ICODE-ADC/lammps
4
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
Files
7ca2dcac62df9bd522e31615cc6a45c87f542bc5
lammps/src/GPU/pair_tersoff_gpu.cpp
W. Michael Brown 37f22c8627 Misc Improvements to GPU Package 
- Optimizations for molecular systems
-   Improved kernel performance and greater CPU overlap
- Reduced GPU to CPU communications for discrete devices
- Switch classic Intel makefiles to use LLVM-based compilers
- Prefetch optimizations supported for OpenCL
- Optimized data repack for quaternions
2023-03-05 21:03:12 -08:00

248 lines
9.0 KiB
C++
Raw Blame History

/* ----------------------------------------------------------------------
LAMMPS - Large-scale Atomic/Molecular Massively Parallel Simulator
https://www.lammps.org/, Sandia National Laboratories
LAMMPS development team: developers@lammps.org
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 author: Trung Dac Nguyen (ndactrung@gmail.com)
------------------------------------------------------------------------- */
#include "pair_tersoff_gpu.h"
#include "atom.h"
#include "comm.h"
#include "domain.h"
#include "error.h"
#include "force.h"
#include "gpu_extra.h"
#include "memory.h"
#include "neigh_list.h"
#include "neighbor.h"
#include "suffix.h"
using namespace LAMMPS_NS;
// External functions from cuda library for atom decomposition
int tersoff_gpu_init(const int ntypes, const int inum, const int nall, const int max_nbors,
const double cell_size, int &gpu_mode, FILE *screen, int *host_map,
const int nelements, int ***host_elem3param, const int nparams,
const double *ts_lam1, const double *ts_lam2, const double *ts_lam3,
const double *ts_powermint, const double *ts_biga, const double *ts_bigb,
const double *ts_bigr, const double *ts_bigd, const double *ts_c1,
const double *ts_c2, const double *ts_c3, const double *ts_c4,
const double *ts_c, const double *ts_d, const double *ts_h,
const double *ts_gamma, const double *ts_beta, const double *ts_powern,
const double *ts_cutsq);
void tersoff_gpu_clear();
int **tersoff_gpu_compute_n(const int ago, const int inum_full, const int nall, double **host_x,
int *host_type, double *sublo, double *subhi, tagint *tag,
int **nspecial, tagint **special, const bool eflag, const bool vflag,
const bool eatom, const bool vatom, int &host_start, int **ilist,
int **jnum, const double cpu_time, bool &success);
void tersoff_gpu_compute(const int ago, const int nlocal, const int nall, const int nlist,
double **host_x, int *host_type, int *ilist, int *numj, int **firstneigh,
const bool eflag, const bool vflag, const bool eatom, const bool vatom,
int &host_start, const double cpu_time, bool &success);
double tersoff_gpu_bytes();
#define MAXLINE 1024
#define DELTA 4
/* ---------------------------------------------------------------------- */
PairTersoffGPU::PairTersoffGPU(LAMMPS *lmp) : PairTersoff(lmp), gpu_mode(GPU_FORCE)
{
cpu_time = 0.0;
suffix_flag |= Suffix::GPU;
GPU_EXTRA::gpu_ready(lmp->modify, lmp->error);
cutghost = nullptr;
ghostneigh = 1;
}
/* ----------------------------------------------------------------------
check if allocated, since class can be destructed when incomplete
------------------------------------------------------------------------- */
PairTersoffGPU::~PairTersoffGPU()
{
tersoff_gpu_clear();
if (allocated) memory->destroy(cutghost);
}
/* ---------------------------------------------------------------------- */
void PairTersoffGPU::compute(int eflag, int vflag)
{
ev_init(eflag, vflag);
int nall = atom->nlocal + atom->nghost;
int inum, host_start;
bool success = true;
int *ilist, *numneigh, **firstneigh;
if (gpu_mode != GPU_FORCE) {
double sublo[3], subhi[3];
if (domain->triclinic == 0) {
sublo[0] = domain->sublo[0];
sublo[1] = domain->sublo[1];
sublo[2] = domain->sublo[2];
subhi[0] = domain->subhi[0];
subhi[1] = domain->subhi[1];
subhi[2] = domain->subhi[2];
} else {
domain->bbox(domain->sublo_lamda, domain->subhi_lamda, sublo, subhi);
}
inum = atom->nlocal;
firstneigh =
tersoff_gpu_compute_n(neighbor->ago, inum, nall, atom->x, atom->type, sublo, subhi,
atom->tag, atom->nspecial, atom->special, eflag, vflag, eflag_atom,
vflag_atom, host_start, &ilist, &numneigh, cpu_time, success);
} else {
inum = list->inum;
ilist = list->ilist;
