pragma omp parallel for num_threads(nthread) schedule(static)
1023 for (thread = 0; thread < nthread; thread++)
1025 int start_t, end_t, n;
1033 start_t = md->start + ((thread+0)*md->homenr)/nthread;
1034 end_t = md->start + ((thread+1)*md->homenr)/nthread;
1036 ekin_sum = ekind->ekin_work[thread];
1037 dekindl_sum = ekind->dekindl_work[thread];
1039 for (gt = 0; gt < opts->ngtc; gt++)
1041 clear_mat(ekin_sum[gt]);
1047 for (n = start_t; n < end_t; n++)
1057 hm = 0.5*md->massT[n];
1059 for (d = 0; (d < DIM); d++)
1061 v_corrt[d] = v[n][d] - grpstat[ga].u[d];
1063 for (d = 0; (d < DIM); d++)
1065 for (m = 0; (m < DIM); m++)
1067 /* if we're computing a full step velocity, v_corrt[d] has v(t). Otherwise, v(t+dt/2) */
1068 ekin_sum[gt][m][d] += hm*v_corrt[m]*v_corrt[d];
1071 if (md->nMassPerturbed && md->bPerturbed[n])
1074 0.5*(md->massB[n] - md->massA[n])*iprod(v_corrt, v_corrt);
1080 for (thread = 0; thread < nthread; thread++)
1082 for (g = 0; g < opts->ngtc; g++)
1086 m_add(tcstat[g].ekinf, ekind->ekin_work[thread][g],
1091 m_add(tcstat[g].ekinh, ekind->ekin_work[thread][g],
1096 ekind->dekindl += *ekind->dekindl_work[thread];
1099 inc_nrnb(nrnb, eNR_EKIN, md->homenr);
1102 static void calc_ke_part_visc(matrix box, rvec x[], rvec v[],
1103 t_grpopts *opts, t_mdatoms *md,
1104 gmx_ekindata_t *ekind,
1105 t_nrnb *nrnb, gmx_bool bEkinAveVel)
1107 int start = md->start, homenr = md->homenr;
1108 int g, d, n, m, gt = 0;
1111 t_grp_tcstat *tcstat = ekind->tcstat;
1112 t_cos_acc *cosacc = &(ekind->cosacc);
1117 for (g = 0; g < opts->ngtc; g++)
1119 copy_mat(ekind->tcstat[g].ekinh, ekind->tcstat[g].ekinh_old);
1120 clear_mat(ekind->tcstat[g].ekinh);
1122 ekind->dekindl_old = ekind->dekindl;
1124 fac = 2*M_PI/box[ZZ][ZZ];
1127 for (n = start; n < start+homenr; n++)
1133 hm = 0.5*md->massT[n];
1135 /* Note that the times of x and v differ by half a step */
1136 /* MRS -- would have to be changed for VV */
1137 cosz = cos(fac*x[n][ZZ]);
1138 /* Calculate the amplitude of the new velocity profile */
1139 mvcos += 2*cosz*md->massT[n]*v[n][XX];
1141 copy_rvec(v[n], v_corrt);
1142 /* Subtract the profile for the kinetic energy */
1143 v_corrt[XX] -= cosz*cosacc->vcos;
1144 for (d = 0; (d < DIM); d++)
1146 for (m = 0; (m < DIM); m++)
1148 /* if we're computing a full step velocity, v_corrt[d] has v(t). Otherwise, v(t+dt/2) */
1151 tcstat[gt].ekinf[m][d] += hm*v_corrt[m]*v_corrt[d];
1155 tcstat[gt].ekinh[m][d] += hm*v_corrt[m]*v_corrt[d];
1159 if (md->nPerturbed && md->bPerturbed[n])
1161 dekindl += 0.5*(md->massB[n] - md->massA[n])*iprod(v_corrt, v_corrt);
1164 ekind->dekindl = dekindl;
1165 cosacc->mvcos = mvcos;
1167 inc_nrnb(nrnb, eNR_EKIN, homenr);
1170 void calc_ke_part(t_state *state, t_grpopts *opts, t_mdatoms *md,
1171 gmx_ekindata_t *ekind, t_nrnb *nrnb, gmx_bool bEkinAveVel, gmx_bool bSaveEkinOld)
1173 if (ekind->cosacc.cos_accel == 0)
1175 calc_ke_part_normal(state->v, opts, md, ekind, nrnb, bEkinAveVel, bSaveEkinOld);
1179 calc_ke_part_visc(state->box, state->x, state->v, opts, md, ekind, nrnb, bEkinAveVel);
1183 extern void init_ekinstate(ekinstate_t *ekinstate, const t_inputrec *ir)
1185 ekinstate->ekin_n = ir->opts.ngtc;
1186 snew(ekinstate->ekinh, ekinstate->ekin_n);
1187 snew(ekinstate->ekinf, ekinstate->ekin_n);
