git-svn-id: svn://svn.icms.temple.edu/lammps-ro/trunk@7148 f3b2605a-c512-4ea7-a41b-209d697bcdaa

This commit is contained in:
sjplimp 2011-10-20 15:01:42 +00:00
parent c6d228775f
commit 89a39c7425
5 changed files with 1211 additions and 0 deletions

View File

@ -12,7 +12,9 @@ if (test $1 = 1) then
cp fix_nve_tri.cpp ..
cp fix_nvt_asphere.cpp ..
cp pair_gayberne.cpp ..
cp pair_line.cpp ..
cp pair_resquared.cpp ..
cp pair_tri.cpp ..
cp compute_erotate_asphere.h ..
cp compute_temp_asphere.h ..
@ -24,7 +26,9 @@ if (test $1 = 1) then
cp fix_nve_tri.h ..
cp fix_nvt_asphere.h ..
cp pair_gayberne.h ..
cp pair_line.h ..
cp pair_resquared.h ..
cp pair_tri.h ..
elif (test $1 = 0) then
@ -38,7 +42,9 @@ elif (test $1 = 0) then
rm -f ../fix_nve_tri.cpp
rm -f ../fix_nvt_asphere.cpp
rm -f ../pair_gayberne.cpp
rm -f ../pair_line.cpp
rm -f ../pair_resquared.cpp
rm -f ../pair_tri.cpp
rm -f ../compute_erotate_asphere.h
rm -f ../compute_temp_asphere.h
@ -50,6 +56,8 @@ elif (test $1 = 0) then
rm -f ../fix_nve_tri.h
rm -f ../fix_nvt_asphere.h
rm -f ../pair_gayberne.h
rm -f ../pair_line.h
rm -f ../pair_resquared.h
rm -f ../pair_tri.h
fi

444
src/ASPHERE/pair_line.cpp Normal file
View File

@ -0,0 +1,444 @@
/* ----------------------------------------------------------------------
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.
------------------------------------------------------------------------- */
#include "math.h"
#include "stdio.h"
#include "stdlib.h"
#include "string.h"
#include "pair_line.h"
#include "atom.h"
#include "atom_vec_line.h"
#include "force.h"
#include "neighbor.h"
#include "neigh_list.h"
#include "memory.h"
#include "error.h"
using namespace LAMMPS_NS;
#define DELTA 10000
/* ---------------------------------------------------------------------- */
PairLine::PairLine(LAMMPS *lmp) : Pair(lmp)
{
avec = (AtomVecLine *) atom->style_match("line");
if (!avec) error->all(FLERR,"Pair line requires atom style line");
dmax = nmax = 0;
discrete = NULL;
dnum = dfirst = NULL;
single_enable = 0;
}
/* ---------------------------------------------------------------------- */
PairLine::~PairLine()
{
memory->sfree(discrete);
memory->destroy(dnum);
memory->destroy(dfirst);
if (allocated) {
memory->destroy(setflag);
memory->destroy(cutsq);
memory->destroy(cut);
memory->destroy(epsilon);
memory->destroy(sigma);
memory->destroy(lj1);
memory->destroy(lj2);
memory->destroy(lj3);
memory->destroy(lj4);
}
}
/* ---------------------------------------------------------------------- */
void PairLine::compute(int eflag, int vflag)
{
int i,j,ii,jj,inum,jnum,itype,jtype;
int ni,nj,npi,npj,ifirst,jfirst;
double xtmp,ytmp,ztmp,delx,dely,delz,evdwl,fpair;
double rsq,r2inv,r6inv,term1,term2,sig,sig3,forcelj;
double xi[2],xj[2],fi[2],fj[2],dxi,dxj,dyi,dyj,ti,tj;
int *ilist,*jlist,*numneigh,**firstneigh;
evdwl = 0.0;
if (eflag || vflag) ev_setup(eflag,vflag);
else evflag = vflag_fdotr = 0;
double **x = atom->x;
double **f = atom->f;
double **torque = atom->torque;
int *line = atom->line;
int *type = atom->type;
int nlocal = atom->nlocal;
int nall = nlocal + atom->nghost;
int newton_pair = force->newton_pair;
inum = list->inum;
ilist = list->ilist;
numneigh = list->numneigh;
firstneigh = list->firstneigh;
// grow discrete list if necessary and initialize
if (nall > nmax) {
memory->destroy(dnum);
memory->destroy(dfirst);
memory->create(dnum,nall,"pair:dnum");
memory->create(dfirst,nall,"pair:dfirst");
}
for (i = 0; i < nall; i++) dnum[i] = 0;
ndiscrete = 0;
// 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];
itype = type[i];
jlist = firstneigh[i];
jnum = numneigh[i];
for (jj = 0; jj < jnum; jj++) {
j = jlist[jj];
j &= NEIGHMASK;
delx = xtmp - x[j][0];
dely = ytmp - x[j][1];
delz = ztmp - x[j][2];
rsq = delx*delx + dely*dely + delz*delz;
jtype = type[j];
if (rsq >= cutsq[itype][jtype]) continue;
