jianmu
/
lammps
forked from lijiext/lammps
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

git-svn-id: svn://svn.icms.temple.edu/lammps-ro/trunk@13937 f3b2605a-c512-4ea7-a41b-209d697bcdaa

This commit is contained in:
sjplimp 2015-08-28 20:00:56 +00:00
parent 962dd0da74
commit ce59e32d1c
189 changed files with 1526 additions and 587 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
src/GPU/fix_gpu.cpp
View File

@ -295,7 +295,7 @@ void FixGPU::post_force(int vflag)
force->pair->virial[5] += lvirial[5];
if (force->pair->vflag_fdotr) force->pair->virial_fdotr_compute();
timer->stamp(TIME_PAIR);
timer->stamp(Timer::PAIR);
}
/* ---------------------------------------------------------------------- */

51
src/KOKKOS/verlet_kokkos.cpp
View File

@ -107,7 +107,7 @@ void VerletKokkos::setup()
atomKK->sync(force->pair->execution_space,force->pair->datamask_read);
atomKK->modified(force->pair->execution_space,force->pair->datamask_modify);
force->pair->compute(eflag,vflag);
timer->stamp(TIME_PAIR);
timer->stamp(Timer::PAIR);
}
else if (force->pair) force->pair->compute_dummy(eflag,vflag);
@ -133,7 +133,7 @@ void VerletKokkos::setup()
atomKK->modified(force->improper->execution_space,force->improper->datamask_modify);
force->improper->compute(eflag,vflag);
}
timer->stamp(TIME_BOND);
timer->stamp(Timer::BOND);
}
if(force->kspace) {
@ -142,7 +142,7 @@ void VerletKokkos::setup()
atomKK->sync(force->kspace->execution_space,force->kspace->datamask_read);
atomKK->modified(force->kspace->execution_space,force->kspace->datamask_modify);
force->kspace->compute(eflag,vflag);
timer->stamp(TIME_KSPACE);
timer->stamp(Timer::KSPACE);
} else force->kspace->compute_dummy(eflag,vflag);
}
@ -209,7 +209,7 @@ void VerletKokkos::setup_minimal(int flag)
atomKK->sync(force->pair->execution_space,force->pair->datamask_read);
atomKK->modified(force->pair->execution_space,force->pair->datamask_modify);
force->pair->compute(eflag,vflag);
timer->stamp(TIME_PAIR);
timer->stamp(Timer::PAIR);
}
else if (force->pair) force->pair->compute_dummy(eflag,vflag);
@ -235,7 +235,7 @@ void VerletKokkos::setup_minimal(int flag)
atomKK->modified(force->improper->execution_space,force->improper->datamask_modify);
force->improper->compute(eflag,vflag);
}
timer->stamp(TIME_BOND);
timer->stamp(Timer::BOND);
}
if(force->kspace) {
@ -244,7 +244,7 @@ void VerletKokkos::setup_minimal(int flag)
atomKK->sync(force->kspace->execution_space,force->kspace->datamask_read);
atomKK->modified(force->kspace->execution_space,force->kspace->datamask_modify);
force->kspace->compute(eflag,vflag);
timer->stamp(TIME_KSPACE);
timer->stamp(Timer::KSPACE);
} else force->kspace->compute_dummy(eflag,vflag);
}
@ -286,9 +286,11 @@ void VerletKokkos::run(int n)
// initial time integration
ktimer.reset();
timer->stamp();
modify->initial_integrate(vflag);
time += ktimer.seconds();
if (n_post_integrate) modify->post_integrate();
timer->stamp(Timer::MODIFY);
// regular communication vs neighbor list rebuild
@ -297,13 +299,17 @@ void VerletKokkos::run(int n)
if (nflag == 0) {
timer->stamp();
comm->forward_comm();
timer->stamp(TIME_COMM);
timer->stamp(Timer::COMM);
} else {
// added debug
//atomKK->sync(Host,ALL_MASK);
//atomKK->modified(Host,ALL_MASK);
if (n_pre_exchange) modify->pre_exchange();
if (n_pre_exchange) {
timer->stamp();
modify->pre_exchange();
timer->stamp(Timer::MODIFY);
}
// debug
//atomKK->sync(Host,ALL_MASK);
//atomKK->modified(Host,ALL_MASK);
@ -330,10 +336,13 @@ void VerletKokkos::run(int n)
if (triclinic) domain->lamda2x(atomKK->nlocal+atomKK->nghost);
timer->stamp(TIME_COMM);
if (n_pre_neighbor) modify->pre_neighbor();
timer->stamp(Timer::COMM);
if (n_pre_neighbor) {
modify->pre_neighbor();
timer->stamp(Timer::MODIFY);
}
neighbor->build();
timer->stamp(TIME_NEIGHBOR);
timer->stamp(Timer::NEIGH);
}
// force computations
@ -342,19 +351,24 @@ void VerletKokkos::run(int n)
// and Pair:ev_tally() needs to be called before any tallying
force_clear();
timer->stamp();
// added for debug
//atomKK->k_x.sync<LMPHostType>();
//atomKK->k_f.sync<LMPHostType>();
//atomKK->k_f.modify<LMPHostType>();
if (n_pre_force) modify->pre_force(vflag);
if (n_pre_force) {
modify->pre_force(vflag);
timer->stamp(Timer::MODIFY);
}
timer->stamp();
if (pair_compute_flag) {
atomKK->sync(force->pair->execution_space,force->pair->datamask_read);
atomKK->modified(force->pair->execution_space,force->pair->datamask_modify);
force->pair->compute(eflag,vflag);
timer->stamp(TIME_PAIR);
timer->stamp(Timer::PAIR);
}
if (atomKK->molecular) {
@ -378,20 +392,20 @@ void VerletKokkos::run(int n)
atomKK->modified(force->improper->execution_space,force->improper->datamask_modify);
force->improper->compute(eflag,vflag);
}
timer->stamp(TIME_BOND);
timer->stamp(Timer::BOND);
}
if (kspace_compute_flag) {
atomKK->sync(force->kspace->execution_space,force->kspace->datamask_read);
atomKK->modified(force->kspace->execution_space,force->kspace->datamask_modify);
force->kspace->compute(eflag,vflag);
timer->stamp(TIME_KSPACE);
timer->stamp(Timer::KSPACE);
}
// reverse communication of forces
if (force->newton) comm->reverse_comm();
timer->stamp(TIME_COMM);
timer->stamp(Timer::COMM);
// force modifications, final time integration, diagnostics
@ -400,6 +414,7 @@ void VerletKokkos::run(int n)
if (n_post_force) modify->post_force(vflag);
modify->final_integrate();
if (n_end_of_step) modify->end_of_step();
timer->stamp(Timer::MODIFY);
time += ktimer.seconds();
@ -410,7 +425,7 @@ void VerletKokkos::run(int n)
timer->stamp();
output->write(ntimestep);
timer->stamp(TIME_OUTPUT);
timer->stamp(Timer::OUTPUT);
}
}
}