1188 snew(ekinstate->ekinh_old, ekinstate->ekin_n);
1189 snew(ekinstate->ekinscalef_nhc, ekinstate->ekin_n);
1190 snew(ekinstate->ekinscaleh_nhc, ekinstate->ekin_n);
1191 snew(ekinstate->vscale_nhc, ekinstate->ekin_n);
1192 ekinstate->dekindl = 0;
1193 ekinstate->mvcos = 0;
1196 void update_ekinstate(ekinstate_t *ekinstate, gmx_ekindata_t *ekind)
1200 for (i = 0; i < ekinstate->ekin_n; i++)
1202 copy_mat(ekind->tcstat[i].ekinh, ekinstate->ekinh[i]);
1203 copy_mat(ekind->tcstat[i].ekinf, ekinstate->ekinf[i]);
1204 copy_mat(ekind->tcstat[i].ekinh_old, ekinstate->ekinh_old[i]);
1205 ekinstate->ekinscalef_nhc[i] = ekind->tcstat[i].ekinscalef_nhc;
1206 ekinstate->ekinscaleh_nhc[i] = ekind->tcstat[i].ekinscaleh_nhc;
1207 ekinstate->vscale_nhc[i] = ekind->tcstat[i].vscale_nhc;
1210 copy_mat(ekind->ekin, ekinstate->ekin_total);
1211 ekinstate->dekindl = ekind->dekindl;
1212 ekinstate->mvcos = ekind->cosacc.mvcos;
1216 void restore_ekinstate_from_state(t_commrec *cr,
1217 gmx_ekindata_t *ekind, ekinstate_t *ekinstate)
1223 for (i = 0; i < ekinstate->ekin_n; i++)
1225 copy_mat(ekinstate->ekinh[i], ekind->tcstat[i].ekinh);
1226 copy_mat(ekinstate->ekinf[i], ekind->tcstat[i].ekinf);
1227 copy_mat(ekinstate->ekinh_old[i], ekind->tcstat[i].ekinh_old);
1228 ekind->tcstat[i].ekinscalef_nhc = ekinstate->ekinscalef_nhc[i];
1229 ekind->tcstat[i].ekinscaleh_nhc = ekinstate->ekinscaleh_nhc[i];
1230 ekind->tcstat[i].vscale_nhc = ekinstate->vscale_nhc[i];
1233 copy_mat(ekinstate->ekin_total, ekind->ekin);
1235 ekind->dekindl = ekinstate->dekindl;
1236 ekind->cosacc.mvcos = ekinstate->mvcos;
1237 n = ekinstate->ekin_n;
1242 gmx_bcast(sizeof(n), &n, cr);
1243 for (i = 0; i < n; i++)
1245 gmx_bcast(DIM*DIM*sizeof(ekind->tcstat[i].ekinh[0][0]),
1246 ekind->tcstat[i].ekinh[0], cr);
1247 gmx_bcast(DIM*DIM*sizeof(ekind->tcstat[i].ekinf[0][0]),
1248 ekind->tcstat[i].ekinf[0], cr);
1249 gmx_bcast(DIM*DIM*sizeof(ekind->tcstat[i].ekinh_old[0][0]),
1250 ekind->tcstat[i].ekinh_old[0], cr);
1252 gmx_bcast(sizeof(ekind->tcstat[i].ekinscalef_nhc),
1253 &(ekind->tcstat[i].ekinscalef_nhc), cr);
1254 gmx_bcast(sizeof(ekind->tcstat[i].ekinscaleh_nhc),
1255 &(ekind->tcstat[i].ekinscaleh_nhc), cr);
1256 gmx_bcast(sizeof(ekind->tcstat[i].vscale_nhc),
1257 &(ekind->tcstat[i].vscale_nhc), cr);
1259 gmx_bcast(DIM*DIM*sizeof(ekind->ekin[0][0]),
1260 ekind->ekin[0], cr);
1262 gmx_bcast(sizeof(ekind->dekindl), &ekind->dekindl, cr);
1263 gmx_bcast(sizeof(ekind->cosacc.mvcos), &ekind->cosacc.mvcos, cr);
1267 void set_deform_reference_box(gmx_update_t upd, gmx_large_int_t step, matrix box)
1269 upd->deformref_step = step;
1270 copy_mat(box, upd->deformref_box);
1273 static void deform(gmx_update_t upd,
1274 int start, int homenr, rvec x[], matrix box, matrix *scale_tot,
1275 const t_inputrec *ir, gmx_large_int_t step)
1277 matrix bnew, invbox, mu;
1281 elapsed_time = (step + 1 - upd->deformref_step)*ir->delta_t;
1282 copy_mat(box, bnew);
1283 for (i = 0; i < DIM; i++)
1285 for (j = 0; j < DIM; j++)
1287 if (ir->deform[i][j] != 0)
1290 upd->deformref_box[i][j] + elapsed_time*ir->deform[i][j];
1294 /* We correct the off-diagonal elements,
1295 * which can grow indefinitely during shearing,
1296 * so the shifts do not get messed up.