// line/line interactions = NxN particles
evdwl = 0.0;
if (line[i] >= 0 && line[j] >= 0) {
if (dnum[i] == 0) discretize(i,sigma[itype][itype]);
npi = dnum[i];
ifirst = dfirst[i];
if (dnum[j] == 0) discretize(j,sigma[jtype][jtype]);
npj = dnum[j];
jfirst = dfirst[j];
for (ni = 0; ni < npi; ni++) {
dxi = discrete[ifirst+ni].dx;
dyi = discrete[ifirst+ni].dy;
for (nj = 0; nj < npj; nj++) {
dxj = discrete[jfirst+nj].dx;
dyj = discrete[jfirst+nj].dy;
xi[0] = x[i][0] + dxi;
xi[1] = x[i][1] + dyi;
xj[0] = x[j][0] + dxj;
xj[1] = x[j][1] + dyj;
delx = xi[0] - xj[0];
dely = xi[1] - xj[1];
rsq = delx*delx + dely*dely;
sig = 0.5 * (discrete[ifirst+ni].sigma+discrete[jfirst+nj].sigma);
sig3 = sig*sig*sig;
term2 = 24.0*epsilon[itype][jtype] * sig3*sig3;
term1 = 2.0 * term2 * sig3*sig3;
r2inv = 1.0/rsq;
r6inv = r2inv*r2inv*r2inv;
forcelj = r6inv * (term1*r6inv - term2);
fpair = forcelj*r2inv;
if (eflag) evdwl += r6inv*(term1/12.0*r6inv-term2/6.0);
fi[0] = delx*fpair;
fi[1] = dely*fpair;
f[i][0] += fi[0];
f[i][1] += fi[1];
torque[i][2] += dxi*fi[1] - dyi*fi[0];
if (newton_pair || j < nlocal) {
fj[0] = -delx*fpair;
fj[1] = -dely*fpair;
f[j][0] += fj[0];
f[j][1] += fj[1];
torque[j][2] += dxj*fj[1] - dyj*fj[0];
}
}
}
// line/particle interaction = Nx1 particles
// convert line into Np particles based on sigma and line length
} else if (line[i] >= 0) {
if (dnum[i] == 0) discretize(i,sigma[itype][itype]);
npi = dnum[i];
ifirst = dfirst[i];
for (ni = 0; ni < npi; ni++) {
dxi = discrete[ifirst+ni].dx;
dyi = discrete[ifirst+ni].dy;
xi[0] = x[i][0] + dxi;
xi[1] = x[i][1] + dyi;
xj[0] = x[j][0];
xj[1] = x[j][1];
delx = xi[0] - xj[0];
dely = xi[1] - xj[1];
rsq = delx*delx + dely*dely;
sig = 0.5 * (discrete[ifirst+ni].sigma+sigma[jtype][jtype]);
sig3 = sig*sig*sig;
term2 = 24.0*epsilon[itype][jtype] * sig3*sig3;
term1 = 2.0 * term2 * sig3*sig3;
r2inv = 1.0/rsq;
r6inv = r2inv*r2inv*r2inv;
forcelj = r6inv * (term1*r6inv - term2);
fpair = forcelj*r2inv;
if (eflag) evdwl += r6inv*(term1/12.0*r6inv-term2/6.0);
fi[0] = delx*fpair;
fi[1] = dely*fpair;
f[i][0] += fi[0];
f[i][1] += fi[1];
torque[i][2] += dxi*fi[1] - dyi*fi[0];
if (newton_pair || j < nlocal) {
fj[0] = -delx*fpair;
fj[1] = -dely*fpair;
f[j][0] += fj[0];
f[j][1] += fj[1];
}
}
// particle/line interaction = Nx1 particles
// convert line into Np particles based on sigma and line length
} else if (line[j] >= 0) {
if (dnum[j] == 0) discretize(j,sigma[jtype][jtype]);
npj = dnum[j];
jfirst = dfirst[j];
for (nj = 0; nj < npj; nj++) {
dxj = discrete[jfirst+nj].dx;
dyj = discrete[jfirst+nj].dy;
xi[0] = x[i][0];
xi[1] = x[i][1];
xj[0] = x[j][0] + dxj;
xj[1] = x[j][1] + dyj;
delx = xi[0] - xj[0];
dely = xi[1] - xj[1];
rsq = delx*delx + dely*dely;
sig = 0.5 * (sigma[itype][itype]+discrete[jfirst+nj].sigma);
sig3 = sig*sig*sig;
term2 = 24.0*epsilon[itype][jtype] * sig3*sig3;
term1 = 2.0 * term2 * sig3*sig3;
r2inv = 1.0/rsq;
r6inv = r2inv*r2inv*r2inv;
forcelj = r6inv * (term1*r6inv - term2);
fpair = forcelj*r2inv;
if (eflag) evdwl += r6inv*(term1/12.0*r6inv-term2/6.0);
fi[0] = delx*fpair;
fi[1] = dely*fpair;
f[i][0] += fi[0];
f[i][1] += fi[1];
if (newton_pair || j < nlocal) {
f[j][0] += fj[0];
f[j][1] += fj[1];
fj[0] = -delx*fpair;
fj[1] = -dely*fpair;
torque[j][2] += dxj*fj[1] - dyj*fj[0];
}
}
// particle/particle interaction = 1x1 particles
} else {
r2inv = 1.0/rsq;
r6inv = r2inv*r2inv*r2inv;
forcelj = r6inv * (lj1[itype][jtype]*r6inv - lj2[itype][jtype]);
fpair = forcelj*r2inv;
if (eflag)
evdwl += r6inv*(lj3[itype][jtype]*r6inv-lj4[itype][jtype]);
f[i][0] += delx*fpair;
f[i][1] += dely*fpair;
f[i][2] += delz*fpair;
if (newton_pair || j < nlocal) {
f[j][0] -= delx*fpair;
f[j][1] -= dely*fpair;
f[j][2] -= delz*fpair;
}
}
if (evflag) ev_tally(i,j,nlocal,newton_pair,
evdwl,0.0,fpair,delx,dely,delz);
}
}
if (vflag_fdotr) virial_fdotr_compute();
