/* ---------------------------------------------------------------------- LAMMPS - Large-scale Atomic/Molecular Massively Parallel Simulator http://lammps.sandia.gov, Sandia National Laboratories Steve Plimpton, sjplimp@sandia.gov Copyright (2003) Sandia Corporation. Under the terms of Contract DE-AC04-94AL85000 with Sandia Corporation, the U.S. Government retains certain rights in this software. This software is distributed under the GNU General Public License. See the README file in the top-level LAMMPS directory. ------------------------------------------------------------------------- */ /* ------------------------------------------------------------------------ Contributing authors: Julien Tranchida (SNL) Aidan Thompson (SNL) ------------------------------------------------------------------------- */ #include #include #include #include "atom.h" #include "comm.h" #include "error.h" #include "force.h" #include "neighbor.h" #include "neigh_list.h" #include "neigh_request.h" #include "math_const.h" #include "memory.h" #include "pair_spin.h" #include "update.h" using namespace LAMMPS_NS; using namespace MathConst; /* ---------------------------------------------------------------------- */ PairSpin::PairSpin(LAMMPS *lmp) : Pair(lmp) { hbar = force->hplanck/MY_2PI; newton_pair_spin = 0; // no newton pair for now single_enable = 0; exch_flag = 0; dmi_flag = 0; me_flag = 0; no_virial_fdotr_compute = 1; } /* ---------------------------------------------------------------------- */ PairSpin::~PairSpin() { if (allocated) { memory->destroy(setflag); memory->destroy(cut_spin_exchange); memory->destroy(J1_mag); memory->destroy(J1_mech); memory->destroy(J2); memory->destroy(J3); memory->destroy(cut_spin_dmi); memory->destroy(DM); memory->destroy(v_dmx); memory->destroy(v_dmy); memory->destroy(v_dmz); memory->destroy(cut_spin_me); memory->destroy(ME); memory->destroy(v_mex); memory->destroy(v_mey); memory->destroy(v_mez); memory->destroy(spi); memory->destroy(spj); memory->destroy(fi); memory->destroy(fj); memory->destroy(fmi); memory->destroy(fmj); memory->destroy(del); memory->destroy(cutsq); } } /* ---------------------------------------------------------------------- */ void PairSpin::compute(int eflag, int vflag) { int i,j,ii,jj,inum,jnum,itype,jtype; double evdwl,ecoul; double xtmp,ytmp,ztmp; double fix,fiy,fiz,fjx,fjy,fjz; double fmix,fmiy,fmiz,fmjx,fmjy,fmjz; double cut_ex_2,cut_dmi_2,cut_me_2,cut_spin_pair_global2; double rsq,rd; int *ilist,*jlist,*numneigh,**firstneigh; evdwl = ecoul = 0.0; if (eflag || vflag) ev_setup(eflag,vflag); else evflag = vflag_fdotr = 0; cut_spin_pair_global2 = cut_spin_pair_global*cut_spin_pair_global; double **x = atom->x; double **f = atom->f; double **fm = atom->fm; double *mumag = atom->mumag; double **sp = atom->sp; int *type = atom->type; int nlocal = atom->nlocal; int newton_pair = force->newton_pair; inum = list->inum; ilist = list->ilist; numneigh = list->numneigh; firstneigh = list->firstneigh; // pair spin computations // loop over neighbors of my atoms for (ii = 0; ii < inum; ii++) { i = ilist[ii]; xtmp = x[i][0]; ytmp = x[i][1]; ztmp = x[i][2]; jlist = firstneigh[i]; jnum = numneigh[i]; spi[0] = sp[i][0]; spi[1] = sp[i][1]; spi[2] = sp[i][2]; // loop on neighbors for (jj = 0; jj < jnum; jj++) { j = jlist[jj]; j &= NEIGHMASK; spj[0] = sp[j][0]; spj[1] = sp[j][1]; spj[2] = sp[j][2]; fi[0] = fi[1] = fi[2] = 0.0; fj[0] = fj[1] = fj[2] = 0.0; fmi[0] = fmi[1] = fmi[2] = 0.0; fmj[0] = fmj[1] = fmj[2] = 0.0; del[0] = del[1] = del[2] = 0.0; del[0] = x[j][0] - xtmp; del[1] = x[j][1] - ytmp; del[2] = x[j][2] - ztmp; // square of inter-atomic distance rsq = del[0]*del[0] + del[1]*del[1] + del[2]*del[2]; itype = type[i]; jtype = type[j]; // exchange interaction if (exch_flag) { cut_ex_2 = cut_spin_exchange[itype][jtype]*cut_spin_exchange[itype][jtype]; if (rsq <= cut_ex_2) { compute_exchange(i,j,rsq,fmi,fmj,spi,spj); compute_exchange_mech(i,j,rsq,del,fi,fj,spi,spj); } } // dm interaction if (dmi_flag){ cut_dmi_2 = cut_spin_dmi[itype][jtype]*cut_spin_dmi[itype][jtype]; if (rsq <= cut_dmi_2){ compute_dmi(i,j,fmi,fmj,spi,spj); } } // me interaction if (me_flag){ cut_me_2 = cut_spin_me[itype][jtype]*cut_spin_me[itype][jtype]; if (rsq <= cut_me_2){ compute_me(i,j,fmi,fmj,spi,spj); } } f[i][0] += fi[0]; f[i][1] += fi[1]; f[i][2] += fi[2]; fm[i][0] += fmi[0]; fm[i][1] += fmi[1]; fm[i][2] += fmi[2]; /* if (newton_pair || j < nlocal) {*/ if (newton_pair_spin) { f[j][0] += fj[0]; f[j][1] += fj[1]; f[j][2] += fj[2]; fm[j][0] += fmj[0]; fm[j][1] += fmj[1]; fm[j][2] += fmj[2]; } if (eflag) { if (rsq <= cut_spin_pair_global2) { evdwl -= mumag[i]*spi[0]*fmi[0]; evdwl -= mumag[i]*spi[1]*fmi[1]; evdwl -= mumag[i]*spi[2]*fmi[2]; } } evdwl *= hbar; if (evflag) ev_tally_xyz(i,j,nlocal,newton_pair, evdwl,ecoul,fi[0],fi[1],fi[2],del[0],del[1],del[2]); } } if (vflag_fdotr) virial_fdotr_compute(); } /* ---------------------------------------------------------------------- */ void PairSpin::compute_exchange(int i, int j, double rsq, double *fmi, double *fmj, double *spi, double *spj) { int *type = atom->type; int itype, jtype; double Jex, ra; itype = type[i]; jtype = type[j]; ra = rsq/J3[itype][jtype]/J3[itype][jtype]; Jex = 4.0*J1_mag[itype][jtype]*ra; Jex *= (1.0-J2[itype][jtype]*ra); Jex *= exp(-ra); fmi[0] += Jex*spj[0]; fmi[1] += Jex*spj[1]; fmi[2] += Jex*spj[2]; fmj[0] += Jex*spi[0]; fmj[1] += Jex*spi[1]; fmj[2] += Jex*spi[2]; } /* ---------------------------------------------------------------------- */ void PairSpin::compute_exchange_mech(int i, int j, double rsq, double *del, double *fi, double *fj, double *spi, double *spj) { int *type = atom->type; int itype, jtype; double Jex, Jex_mech, ra, rr; itype = type[i]; jtype = type[j]; Jex = J1_mech[itype][jtype]; ra = rsq/J3[itype][jtype]/J3[itype][jtype]; rr = sqrt(rsq)/J3[itype][jtype]/J3[itype][jtype]; Jex_mech = 1.0-ra-J2[itype][jtype]*ra*(2.0-ra); Jex_mech *= 8.0*Jex*rr*exp(-ra); Jex_mech *= (spi[0]*spj[0]+spi[1]*spj[1]+spi[2]*spj[2]); fi[0] += Jex_mech*del[0]; fi[1] += Jex_mech*del[1]; fi[2] += Jex_mech*del[2]; fj[0] -= Jex_mech*del[0]; fj[1] -= Jex_mech*del[1]; fj[2] -= Jex_mech*del[2]; } /* ---------------------------------------------------------------------- */ void PairSpin::compute_dmi(int i, int j, double *fmi, double *fmj, double *spi, double *spj) { int *type = atom->type; int itype, jtype; double dmix,dmiy,dmiz; itype = type[i]; jtype = type[j]; dmix = DM[itype][jtype]*v_dmx[itype][jtype]; dmiy = DM[itype][jtype]*v_dmy[itype][jtype]; dmiz = DM[itype][jtype]*v_dmz[itype][jtype]; fmi[0] += spj[1]*dmiz-spj[2]*dmiy; fmi[1] += spj[2]*dmix-spj[0]*dmiz; fmi[2] += spj[0]*dmiy-spj[1]*dmix; fmj[0] -= spi[1]*dmiz-spi[2]*dmiy; fmj[1] -= spi[2]*dmix-spi[0]*dmiz; fmj[2] -= spi[0]*dmiy-spi[1]*dmix; } /* ---------------------------------------------------------------------- */ void PairSpin::compute_me(int i, int j, double *fmi, double *fmj, double *spi, double *spj) { int *type = atom->type; int itype, jtype; itype = type[i]; jtype = type[j]; double **sp = atom->sp; double **x = atom->x; double meix,meiy,meiz; double rx, ry, rz, inorm; rx = x[j][0] - x[i][0]; ry = x[j][1] - x[i][1]; rz = x[j][2] - x[i][2]; inorm = 1.0/sqrt(rx*rx+ry*ry+rz*rz); rx *= inorm; ry *= inorm; rz *= inorm; meix = v_mey[itype][jtype]*rz - v_mez[itype][jtype]*ry; meiy = v_mez[itype][jtype]*rx - v_mex[itype][jtype]*rz; meiz = v_mex[itype][jtype]*ry - v_mey[itype][jtype]*rx; meix *= ME[itype][jtype]; meiy *= ME[itype][jtype]; meiz *= ME[itype][jtype]; fmi[0] += spj[1]*meiz - spj[2]*meiy; fmi[1] += spj[2]*meix - spj[0]*meiz; fmi[2] += spj[0]*meiy - spj[1]*meix; fmj[0] -= spi[1]*meiz - spi[2]*meiy; fmj[1] -= spi[2]*meix - spi[0]*meiz; fmj[2] -= spi[0]*meiy - spi[1]*meix; } /* ---------------------------------------------------------------------- allocate all arrays ------------------------------------------------------------------------- */ void PairSpin::allocate() { allocated = 1; int n = atom->ntypes; memory->create(setflag,n+1,n+1,"pair:setflag"); for (int i = 1; i <= n; i++) for (int j = i; j <= n; j++) setflag[i][j] = 0; memory->create(cut_spin_exchange,n+1,n+1,"pair:cut_spin_exchange"); memory->create(J1_mag,n+1,n+1,"pair:J1_mag"); memory->create(J1_mech,n+1,n+1,"pair:J1_mech"); memory->create(J2,n+1,n+1,"pair:J2"); memory->create(J3,n+1,n+1,"pair:J3"); memory->create(cut_spin_dmi,n+1,n+1,"pair:cut_spin_dmi"); memory->create(DM,n+1,n+1,"pair:DM"); memory->create(v_dmx,n+1,n+1,"pair:DM_vector_x"); memory->create(v_dmy,n+1,n+1,"pair:DM_vector_y"); memory->create(v_dmz,n+1,n+1,"pair:DM_vector_z"); memory->create(cut_spin_me,n+1,n+1,"pair:cut_spin_me"); memory->create(ME,n+1,n+1,"pair:ME"); memory->create(v_mex,n+1,n+1,"pair:ME_vector_x"); memory->create(v_mey,n+1,n+1,"pair:ME_vector_y"); memory->create(v_mez,n+1,n+1,"pair:ME_vector_z"); memory->create(spi,3,"pair:spi"); memory->create(spj,3,"pair:spj"); memory->create(fi,3,"pair:fi"); memory->create(fj,3,"pair:fj"); memory->create(fmi,3,"pair:fmi"); memory->create(fmj,3,"pair:fmj"); memory->create(del,3,"pair:del"); memory->create(cutsq,n+1,n+1,"pair:cutsq"); } /* ---------------------------------------------------------------------- global settings ------------------------------------------------------------------------- */ void PairSpin::settings(int narg, char **arg) { if (narg < 1 || narg > 2) error->all(FLERR,"Incorrect number of args in pair_style pair/spin command"); if (strcmp(update->unit_style,"metal") != 0) error->all(FLERR,"Spin simulations require metal unit style"); cut_spin_pair_global = force->numeric(FLERR,arg[0]); // reset cutoffs that have been explicitly set if (allocated) { int i,j; for (i = 1; i <= atom->ntypes; i++) for (j = i+1; j <= atom->ntypes; j++) if (setflag[i][j]) { cut_spin_exchange[i][j] = cut_spin_pair_global; cut_spin_dmi[i][j] = cut_spin_pair_global; cut_spin_me[i][j] = cut_spin_pair_global; } } } /* ---------------------------------------------------------------------- set coeffs for one