1298 for (i = 1; i < DIM; i++)
1300 for (j = i-1; j >= 0; j--)
1302 while (bnew[i][j] - box[i][j] > 0.5*bnew[j][j])
1304 rvec_dec(bnew[i], bnew[j]);
1306 while (bnew[i][j] - box[i][j] < -0.5*bnew[j][j])
1308 rvec_inc(bnew[i], bnew[j]);
1312 m_inv_ur0(box, invbox);
1313 copy_mat(bnew, box);
1314 mmul_ur0(box, invbox, mu);
1316 for (i = start; i < start+homenr; i++)
1318 x[i][XX] = mu[XX][XX]*x[i][XX]+mu[YY][XX]*x[i][YY]+mu[ZZ][XX]*x[i][ZZ];
1319 x[i][YY] = mu[YY][YY]*x[i][YY]+mu[ZZ][YY]*x[i][ZZ];
1320 x[i][ZZ] = mu[ZZ][ZZ]*x[i][ZZ];
1324 /* The transposes of the scaling matrices are stored,
1325 * so we need to do matrix multiplication in the inverse order.
1327 mmul_ur0(*scale_tot, mu, *scale_tot);
1331 static void combine_forces(int nstcalclr,
1332 gmx_constr_t constr,
1333 t_inputrec *ir, t_mdatoms *md, t_idef *idef,
1335 gmx_large_int_t step,
1336 t_state *state, gmx_bool bMolPBC,
1337 int start, int nrend,
1338 rvec f[], rvec f_lr[],
1343 /* f contains the short-range forces + the long range forces
1344 * which are stored separately in f_lr.
1347 if (constr != NULL && !(ir->eConstrAlg == econtSHAKE && ir->epc == epcNO))
1349 /* We need to constrain the LR forces separately,
1350 * because due to the different pre-factor for the SR and LR
1351 * forces in the update algorithm, we can not determine
1352 * the constraint force for the coordinate constraining.
1353 * Constrain only the additional LR part of the force.
1355 /* MRS -- need to make sure this works with trotter integration -- the constraint calls may not be right.*/
1356 constrain(NULL, FALSE, FALSE, constr, idef, ir, NULL, cr, step, 0, md,
1357 state->x, f_lr, f_lr, bMolPBC, state->box, state->lambda[efptBONDED], NULL,
1358 NULL, NULL, nrnb, econqForce, ir->epc == epcMTTK, state->veta, state->veta);
1361 /* Add nstcalclr-1 times the LR force to the sum of both forces
1362 * and store the result in forces_lr.