}
/* ----------------------------------------------------------------------
allocate all arrays
------------------------------------------------------------------------- */
void PairLine::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(cutsq,n+1,n+1,"pair:cutsq");
memory->create(cut,n+1,n+1,"pair:cut");
memory->create(epsilon,n+1,n+1,"pair:epsilon");
memory->create(sigma,n+1,n+1,"pair:sigma");
memory->create(lj1,n+1,n+1,"pair:lj1");
memory->create(lj2,n+1,n+1,"pair:lj2");
memory->create(lj3,n+1,n+1,"pair:lj3");
memory->create(lj4,n+1,n+1,"pair:lj4");
}
/* ----------------------------------------------------------------------
global settings
------------------------------------------------------------------------- */
void PairLine::settings(int narg, char **arg)
{
if (narg != 1) error->all(FLERR,"Illegal pair_style command");
cut_global = force->numeric(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[i][j] = cut_global;
}
}
/* ----------------------------------------------------------------------
set coeffs for one or more type pairs
------------------------------------------------------------------------- */
void PairLine::coeff(int narg, char **arg)
{
if (narg < 4 || narg > 5)
error->all(FLERR,"Incorrect args for pair coefficients");
if (!allocated) allocate();
int ilo,ihi,jlo,jhi;
force->bounds(arg[0],atom->ntypes,ilo,ihi);
force->bounds(arg[1],atom->ntypes,jlo,jhi);
double epsilon_one = force->numeric(arg[2]);
double sigma_one = force->numeric(arg[3]);
double cut_one = cut_global;
if (narg == 5) cut_one = force->numeric(arg[4]);
int count = 0;
for (int i = ilo; i <= ihi; i++) {
for (int j = MAX(jlo,i); j <= jhi; j++) {
epsilon[i][j] = epsilon_one;
sigma[i][j] = sigma_one;
cut[i][j] = cut_one;
setflag[i][j] = 1;
count++;
}
}
if (count == 0) error->all(FLERR,"Incorrect args for pair coefficients");
}
/* ----------------------------------------------------------------------
init for one type pair i,j and corresponding j,i
------------------------------------------------------------------------- */
double PairLine::init_one(int i, int j)
{
if (setflag[i][j] == 0) {
epsilon[i][j] = mix_energy(epsilon[i][i],epsilon[j][j],
sigma[i][i],sigma[j][j]);
sigma[i][j] = mix_distance(sigma[i][i],sigma[j][j]);
cut[i][j] = mix_distance(cut[i][i],cut[j][j]);
}
lj1[i][j] = 48.0 * epsilon[i][j] * pow(sigma[i][j],12.0);
lj2[i][j] = 24.0 * epsilon[i][j] * pow(sigma[i][j],6.0);
lj3[i][j] = 4.0 * epsilon[i][j] * pow(sigma[i][j],12.0);
lj4[i][j] = 4.0 * epsilon[i][j] * pow(sigma[i][j],6.0);
epsilon[j][i] = epsilon[i][j];
sigma[j][i] = sigma[i][j];
lj1[j][i] = lj1[i][j];
lj2[j][i] = lj2[i][j];
lj3[j][i] = lj3[i][j];
lj4[j][i] = lj4[i][j];
return cut[i][j];
}
/* ----------------------------------------------------------------------
discretize line segment I into N sub-segments no more than sigma in length
store new discrete particles in Discrete list
------------------------------------------------------------------------- */
void PairLine::discretize(int i, double sigma)
{
AtomVecLine::Bonus *bonus = avec->bonus;
double length = bonus[atom->line[i]].length;
double theta = bonus[atom->line[i]].theta;
int n = static_cast<int> (length/sigma) + 1;
dnum[i] = n;
dfirst[i] = ndiscrete;
if (ndiscrete + n > dmax) {
dmax += DELTA;
discrete = (Discrete *)
memory->srealloc(discrete,dmax*sizeof(Discrete),"pair:discrete");
}
double *x = atom->x[i];
sigma = length/n;
double delta;
for (int m = 0; m < n; m++) {
delta = -0.5 + (2*m+1)/(2.0*n);
discrete[ndiscrete].dx = delta*length*cos(theta);
discrete[ndiscrete].dy = delta*length*sin(theta);
discrete[ndiscrete].sigma = sigma;
ndiscrete++;
}
}

61
src/ASPHERE/pair_line.h Normal file
View File

@ -0,0 +1,61 @@
/* ----------------------------------------------------------------------
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.