numneigh = list->numneigh;
firstneigh = list->firstneigh;
tersoff_gpu_compute(neighbor->ago, inum, nall, inum + list->gnum, atom->x, atom->type, ilist,
numneigh, firstneigh, eflag, vflag, eflag_atom, vflag_atom, host_start,
cpu_time, success);
}
if (!success) error->one(FLERR, "Insufficient memory on accelerator");
if (atom->molecular != Atom::ATOMIC && neighbor->ago == 0)
neighbor->build_topology();
}
/* ---------------------------------------------------------------------- */
void PairTersoffGPU::allocate()
{
PairTersoff::allocate();
int n = atom->ntypes;
memory->create(cutghost, n + 1, n + 1, "pair:cutghost");
}
/* ----------------------------------------------------------------------
init specific to this pair style
------------------------------------------------------------------------- */
void PairTersoffGPU::init_style()
{
double cell_size = cutmax + neighbor->skin;
if (atom->tag_enable == 0) error->all(FLERR, "Pair style tersoff/gpu requires atom IDs");
double *lam1, *lam2, *lam3, *powermint;
double *biga, *bigb, *bigr, *bigd;
double *c1, *c2, *c3, *c4;
double *c, *d, *h, *gamma;
double *beta, *powern, *_cutsq;
lam1 = lam2 = lam3 = powermint = nullptr;
biga = bigb = bigr = bigd = nullptr;
c1 = c2 = c3 = c4 = nullptr;
c = d = h = gamma = nullptr;
beta = powern = _cutsq = nullptr;
memory->create(lam1, nparams, "pair:lam1");
memory->create(lam2, nparams, "pair:lam2");
memory->create(lam3, nparams, "pair:lam3");
memory->create(powermint, nparams, "pair:powermint");
memory->create(biga, nparams, "pair:biga");
memory->create(bigb, nparams, "pair:bigb");
memory->create(bigr, nparams, "pair:bigr");
memory->create(bigd, nparams, "pair:bigd");
memory->create(c1, nparams, "pair:c1");
memory->create(c2, nparams, "pair:c2");
memory->create(c3, nparams, "pair:c3");
memory->create(c4, nparams, "pair:c4");
memory->create(c, nparams, "pair:c");
memory->create(d, nparams, "pair:d");
memory->create(h, nparams, "pair:h");
memory->create(gamma, nparams, "pair:gamma");
memory->create(beta, nparams, "pair:beta");
memory->create(powern, nparams, "pair:powern");
memory->create(_cutsq, nparams, "pair:_cutsq");
for (int i = 0; i < nparams; i++) {
lam1[i] = params[i].lam1;
lam2[i] = params[i].lam2;
lam3[i] = params[i].lam3;
powermint[i] = params[i].powermint;
biga[i] = params[i].biga;
bigb[i] = params[i].bigb;
bigr[i] = params[i].bigr;
bigd[i] = params[i].bigd;
c1[i] = params[i].c1;
c2[i] = params[i].c2;
c3[i] = params[i].c3;
c4[i] = params[i].c4;
c[i] = params[i].c;
d[i] = params[i].d;
h[i] = params[i].h;
gamma[i] = params[i].gamma;
beta[i] = params[i].beta;
powern[i] = params[i].powern;
_cutsq[i] = params[i].cutsq;
}
int mnf = 5e-2 * neighbor->oneatom;
int success = tersoff_gpu_init(atom->ntypes + 1, atom->nlocal, atom->nlocal + atom->nghost, mnf,
cell_size, gpu_mode, screen, map, nelements, elem3param, nparams,
lam1, lam2, lam3, powermint, biga, bigb, bigr, bigd, c1, c2, c3,
c4, c, d, h, gamma, beta, powern, _cutsq);
memory->destroy(lam1);
memory->destroy(lam2);
memory->destroy(lam3);
memory->destroy(powermint);
memory->destroy(biga);
memory->destroy(bigb);
memory->destroy(bigr);
memory->destroy(bigd);
memory->destroy(c1);
memory->destroy(c2);
memory->destroy(c3);
memory->destroy(c4);
memory->destroy(c);
memory->destroy(d);
memory->destroy(h);
memory->destroy(gamma);
memory->destroy(beta);
memory->destroy(powern);
memory->destroy(_cutsq);
GPU_EXTRA::check_flag(success, error, world);
if (gpu_mode == GPU_FORCE)
neighbor->add_request(this, NeighConst::REQ_FULL | NeighConst::REQ_GHOST);
if (comm->get_comm_cutoff() < (2.0 * cutmax + neighbor->skin)) {
comm->cutghostuser = 2.0 * cutmax + neighbor->skin;
if (comm->me == 0)
error->warning(FLERR, "Increasing communication cutoff to {:.8} for GPU pair style",
comm->cutghostuser);
}
}
/* ----------------------------------------------------------------------
init for one type pair i,j and corresponding j,i
------------------------------------------------------------------------- */
double PairTersoffGPU::init_one(int i, int j)
{
if (setflag[i][j] == 0) error->all(FLERR, "All pair coeffs are not set");
cutghost[i][j] = cutmax;
cutghost[j][i] = cutmax;
return cutmax;
}
Reference in New Issue View Git Blame Copy Permalink