1364 nm1 = nstcalclr - 1;
1365 for (i = start; i < nrend; i++)
1367 for (d = 0; d < DIM; d++)
1369 f_lr[i][d] = f[i][d] + nm1*f_lr[i][d];
1374 void update_tcouple(gmx_large_int_t step,
1375 t_inputrec *inputrec,
1377 gmx_ekindata_t *ekind,
1383 gmx_bool bTCouple = FALSE;
1385 int i, start, end, homenr, offset;
1387 /* if using vv with trotter decomposition methods, we do this elsewhere in the code */
1388 if (inputrec->etc != etcNO &&
1389 !(IR_NVT_TROTTER(inputrec) || IR_NPT_TROTTER(inputrec) || IR_NPH_TROTTER(inputrec)))
1391 /* We should only couple after a step where energies were determined (for leapfrog versions)
1392 or the step energies are determined, for velocity verlet versions */
1394 if (EI_VV(inputrec->eI))
1402 bTCouple = (inputrec->nsttcouple == 1 ||
1403 do_per_step(step+inputrec->nsttcouple-offset,
1404 inputrec->nsttcouple));
1409 dttc = inputrec->nsttcouple*inputrec->delta_t;
1411 switch (inputrec->etc)
1416 berendsen_tcoupl(inputrec, ekind, dttc);
1419 nosehoover_tcoupl(&(inputrec->opts), ekind, dttc,
1420 state->nosehoover_xi, state->nosehoover_vxi, MassQ);
1423 vrescale_tcoupl(inputrec, ekind, dttc,
1424 state->therm_integral, upd->sd->gaussrand[0]);
1427 /* rescale in place here */
1428 if (EI_VV(inputrec->eI))
1430 rescale_velocities(ekind, md, md->start, md->start+md->homenr, state->v);
1435 /* Set the T scaling lambda to 1 to have no scaling */
1436 for (i = 0; (i < inputrec->opts.ngtc); i++)
1438 ekind->tcstat[i].lambda = 1.0;
1443 void update_pcouple(FILE *fplog,
1444 gmx_large_int_t step,
1445 t_inputrec *inputrec,
1451 gmx_bool bPCouple = FALSE;
1455 /* if using Trotter pressure, we do this in coupling.c, so we leave it false. */
1456 if (inputrec->epc != epcNO && (!(IR_NPT_TROTTER(inputrec) || IR_NPH_TROTTER(inputrec))))
1458 /* We should only couple after a step where energies were determined */
1459 bPCouple = (inputrec->nstpcouple == 1 ||
1460 do_per_step(step+inputrec->nstpcouple-1,
1461 inputrec->nstpcouple));
1464 clear_mat(pcoupl_mu);
1465 for (i = 0; i < DIM; i++)
1467 pcoupl_mu[i][i] = 1.0;
1474 dtpc = inputrec->nstpcouple*inputrec->delta_t;
1476 switch (inputrec->epc)
1478 /* We can always pcoupl, even if we did not sum the energies
1479 * the previous step, since state->pres_prev is only updated
1480 * when the energies have been summed.
1484 case (epcBERENDSEN):
1487 berendsen_pcoupl(fplog, step, inputrec, dtpc, state->pres_prev, state->box,
1491 case (epcPARRINELLORAHMAN):
1492 parrinellorahman_pcoupl(fplog, step, inputrec, dtpc, state->pres_prev,
1493 state->box, state->box_rel, state->boxv,
1494 M, pcoupl_mu, bInitStep);
1502 static rvec *get_xprime(const t_state *state, gmx_update_t upd)
1504 if (state->nalloc > upd->xp_nalloc)
1506 upd->xp_nalloc = state->nalloc;
1507 srenew(upd->xp, upd->xp_nalloc);
1513 void update_constraints(FILE *fplog,
1514 gmx_large_int_t step,
1515 real *dvdlambda, /* the contribution to be added to the bonded interactions */
1516 t_inputrec *inputrec, /* input record and box stuff */
1517 gmx_ekindata_t *ekind,
1522 rvec force[], /* forces on home particles */
1527 gmx_wallcycle_t wcycle,
1529 gmx_constr_t constr,
1530 gmx_bool bFirstHalf,