------------------------------------------------------------------------- */
#ifdef PAIR_CLASS
PairStyle(line,PairLine)
#else
#ifndef LMP_PAIR_LINE_H
#define LMP_PAIR_LINE_H
#include "pair.h"
namespace LAMMPS_NS {
class PairLine : public Pair {
public:
PairLine(class LAMMPS *);
~PairLine();
void compute(int, int);
void settings(int, char **);
void coeff(int, char **);
double init_one(int, int);
protected:
double cut_global;
double **cut;
double **epsilon,**sigma;
double **lj1,**lj2,**lj3,**lj4;
class AtomVecLine *avec;
struct Discrete {
double dx,dy;
double sigma;
};
Discrete *discrete; // list of all discretes for all lines
int ndiscrete; // number of discretes in list
int dmax; // allocated size of discrete list
int *dnum; // number of discretes per line, 0 if uninit
int *dfirst; // index of first discrete per each line
int nmax; // allocated size of dnum,dfirst vectors
void allocate();
void discretize(int, double);
};
}
#endif
#endif

637
src/ASPHERE/pair_tri.cpp Normal file
View File

@ -0,0 +1,637 @@
/* ----------------------------------------------------------------------
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.
------------------------------------------------------------------------- */
#include "math.h"
#include "stdio.h"
#include "stdlib.h"
#include "string.h"
#include "pair_tri.h"
#include "math_extra.h"
#include "atom.h"
#include "atom_vec_tri.h"
#include "force.h"
#include "neighbor.h"
#include "neigh_list.h"
#include "memory.h"
#include "error.h"
using namespace LAMMPS_NS;
#define DELTA 20
/* ---------------------------------------------------------------------- */
PairTri::PairTri(LAMMPS *lmp) : Pair(lmp)
{
avec = (AtomVecTri *) atom->style_match("tri");
if (!avec) error->all(FLERR,"Pair tri requires atom style tri");
dmax = nmax = 0;
discrete = NULL;
dnum = dfirst = NULL;
single_enable = 0;
}
/* ---------------------------------------------------------------------- */
PairTri::~PairTri()
{
memory->sfree(discrete);
memory->destroy(dnum);
memory->destroy(dfirst);
if (allocated) {
memory->destroy(setflag);
memory->destroy(cutsq);
memory->destroy(cut);
memory->destroy(epsilon);
memory->destroy(sigma);
memory->destroy(lj1);
memory->destroy(lj2);
memory->destroy(lj3);
memory->destroy(lj4);
}
}
/* ---------------------------------------------------------------------- */
void PairTri::compute(int eflag, int vflag)
{
int i,j,ii,jj,inum,jnum,itype,jtype;
int ni,nj,npi,npj,ifirst,jfirst;
double xtmp,ytmp,ztmp,delx,dely,delz,evdwl,fpair;
double rsq,r2inv,r6inv,term1,term2,sig,sig3,forcelj;
double dxi,dxj,dyi,dyj,dzi,dzj;
double xi[3],xj[3],fi[3],fj[3],ti[3],tj[3],p[3][3];
double dc1[3],dc2[3],dc3[3];
int *ilist,*jlist,*numneigh,**firstneigh;
evdwl = 0.0;
if (eflag || vflag) ev_setup(eflag,vflag);
else evflag = vflag_fdotr = 0;
AtomVecTri::Bonus *bonus = avec->bonus;
double **x = atom->x;
double **f = atom->f;
double **torque = atom->torque;
int *tri = atom->tri;
int *type = atom->type;
int nlocal = atom->nlocal;
int nall = nlocal + atom->nghost;
int newton_pair = force->newton_pair;
inum = list->inum;
ilist = list->ilist;
numneigh = list->numneigh;
firstneigh = list->firstneigh;
// grow discrete list if necessary and initialize
if (nall > nmax) {
memory->destroy(dnum);
memory->destroy(dfirst);
memory->create(dnum,nall,"pair:dnum");
memory->create(dfirst,nall,"pair:dfirst");
}
for (i = 0; i < nall; i++) dnum[i] = 0;
ndiscrete = 0;
// 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];
itype = type[i];
jlist = firstneigh[i];
jnum = numneigh[i];