or more type spin pairs (only one for now) ------------------------------------------------------------------------- */ void PairSpin::coeff(int narg, char **arg) { const double hbar = force->hplanck/MY_2PI; if (!allocated) allocate(); if (strcmp(arg[2],"exchange")==0){ if (narg != 7) error->all(FLERR,"Incorrect args in pair_style command"); exch_flag = 1; int ilo,ihi,jlo,jhi; force->bounds(FLERR,arg[0],atom->ntypes,ilo,ihi); force->bounds(FLERR,arg[1],atom->ntypes,jlo,jhi); const double rij = force->numeric(FLERR,arg[3]); const double j1 = (force->numeric(FLERR,arg[4])); const double j2 = force->numeric(FLERR,arg[5]); const double j3 = force->numeric(FLERR,arg[6]); int count = 0; for (int i = ilo; i <= ihi; i++) { for (int j = MAX(jlo,i); j <= jhi; j++) { cut_spin_exchange[i][j] = rij; J1_mag[i][j] = j1/hbar; J1_mech[i][j] = j1; J2[i][j] = j2; J3[i][j] = j3; setflag[i][j] = 1; count++; } } if (count == 0) error->all(FLERR,"Incorrect args in pair_style command"); } else if (strcmp(arg[2],"dmi")==0) { if (narg != 8) error->all(FLERR,"Incorrect args in pair_style command"); dmi_flag = 1; int ilo,ihi,jlo,jhi; force->bounds(FLERR,arg[0],atom->ntypes,ilo,ihi); force->bounds(FLERR,arg[1],atom->ntypes,jlo,jhi); const double rij = force->numeric(FLERR,arg[3]); const double dm = (force->numeric(FLERR,arg[4]))/hbar; double dmx = force->numeric(FLERR,arg[5]); double dmy = force->numeric(FLERR,arg[6]); double dmz = force->numeric(FLERR,arg[7]); double inorm = 1.0/(dmx*dmx+dmy*dmy+dmz*dmz); dmx *= inorm; dmy *= inorm; dmz *= inorm; int count = 0; for (int i = ilo; i <= ihi; i++) { for (int j = MAX(jlo,i); j <= jhi; j++) { cut_spin_dmi[i][j] = rij; DM[i][j] = dm; v_dmx[i][j] = dmx; v_dmy[i][j] = dmy; v_dmz[i][j] = dmz; setflag[i][j] = 1; count++; } } if (count == 0) error->all(FLERR,"Incorrect args in pair_style command"); } else if (strcmp(arg[2],"me")==0) { if (narg != 8) error->all(FLERR,"Incorrect args in pair_style command"); me_flag = 1; int ilo,ihi,jlo,jhi; force->bounds(FLERR,arg[0],atom->ntypes,ilo,ihi); force->bounds(FLERR,arg[1],atom->ntypes,jlo,jhi); const double rij = force->numeric(FLERR,arg[3]); const double me = (force->numeric(FLERR,arg[4]))/hbar; double mex = force->numeric(FLERR,arg[5]); double mey = force->numeric(FLERR,arg[6]); double mez = force->numeric(FLERR,arg[7]); double inorm = 1.0/(mex*mex+mey*mey+mez*mez); mex *= inorm; mey *= inorm; mez *= inorm; int count = 0; for (int i = ilo; i <= ihi; i++) { for (int j = MAX(jlo,i); j <= jhi; j++) { cut_spin_me[i][j] = rij; DM[i][j] = me; v_mex[i][j] = mex; v_mey[i][j] = mey; v_mez[i][j] = mez; setflag[i][j] = 1; count++; } } if (count == 0) error->all(FLERR,"Incorrect args in pair_style command"); } else error->all(FLERR,"Incorrect args in pair_style command"); // check if Jex [][] still works for Ferrimagnetic exchange } /* ---------------------------------------------------------------------- init specific to this pair style ------------------------------------------------------------------------- */ void PairSpin::init_style() { if (!atom->sp_flag || !atom->mumag_flag) error->all(FLERR,"Pair spin requires atom attributes sp, mumag"); neighbor->request(this,instance_me); #define FULLNEI #if defined FULLNEI int irequest = neighbor->request(this,instance_me); neighbor->requests[irequest]->half = 0; neighbor->requests[irequest]->full = 1; #endif } /* ---------------------------------------------------------------------- init for one type pair i,j and corresponding j,i ------------------------------------------------------------------------- */ double PairSpin::init_one(int i, int j) { if (setflag[i][j] == 0) error->all(FLERR,"All pair coeffs are not set"); return cut_spin_pair_global; } /* ---------------------------------------------------------------------- proc 0 writes to restart file ------------------------------------------------------------------------- */ void PairSpin::write_restart(FILE *fp) { write_restart_settings(fp); int i,j; for (i = 1; i <= atom->ntypes; i++) for (j = i; j <= atom->ntypes; j++) { fwrite(&setflag[i][j],sizeof(int),1,fp); if (setflag[i][j]) { if (exch_flag){ fwrite(&J1_mag[i][j],sizeof(double),1,fp); fwrite(&J1_mech[i][j],sizeof(double),1,fp); fwrite(&J2[i][j],sizeof(double),1,fp); fwrite(&J3[i][j],sizeof(double),1,fp); fwrite(&cut_spin_exchange[i][j],sizeof(double),1,fp); } if (dmi_flag) { fwrite(&DM[i][j],sizeof(double),1,fp); fwrite(&v_dmx[i][j],sizeof(double),1,fp); fwrite(&v_dmy[i][j],sizeof(double),1,fp); fwrite(&v_dmz[i][j],sizeof(double),1,fp); fwrite(&cut_spin_dmi[i][j],sizeof(double),1,fp); } if (me_flag) { fwrite(&ME[i][j],sizeof(double),1,fp); fwrite(&v_mex[i][j],sizeof(double),1,fp); fwrite(&v_mey[i][j],sizeof(double),1,fp); fwrite(&v_mez[i][j],sizeof(double),1,fp); fwrite(&cut_spin_me[i][j],sizeof(double),1,fp); } } } } /* ---------------------------------------------------------------------- proc 0 reads from restart file, bcasts ------------------------------------------------------------------------- */ void PairSpin::read_restart(FILE *fp) { read_restart_settings(fp); allocate(); int i,j; int me = comm->me; for (i = 1; i <= atom->ntypes; i++) { for (j = i; j <= atom->ntypes; j++) { if (me == 0) fread(&setflag[i][j],sizeof(int),1,fp); MPI_Bcast(&setflag[i][j],1,MPI_INT,0,world); if (setflag[i][j]) { if (me == 0) { fread(&J1_mag[i][j],sizeof(double),1,fp); fread(&J1_mech[i][j],sizeof(double),1,fp); fread(&J2[i][j],sizeof(double),1,fp); fread(&J2[i][j],sizeof(double),1,fp); fread(&cut_spin_exchange[i][j],sizeof(double),1,fp); } MPI_Bcast(&J1_mag[i][j],1,MPI_DOUBLE,0,world); MPI_Bcast(&J1_mech[i][j],1,MPI_DOUBLE,0,world); MPI_Bcast(&J2[i][j],1,MPI_DOUBLE,0,world); MPI_Bcast(&J3[i][j],1,MPI_DOUBLE,0,world); MPI_Bcast(&cut_spin_exchange[i][j],1,MPI_DOUBLE,0,world); } } } } /* ---------------------------------------------------------------------- proc 0 writes to restart file ------------------------------------------------------------------------- */ void PairSpin::write_restart_settings(FILE *fp) { fwrite(&cut_spin_pair_global,sizeof(double),1,fp); fwrite(&offset_flag,sizeof(int),1,fp); fwrite(&mix_flag,sizeof(int),1,fp); } /* ---------------------------------------------------------------------- proc 0 reads from restart file, bcasts ------------------------------------------------------------------------- */ void PairSpin::read_restart_settings(FILE *fp) { if (comm->me == 0) { fread(&cut_spin_pair_global,sizeof(double),1,fp); fread(&offset_flag,sizeof(int),1,fp); fread(&mix_flag,sizeof(int),1,fp); } MPI_Bcast(&cut_spin_pair_global,1,MPI_DOUBLE,0,world); MPI_Bcast(&offset_flag,1,MPI_INT,0,world); MPI_Bcast(&mix_flag,1,MPI_INT,0,world); }