1534 gmx_bool bExtended, bLastStep, bLog = FALSE, bEner = FALSE, bDoConstr = FALSE;
1537 int start, homenr, nrend, i, n, m, g, d;
1539 rvec *vbuf, *xprime = NULL;
1546 if (bFirstHalf && !EI_VV(inputrec->eI))
1551 /* for now, SD update is here -- though it really seems like it
1552 should be reformulated as a velocity verlet method, since it has two parts */
1555 homenr = md->homenr;
1556 nrend = start+homenr;
1558 dt = inputrec->delta_t;
1563 * APPLY CONSTRAINTS:
1566 * When doing PR pressure coupling we have to constrain the
1567 * bonds in each iteration. If we are only using Nose-Hoover tcoupling
1568 * it is enough to do this once though, since the relative velocities
1569 * after this will be normal to the bond vector
1574 /* clear out constraints before applying */
1575 clear_mat(vir_part);
1577 xprime = get_xprime(state, upd);
1579 bLastStep = (step == inputrec->init_step+inputrec->nsteps);
1580 bLog = (do_per_step(step, inputrec->nstlog) || bLastStep || (step < 0));
1581 bEner = (do_per_step(step, inputrec->nstenergy) || bLastStep);
1582 /* Constrain the coordinates xprime */
1583 wallcycle_start(wcycle, ewcCONSTR);
1584 if (EI_VV(inputrec->eI) && bFirstHalf)
1586 constrain(NULL, bLog, bEner, constr, idef,
1587 inputrec, ekind, cr, step, 1, md,
1588 state->x, state->v, state->v,
1589 bMolPBC, state->box,
1590 state->lambda[efptBONDED], dvdlambda,
1591 NULL, bCalcVir ? &vir_con : NULL, nrnb, econqVeloc,
1592 inputrec->epc == epcMTTK, state->veta, vetanew);
1596 constrain(NULL, bLog, bEner, constr, idef,
1597 inputrec, ekind, cr, step, 1, md,
1598 state->x, xprime, NULL,
1599 bMolPBC, state->box,
1600 state->lambda[efptBONDED], dvdlambda,
1601 state->v, bCalcVir ? &vir_con : NULL, nrnb, econqCoord,
1602 inputrec->epc == epcMTTK, state->veta, state->veta);
1604 wallcycle_stop(wcycle, ewcCONSTR);
1608 dump_it_all(fplog, "After Shake",
1609 state->natoms, state->x, xprime, state->v, force);
1613 if (inputrec->eI == eiSD2)
1615 /* A correction factor eph is needed for the SD constraint force */
1616 /* Here we can, unfortunately, not have proper corrections
1617 * for different friction constants, so we use the first one.
1619 for (i = 0; i < DIM; i++)
1621 for (m = 0; m < DIM; m++)
1623 vir_part[i][m] += upd->sd->sdc[0].eph*vir_con[i][m];
1629 m_add(vir_part, vir_con, vir_part);
1633 pr_rvecs(debug, 0, "constraint virial", vir_part, DIM);
1639 if ((inputrec->eI == eiSD2) && !(bFirstHalf))
1641 xprime = get_xprime(state, upd);
1643 nth = gmx_omp_nthreads_get(emntUpdate);
1645 #pragma omp parallel for num_threads(nth) schedule(static)
1646 for (th = 0; th < nth; th++)
1648 int start_th, end_th;
1650 start_th = start + ((nrend-start)* th )/nth;
1651 end_th = start + ((nrend-start)*(th+1))/nth;
1653 /* The second part of the SD integration */
1654 do_update_sd2(upd->sd, upd->sd->gaussrand[th],
1655 FALSE, start_th, end_th,
1656 inputrec->opts.acc, inputrec->opts.nFreeze,
1657 md->invmass, md->ptype,
1658 md->cFREEZE, md->cACC, md->cTC,
1659 state->x, xprime, state->v, force, state->sd_X,
1660 inputrec->opts.tau_t,
1663 inc_nrnb(nrnb, eNR_UPDATE, homenr);
1667 /* Constrain the coordinates xprime */
1668 wallcycle_start(wcycle, ewcCONSTR);
1669 constrain(NULL, bLog, bEner, constr, idef,