for (jj = 0; jj < jnum; jj++) {
j = jlist[jj];
j &= NEIGHMASK;
delx = xtmp - x[j][0];
dely = ytmp - x[j][1];
delz = ztmp - x[j][2];
rsq = delx*delx + dely*dely + delz*delz;
jtype = type[j];
if (rsq >= cutsq[itype][jtype]) continue;
// tri/tri interactions = NxN particles
// c1,c2,c3 = corner pts of triangle I or J
evdwl = 0.0;
if (tri[i] >= 0 && tri[j] >= 0) {
if (dnum[i] == 0) {
MathExtra::quat_to_mat(bonus[tri[i]].quat,p);
MathExtra::matvec(p,bonus[tri[i]].c1,dc1);
MathExtra::matvec(p,bonus[tri[i]].c2,dc2);
MathExtra::matvec(p,bonus[tri[i]].c3,dc3);
dfirst[i] = ndiscrete;
discretize(i,sigma[itype][itype],dc1,dc2,dc3);
dnum[i] = ndiscrete - dfirst[i];
}
npi = dnum[i];
ifirst = dfirst[i];
if (dnum[j] == 0) {
MathExtra::quat_to_mat(bonus[tri[j]].quat,p);
MathExtra::matvec(p,bonus[tri[j]].c1,dc1);
MathExtra::matvec(p,bonus[tri[j]].c2,dc2);
MathExtra::matvec(p,bonus[tri[j]].c3,dc3);
dfirst[j] = ndiscrete;
discretize(j,sigma[jtype][jtype],dc1,dc2,dc3);
dnum[j] = ndiscrete - dfirst[j];
}
npj = dnum[j];
jfirst = dfirst[j];
for (ni = 0; ni < npi; ni++) {
dxi = discrete[ifirst+ni].dx;
dyi = discrete[ifirst+ni].dy;
dzi = discrete[ifirst+ni].dz;
for (nj = 0; nj < npj; nj++) {
dxj = discrete[jfirst+nj].dx;
dyj = discrete[jfirst+nj].dy;
dzj = discrete[jfirst+nj].dz;
xi[0] = x[i][0] + dxi;
xi[1] = x[i][1] + dyi;
xi[2] = x[i][2] + dzi;
xj[0] = x[j][0] + dxj;
xj[1] = x[j][1] + dyj;
xj[2] = x[j][2] + dzj;
delx = xi[0] - xj[0];
dely = xi[1] - xj[1];
delz = xi[2] - xj[2];
rsq = delx*delx + dely*dely + delz*delz;
sig = 0.5 * (discrete[ifirst+ni].sigma+discrete[jfirst+nj].sigma);
sig3 = sig*sig*sig;
term2 = 24.0*epsilon[itype][jtype] * sig3*sig3;
term1 = 2.0 * term2 * sig3*sig3;
r2inv = 1.0/rsq;
r6inv = r2inv*r2inv*r2inv;
forcelj = r6inv * (term1*r6inv - term2);
fpair = forcelj*r2inv;
if (eflag) evdwl += r6inv*(term1/12.0*r6inv-term2/6.0);
fi[0] = delx*fpair;
fi[1] = dely*fpair;
fi[2] = delz*fpair;
f[i][0] += fi[0];
f[i][1] += fi[1];
f[i][2] += fi[2];
ti[0] = dyi*fi[2] - dzi*fi[1];
ti[1] = dzi*fi[0] - dxi*fi[2];
ti[2] = dxi*fi[1] - dyi*fi[0];
torque[i][0] += ti[0];
torque[i][1] += ti[1];
torque[i][2] += ti[2];
if (newton_pair || j < nlocal) {
fj[0] = -delx*fpair;
fj[1] = -dely*fpair;
fj[2] = -delz*fpair;
f[j][0] += fj[0];
f[j][1] += fj[1];
f[j][2] += fj[2];
tj[0] = dyj*fj[2] - dzj*fj[1];
tj[1] = dzj*fj[0] - dxj*fj[2];
tj[2] = dxj*fj[1] - dyj*fj[0];
torque[j][0] += tj[0];
torque[j][1] += tj[1];
torque[j][2] += tj[2];
}
}
}
// tri/particle interaction = Nx1 particles
// c1,c2,c3 = corner pts of triangle I
} else if (tri[i] >= 0) {
if (dnum[i] == 0) {
MathExtra::quat_to_mat(bonus[tri[i]].quat,p);
MathExtra::matvec(p,bonus[tri[i]].c1,dc1);
MathExtra::matvec(p,bonus[tri[i]].c2,dc2);
MathExtra::matvec(p,bonus[tri[i]].c3,dc3);
dfirst[i] = ndiscrete;
discretize(i,sigma[itype][itype],dc1,dc2,dc3);
dnum[i] = ndiscrete - dfirst[i];
}
npi = dnum[i];
ifirst = dfirst[i];
for (ni = 0; ni < npi; ni++) {
dxi = discrete[ifirst+ni].dx;
dyi = discrete[ifirst+ni].dy;
dzi = discrete[ifirst+ni].dz;
xi[0] = x[i][0] + dxi;
xi[1] = x[i][1] + dyi;
xi[2] = x[i][2] + dzi;
xj[0] = x[j][0];
xj[1] = x[j][1];
xj[2] = x[j][2];
delx = xi[0] - xj[0];
dely = xi[1] - xj[1];
delz = xi[2] - xj[2];
rsq = delx*delx + dely*dely + delz*delz;
sig = 0.5 * (discrete[ifirst+ni].sigma+sigma[jtype][jtype]);
sig3 = sig*sig*sig;
term2 = 24.0*epsilon[itype][jtype] * sig3*sig3;
term1 = 2.0 * term2 * sig3*sig3;
r2inv = 1.0/rsq;
r6inv = r2inv*r2inv*r2inv;
forcelj = r6inv * (term1*r6inv - term2);
fpair = forcelj*r2inv;