1670 inputrec, NULL, cr, step, 1, md,
1671 state->x, xprime, NULL,
1672 bMolPBC, state->box,
1673 state->lambda[efptBONDED], dvdlambda,
1674 NULL, NULL, nrnb, econqCoord, FALSE, 0, 0);
1675 wallcycle_stop(wcycle, ewcCONSTR);
1679 /* We must always unshift after updating coordinates; if we did not shake
1680 x was shifted in do_force */
1682 if (!(bFirstHalf)) /* in the first half of vv, no shift. */
1684 if (graph && (graph->nnodes > 0))
1686 unshift_x(graph, state->box, state->x, upd->xp);
1687 if (TRICLINIC(state->box))
1689 inc_nrnb(nrnb, eNR_SHIFTX, 2*graph->nnodes);
1693 inc_nrnb(nrnb, eNR_SHIFTX, graph->nnodes);
1698 #pragma omp parallel for num_threads(gmx_omp_nthreads_get(emntUpdate)) schedule(static)
1699 for (i = start; i < nrend; i++)
1701 copy_rvec(upd->xp[i], state->x[i]);
1705 dump_it_all(fplog, "After unshift",
1706 state->natoms, state->x, upd->xp, state->v, force);
1708 /* ############# END the update of velocities and positions ######### */
1711 void update_box(FILE *fplog,
1712 gmx_large_int_t step,
1713 t_inputrec *inputrec, /* input record and box stuff */
1716 rvec force[], /* forces on home particles */
1722 gmx_bool bExtended, bLastStep, bLog = FALSE, bEner = FALSE;
1725 int start, homenr, nrend, i, n, m, g;
1729 homenr = md->homenr;
1730 nrend = start+homenr;
1733 (inputrec->etc == etcNOSEHOOVER) ||
1734 (inputrec->epc == epcPARRINELLORAHMAN) ||
1735 (inputrec->epc == epcMTTK);
1737 dt = inputrec->delta_t;
1741 /* now update boxes */
1742 switch (inputrec->epc)
1746 case (epcBERENDSEN):
1747 berendsen_pscale(inputrec, pcoupl_mu, state->box, state->box_rel,
1748 start, homenr, state->x, md->cFREEZE, nrnb);
1750 case (epcPARRINELLORAHMAN):
1751 /* The box velocities were updated in do_pr_pcoupl in the update
1752 * iteration, but we dont change the box vectors until we get here
1753 * since we need to be able to shift/unshift above.
1755 for (i = 0; i < DIM; i++)
1757 for (m = 0; m <= i; m++)
1759 state->box[i][m] += dt*state->boxv[i][m];
1762 preserve_box_shape(inputrec, state->box_rel, state->box);
1764 /* Scale the coordinates */
1765 for (n = start; (n < start+homenr); n++)
1767 tmvmul_ur0(pcoupl_mu, state->x[n], state->x[n]);
1771 switch (inputrec->epct)
1773 case (epctISOTROPIC):
1774 /* DIM * eta = ln V. so DIM*eta_new = DIM*eta_old + DIM*dt*veta =>
1775 ln V_new = ln V_old + 3*dt*veta => V_new = V_old*exp(3*dt*veta) =>
1776 Side length scales as exp(veta*dt) */
1778 msmul(state->box, exp(state->veta*dt), state->box);
1780 /* Relate veta to boxv. veta = d(eta)/dT = (1/DIM)*1/V dV/dT.
1781 o If we assume isotropic scaling, and box length scaling
1782 factor L, then V = L^DIM (det(M)). So dV/dt = DIM
1783 L^(DIM-1) dL/dt det(M), and veta = (1/L) dL/dt. The
1784 determinant of B is L^DIM det(M), and the determinant
1785 of dB/dt is (dL/dT)^DIM det (M). veta will be
1786 (det(dB/dT)/det(B))^(1/3). Then since M =
1787 B_new*(vol_new)^(1/3), dB/dT_new = (veta_new)*B(new). */
1789 msmul(state->box, state->veta, state->boxv);
1799 if ((!(IR_NPT_TROTTER(inputrec) || IR_NPH_TROTTER(inputrec))) && scale_tot)
1801 /* The transposes of the scaling matrices are stored,
1802 * therefore we need to reverse the order in the multiplication.
1804 mmul_ur0(*scale_tot, pcoupl_mu, *scale_tot);
1807 if (DEFORM(*inputrec))
1809 deform(upd, start, homenr, state->x, state->box, scale_tot, inputrec, step);
1812 dump_it_all(fplog, "After update",
1813 state->natoms, state->x, upd->xp, state->v, force);
1816 void update_coords(FILE *fplog,
1817 gmx_large_int_t step,
1818 t_inputrec *inputrec, /* input record and box stuff */
1822 rvec *f, /* forces on home particles */
1826 gmx_ekindata_t *ekind,
1831 t_commrec *cr, /* these shouldn't be here -- need to think about it */
1833 gmx_constr_t constr,
1836 gmx_bool bNH, bPR, bLastStep, bLog = FALSE, bEner = FALSE;
1838 real *imass, *imassin;
1841 int start, homenr, nrend, i, j, d, n, m, g;
1842 int blen0, blen1, iatom, jatom, nshake, nsettle, nconstr, nexpand;
1845 rvec *vcom, *xcom, *vall, *xall, *xin, *vin, *forcein, *fall, *xpall, *xprimein, *xprime;
1848 /* Running the velocity half does nothing except for velocity verlet */
1849 if ((UpdatePart == etrtVELOCITY1 || UpdatePart == etrtVELOCITY2) &&
1850 !EI_VV(inputrec->eI))
1852 gmx_incons("update_coords called for velocity without VV integrator");
1856 homenr = md->homenr;
1857 nrend = start+homenr;
1859 xprime = get_xprime(state, upd);
1861 dt = inputrec->delta_t;
1864 /* We need to update the NMR restraint history when time averaging is used */
1865 if (state->flags & (1<<estDISRE_RM3TAV))
1867 update_disres_history(fcd, &state->hist);
1869 if (state->flags & (1<<estORIRE_DTAV))
1871 update_orires_history(fcd, &state->hist);
1875 bNH = inputrec->etc == etcNOSEHOOVER;
1876 bPR = ((inputrec->epc == epcPARRINELLORAHMAN) || (inputrec->epc == epcMTTK));
1878 if (bDoLR && inputrec->nstcalclr > 1 && !EI_VV(inputrec->eI)) /* get this working with VV? */
1880 /* Store the total force + nstcalclr-1 times the LR force
1881 * in forces_lr, so it can be used in a normal update algorithm
1882 * to produce twin time stepping.
1884 /* is this correct in the new construction? MRS */
1885 combine_forces(inputrec->nstcalclr, constr, inputrec, md, idef, cr,
1886 step, state, bMolPBC,
1887 start, nrend, f, f_lr, nrnb);
1895 /* ############# START The update of velocities and positions ######### */
1897 dump_it_all(fplog, "Before update",
1898 state->natoms, state->x, xprime, state->v, force);
1900 if (inputrec->eI == eiSD2)
1902 check_sd2_work_data_allocation(upd->sd, nrend);
1904 do_update_sd2_Tconsts(upd->sd,
1905 inputrec->opts.ngtc,
1906 inputrec->opts.tau_t,
1907 inputrec->opts.ref_t);
1909 if (inputrec->eI == eiBD)
1911 do_update_bd_Tconsts(dt, inputrec->bd_fric,
1912 inputrec->opts.ngtc, inputrec->opts.ref_t,
1916 nth = gmx_omp_nthreads_get(emntUpdate);
1918 #pragma omp parallel for num_threads(nth) schedule(static) private(alpha)
1919 for (th = 0; th < nth; th++)
1921 int start_th, end_th;
1923 start_th = start + ((nrend-start)* th )/nth;
1924 end_th = start + ((nrend-start)*(th+1))/nth;
1926 switch (inputrec->eI)
1929 if (ekind->cosacc.cos_accel == 0)
1931 do_update_md(start_th, end_th, dt,
1932 ekind->tcstat, state->nosehoover_vxi,
1933 ekind->bNEMD, ekind->grpstat, inputrec->opts.acc,
1934 inputrec->opts.nFreeze,
1935 md->invmass, md->ptype,
1936 md->cFREEZE, md->cACC, md->cTC,
1937 state->x, xprime, state->v, force, M,
1942 do_update_visc(start_th, end_th, dt,
1943 ekind->tcstat, state->nosehoover_vxi,
1944 md->invmass, md->ptype,
1945 md->cTC, state->x, xprime, state->v, force, M,
1947 ekind->cosacc.cos_accel,
1953 do_update_sd1(upd->sd, upd->sd->gaussrand[th],