if (eflag) evdwl += r6inv*(term1/12.0*r6inv-term2/6.0);
fi[0] = delx*fpair;
fi[1] = dely*fpair;
fi[2] = delz*fpair;
f[i][0] += fi[0];
f[i][1] += fi[1];
f[i][2] += fi[2];
ti[0] = dyi*fi[2] - dzi*fi[1];
ti[1] = dzi*fi[0] - dxi*fi[2];
ti[2] = dxi*fi[1] - dyi*fi[0];
torque[i][2] += ti[0];
torque[i][1] += ti[1];
torque[i][2] += ti[2];
if (newton_pair || j < nlocal) {
fj[0] = -delx*fpair;
fj[1] = -dely*fpair;
fj[2] = -delz*fpair;
f[j][0] += fj[0];
f[j][1] += fj[1];
f[j][2] += fj[2];
}
}
// particle/tri interaction = Nx1 particles
// c1,c2,c3 = corner pts of triangle J
} else if (tri[j] >= 0) {
if (dnum[j] == 0) {
MathExtra::quat_to_mat(bonus[tri[j]].quat,p);
MathExtra::matvec(p,bonus[tri[j]].c1,dc1);
MathExtra::matvec(p,bonus[tri[j]].c2,dc2);
MathExtra::matvec(p,bonus[tri[j]].c3,dc3);
dfirst[j] = ndiscrete;
discretize(j,sigma[jtype][jtype],dc1,dc2,dc3);
dnum[j] = ndiscrete - dfirst[j];
}
npj = dnum[j];
jfirst = dfirst[j];
for (nj = 0; nj < npj; nj++) {
dxj = discrete[jfirst+nj].dx;
dyj = discrete[jfirst+nj].dy;
dzj = discrete[jfirst+nj].dz;
xi[0] = x[i][0];
xi[1] = x[i][1];
xi[2] = x[i][2];
xj[0] = x[j][0] + dxj;
xj[1] = x[j][1] + dyj;
xj[2] = x[j][2] + dzj;
delx = xi[0] - xj[0];
dely = xi[1] - xj[1];
delz = xi[2] - xj[2];
rsq = delx*delx + dely*dely + delz*delz;
sig = 0.5 * (sigma[itype][itype]+discrete[jfirst+nj].sigma);
sig3 = sig*sig*sig;
term2 = 24.0*epsilon[itype][jtype] * sig3*sig3;
term1 = 2.0 * term2 * sig3*sig3;
r2inv = 1.0/rsq;
r6inv = r2inv*r2inv*r2inv;
forcelj = r6inv * (term1*r6inv - term2);
fpair = forcelj*r2inv;
if (eflag) evdwl += r6inv*(term1/12.0*r6inv-term2/6.0);
fi[0] = delx*fpair;
fi[1] = dely*fpair;
fi[2] = delz*fpair;
f[i][0] += fi[0];
f[i][1] += fi[1];
f[i][2] += fi[2];
if (newton_pair || j < nlocal) {
fj[0] = -delx*fpair;
fj[1] = -dely*fpair;
fj[2] = -delz*fpair;
f[j][0] += fj[0];
f[j][1] += fj[1];
f[j][2] += fj[2];
tj[0] = dyj*fj[2] - dzj*fj[1];
tj[1] = dzj*fj[0] - dxj*fj[2];
tj[2] = dxj*fj[1] - dyj*fj[0];
torque[j][0] += tj[0];
torque[j][1] += tj[1];
torque[j][2] += tj[2];
}
}
// particle/particle interaction = 1x1 particles
} else {
r2inv = 1.0/rsq;
r6inv = r2inv*r2inv*r2inv;
forcelj = r6inv * (lj1[itype][jtype]*r6inv - lj2[itype][jtype]);
fpair = forcelj*r2inv;
if (eflag)
evdwl += r6inv*(lj3[itype][jtype]*r6inv-lj4[itype][jtype]);
f[i][0] += delx*fpair;
f[i][1] += dely*fpair;
f[i][2] += delz*fpair;
if (newton_pair || j < nlocal) {
f[j][0] -= delx*fpair;
f[j][1] -= dely*fpair;
f[j][2] -= delz*fpair;
}
}
if (evflag) ev_tally(i,j,nlocal,newton_pair,
evdwl,0.0,fpair,delx,dely,delz);
}
}
if (vflag_fdotr) virial_fdotr_compute();
}
/* ----------------------------------------------------------------------
allocate all arrays
------------------------------------------------------------------------- */
void PairTri::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(cutsq,n+1,n+1,"pair:cutsq");
memory->create(cut,n+1,n+1,"pair:cut");
memory->create(epsilon,n+1,n+1,"pair:epsilon");
memory->create(sigma,n+1,n+1,"pair:sigma");
memory->create(lj1,n+1,n+1,"pair:lj1");
memory->create(lj2,n+1,n+1,"pair:lj2");
memory->create(lj3,n+1,n+1,"pair:lj3");
memory->create(lj4,n+1,n+1,"pair:lj4");
}
/* ----------------------------------------------------------------------
global settings
------------------------------------------------------------------------- */
void PairTri::settings(int narg, char **arg)
{
if (narg != 1) error->all(FLERR,"Illegal pair_style command");
cut_global = force->numeric(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[i][j] = cut_global;
}
}
/* ----------------------------------------------------------------------
set coeffs for one or more type pairs
------------------------------------------------------------------------- */
void PairTri::coeff(int narg, char **arg)
{
if (narg < 4 || narg > 5)
error->all(FLERR,"Incorrect args for pair coefficients");
if (!allocated) allocate();
int ilo,ihi,jlo,jhi;
force->bounds(arg[0],atom->ntypes,ilo,ihi);
force->bounds(arg[1],atom->ntypes,jlo,jhi);
double epsilon_one = force->numeric(arg[2]);
double sigma_one = force->numeric(arg[3]);
double cut_one = cut_global;
if (narg == 5) cut_one = force->numeric(arg[4]);
int count = 0;
for (int i = ilo; i <= ihi; i++) {
for (int j = MAX(jlo,i); j <= jhi; j++) {
epsilon[i][j] = epsilon_one;
sigma[i][j] = sigma_one;
cut[i][j] = cut_one;
setflag[i][j] = 1;
count++;
}
}
if (count == 0) error->all(FLERR,"Incorrect args for pair coefficients");
}
/* ----------------------------------------------------------------------
init for one type pair i,j and corresponding j,i
------------------------------------------------------------------------- */
double PairTri::init_one(int i, int j)
{
if (setflag[i][j] == 0) {
epsilon[i][j] = mix_energy(epsilon[i][i],epsilon[j][j],
sigma[i][i],sigma[j][j]);
sigma[i][j] = mix_distance(sigma[i][i],sigma[j][j]);
cut[i][j] = mix_distance(cut[i][i],cut[j][j]);
}
lj1[i][j] = 48.0 * epsilon[i][j] * pow(sigma[i][j],12.0);
lj2[i][j] = 24.0 * epsilon[i][j] * pow(sigma[i][j],6.0);
lj3[i][j] = 4.0 * epsilon[i][j] * pow(sigma[i][j],12.0);
lj4[i][j] = 4.0 * epsilon[i][j] * pow(sigma[i][j],6.0);
epsilon[j][i] = epsilon[i][j];
sigma[j][i] = sigma[i][j];
lj1[j][i] = lj1[i][j];
lj2[j][i] = lj2[i][j];
lj3[j][i] = lj3[i][j];
lj4[j][i] = lj4[i][j];
return cut[i][j];
}
/* ----------------------------------------------------------------------
recursively discretize triangle I with displaced corners c1,c2,c3
into N sub-tris no more than sigma in size
recurse by making 2 tris via bisecting longest side
store new discrete particles in Discrete list
------------------------------------------------------------------------- */
void PairTri::discretize(int i, double sigma,
double *c1, double *c2, double *c3)
{
double c1c2[3],c2c3[3],c1c3[3];
double centroid[3],dc1[3],dc2[3],dc3[3];
centroid[0] = (c1[0] + c2[0] + c3[0]) / 3.0;
centroid[1] = (c1[1] + c2[1] + c3[1]) / 3.0;
centroid[2] = (c1[2] + c2[2] + c3[2]) / 3.0;
MathExtra::sub3(c1,centroid,dc1);
MathExtra::sub3(c2,centroid,dc2);
MathExtra::sub3(c3,centroid,dc3);
double sigmasq = 0.25 * sigma*sigma;
double len1sq = MathExtra::lensq3(dc1);
double len2sq = MathExtra::lensq3(dc2);
double len3sq = MathExtra::lensq3(dc3);
// if sigma sphere overlaps all corner points, add particle at centroid
if (len1sq <= sigmasq && len2sq <= sigmasq & len3sq <= sigmasq) {
if (ndiscrete == dmax) {
dmax += DELTA;
discrete = (Discrete *)
memory->srealloc(discrete,dmax*sizeof(Discrete),"pair:discrete");
}
discrete[ndiscrete].dx = centroid[0];
discrete[ndiscrete].dy = centroid[1];
discrete[ndiscrete].dz = centroid[2];
sigmasq = MAX(len1sq,len2sq);
sigmasq = MAX(sigmasq,len3sq);
discrete[ndiscrete].sigma = 2.0 * sqrt(sigmasq);
ndiscrete++;
return;
}
// else break triangle into 2 sub-triangles and recurse
double c12[3],c23[3],c13[3],mid[3];
MathExtra::sub3(c2,c3,c23);
len1sq = MathExtra::lensq3(c23);
MathExtra::sub3(c1,c3,c13);
len2sq = MathExtra::lensq3(c13);
MathExtra::sub3(c1,c2,c12);
len3sq = MathExtra::lensq3(c12);
double maxsq = MAX(len1sq,len2sq);
maxsq = MAX(maxsq,len3sq);