1954 start_th, end_th, dt,
1955 inputrec->opts.acc, inputrec->opts.nFreeze,
1956 md->invmass, md->ptype,
1957 md->cFREEZE, md->cACC, md->cTC,
1958 state->x, xprime, state->v, force,
1959 inputrec->opts.ngtc, inputrec->opts.tau_t, inputrec->opts.ref_t);
1962 /* The SD update is done in 2 parts, because an extra constraint step
1965 do_update_sd2(upd->sd, upd->sd->gaussrand[th],
1966 bInitStep, start_th, end_th,
1967 inputrec->opts.acc, inputrec->opts.nFreeze,
1968 md->invmass, md->ptype,
1969 md->cFREEZE, md->cACC, md->cTC,
1970 state->x, xprime, state->v, force, state->sd_X,
1971 inputrec->opts.tau_t,
1975 do_update_bd(start_th, end_th, dt,
1976 inputrec->opts.nFreeze, md->invmass, md->ptype,
1977 md->cFREEZE, md->cTC,
1978 state->x, xprime, state->v, force,
1980 upd->sd->bd_rf, upd->sd->gaussrand[th]);
1984 alpha = 1.0 + DIM/((double)inputrec->opts.nrdf[0]); /* assuming barostat coupled to group 0. */
1989 do_update_vv_vel(start_th, end_th, dt,
1990 inputrec->opts.acc, inputrec->opts.nFreeze,
1991 md->invmass, md->ptype,
1992 md->cFREEZE, md->cACC,
1994 (bNH || bPR), state->veta, alpha);
1997 do_update_vv_pos(start_th, end_th, dt,
1998 inputrec->opts.nFreeze,
1999 md->ptype, md->cFREEZE,
2000 state->x, xprime, state->v,
2001 (bNH || bPR), state->veta);
2006 gmx_fatal(FARGS, "Don't know how to update coordinates");
2014 void correct_ekin(FILE *log, int start, int end, rvec v[], rvec vcm, real mass[],
2015 real tmass, tensor ekin)
2018 * This is a debugging routine. It should not be called for production code
2020 * The kinetic energy should calculated according to:
2021 * Ekin = 1/2 m (v-vcm)^2
2022 * However the correction is not always applied, since vcm may not be
2023 * known in time and we compute
2024 * Ekin' = 1/2 m v^2 instead
2025 * This can be corrected afterwards by computing
2026 * Ekin = Ekin' + 1/2 m ( -2 v vcm + vcm^2)
2028 * Ekin = Ekin' - m v vcm + 1/2 m vcm^2
2035 /* Local particles */
2038 /* Processor dependent part. */
2040 for (i = start; (i < end); i++)
2044 for (j = 0; (j < DIM); j++)
2050 svmul(1/tmass, vcm, vcm);
2051 svmul(0.5, vcm, hvcm);
2053 for (j = 0; (j < DIM); j++)
2055 for (k = 0; (k < DIM); k++)
2057 dekin[j][k] += vcm[k]*(tm*hvcm[j]-mv[j]);
2060 pr_rvecs(log, 0, "dekin", dekin, DIM);
2061 pr_rvecs(log, 0, " ekin", ekin, DIM);
2062 fprintf(log, "dekin = %g, ekin = %g vcm = (%8.4f %8.4f %8.4f)\n",
2063 trace(dekin), trace(ekin), vcm[XX], vcm[YY], vcm[ZZ]);
2064 fprintf(log, "mv = (%8.4f %8.4f %8.4f)\n",
2065 mv[XX], mv[YY], mv[ZZ]);
2068 extern gmx_bool update_randomize_velocities(t_inputrec *ir, gmx_large_int_t step, t_mdatoms *md, t_state *state, gmx_update_t upd, t_idef *idef, gmx_constr_t constr)
2072 real rate = (ir->delta_t)/ir->opts.tau_t[0];
2073 /* proceed with andersen if 1) it's fixed probability per
2074 particle andersen or 2) it's massive andersen and it's tau_t/dt */
2075 if ((ir->etc == etcANDERSEN) || do_per_step(step, (int)(1.0/rate)))
2077 srenew(upd->randatom, state->nalloc);
2078 srenew(upd->randatom_list, state->nalloc);
2079 if (upd->randatom_list_init == FALSE)
2081 for (i = 0; i < state->nalloc; i++)
2083 upd->randatom[i] = FALSE;
2084 upd->randatom_list[i] = 0;
2086 upd->randatom_list_init = TRUE;
2088 andersen_tcoupl(ir, md, state, upd->sd->gaussrand[0], rate,
2089 (ir->etc == etcANDERSEN) ? idef : NULL,
2090 constr ? get_nblocks(constr) : 0,
2091 constr ? get_sblock(constr) : NULL,
2092 upd->randatom, upd->randatom_list,
2093 upd->sd->randomize_group, upd->sd->boltzfac);