if (len1sq == maxsq) {
MathExtra::add3(c2,c3,mid);
MathExtra::scale3(0.5,mid);
discretize(i,sigma,c1,c2,mid);
discretize(i,sigma,c1,c3,mid);
} else if (len2sq == maxsq) {
MathExtra::add3(c1,c3,mid);
MathExtra::scale3(0.5,mid);
discretize(i,sigma,c2,c1,mid);
discretize(i,sigma,c2,c3,mid);
} else {
MathExtra::add3(c1,c2,mid);
MathExtra::scale3(0.5,mid);
discretize(i,sigma,c3,c1,mid);
discretize(i,sigma,c3,c2,mid);
}
}
/* ----------------------------------------------------------------------
recursively discretize triangle I with displaced corners c1,c2,c3
into N sub-tris no more than sigma in size
recurse by making 6 tris via centroid
store new discrete particles in Discrete list
------------------------------------------------------------------------- */
/*
void PairTri::discretize(int i, double sigma,
double *c1, double *c2, double *c3)
{
double centroid[3],dc1[3],dc2[3],dc3[3];
centroid[0] = (c1[0] + c2[0] + c3[0]) / 3.0;
centroid[1] = (c1[1] + c2[1] + c3[1]) / 3.0;
centroid[2] = (c1[2] + c2[2] + c3[2]) / 3.0;
MathExtra::sub3(c1,centroid,dc1);
MathExtra::sub3(c2,centroid,dc2);
MathExtra::sub3(c3,centroid,dc3);
double sigmasq = 0.25 * sigma*sigma;
double len1sq = MathExtra::lensq3(dc1);
double len2sq = MathExtra::lensq3(dc2);
double len3sq = MathExtra::lensq3(dc3);
// if sigma sphere overlaps all corner points, add particle at centroid
if (len1sq <= sigmasq && len2sq <= sigmasq & len3sq <= sigmasq) {
if (ndiscrete == dmax) {
dmax += DELTA;
discrete = (Discrete *)
memory->srealloc(discrete,dmax*sizeof(Discrete),"pair:discrete");
}
discrete[ndiscrete].dx = centroid[0];
discrete[ndiscrete].dy = centroid[1];
discrete[ndiscrete].dz = centroid[2];
sigmasq = MAX(len1sq,len2sq);
sigmasq = MAX(sigmasq,len3sq);
discrete[ndiscrete].sigma = 2.0 * sqrt(sigmasq);
ndiscrete++;
return;
}
// else break triangle into 6 sub-triangles and recurse
double c1c2mid[3],c2c3mid[3],c1c3mid[3];
MathExtra::add3(c1,c2,c1c2mid);
MathExtra::scale3(0.5,c1c2mid);
MathExtra::add3(c2,c3,c2c3mid);
MathExtra::scale3(0.5,c2c3mid);
MathExtra::add3(c1,c3,c1c3mid);
MathExtra::scale3(0.5,c1c3mid);
discretize(i,sigma,c1,c1c2mid,centroid);
discretize(i,sigma,c1,c1c3mid,centroid);
discretize(i,sigma,c2,c2c3mid,centroid);
discretize(i,sigma,c2,c1c2mid,centroid);
discretize(i,sigma,c3,c1c3mid,centroid);
discretize(i,sigma,c3,c2c3mid,centroid);
}
*/

61
src/ASPHERE/pair_tri.h Normal file
View File

@ -0,0 +1,61 @@
/* ----------------------------------------------------------------------
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.
------------------------------------------------------------------------- */
#ifdef PAIR_CLASS
PairStyle(tri,PairTri)
#else
#ifndef LMP_PAIR_TRI_H
#define LMP_PAIR_TRI_H
#include "pair.h"
namespace LAMMPS_NS {
class PairTri : public Pair {
public:
PairTri(class LAMMPS *);
~PairTri();
void compute(int, int);
void settings(int, char **);
void coeff(int, char **);
double init_one(int, int);
protected:
double cut_global;
double **cut;
double **epsilon,**sigma;
double **lj1,**lj2,**lj3,**lj4;
class AtomVecTri *avec;
struct Discrete {
double dx,dy,dz;
double sigma;
};
Discrete *discrete; // list of all discretes for all lines
int ndiscrete; // number of discretes in list
int dmax; // allocated size of discrete list
int *dnum; // number of discretes per line, 0 if uninit
int *dfirst; // index of first discrete per each line
int nmax; // allocated size of dnum,dfirst vectors
void allocate();
void discretize(int, double, double *, double *, double *);
};
}
#endif
#endif