4
src/KSPACE/fix_tune_kspace.cpp
View File

@ -159,7 +159,7 @@ void FixTuneKspace::pre_exchange()
adjust_rcut(time);
}
last_spcpu = timer->elapsed(TIME_LOOP);
last_spcpu = timer->elapsed(Timer::TOTAL);
}
/* ----------------------------------------------------------------------
@ -177,7 +177,7 @@ double FixTuneKspace::get_timing_info()
dvalue = 0.0;
firststep = 1;
} else {
new_cpu = timer->elapsed(TIME_LOOP);
new_cpu = timer->elapsed(Timer::TOTAL);
double cpu_diff = new_cpu - last_spcpu;
int step_diff = new_step - last_step;
if (step_diff > 0.0) dvalue = cpu_diff/step_diff;

8
src/REPLICA/neb.cpp
View File

@ -232,7 +232,7 @@ void NEB::run()
// damped dynamic min styles insure all replicas converge together
timer->init();
timer->barrier_start(TIME_LOOP);
timer->barrier_start();
while (update->minimize->niter < n1steps) {
update->minimize->run(nevery);
@ -240,7 +240,7 @@ void NEB::run()
if (update->minimize->stop_condition) break;
}
timer->barrier_stop(TIME_LOOP);
timer->barrier_stop();
update->minimize->cleanup();
@ -302,7 +302,7 @@ void NEB::run()
// damped dynamic min styles insure all replicas converge together
timer->init();
timer->barrier_start(TIME_LOOP);
timer->barrier_start();
while (update->minimize->niter < n2steps) {
update->minimize->run(nevery);
@ -310,7 +310,7 @@ void NEB::run()
if (update->minimize->stop_condition) break;
}
timer->barrier_stop(TIME_LOOP);
timer->barrier_stop();
update->minimize->cleanup();

84
src/REPLICA/prd.cpp
View File

@ -278,8 +278,8 @@ void PRD::command(int narg, char **arg)
share_event(0,0,0);
timer->init();
timer->barrier_start(TIME_LOOP);
time_start = timer->array[TIME_LOOP];
timer->barrier_start();
time_start = timer->get_wall(Timer::TOTAL);
log_event();
@ -305,8 +305,8 @@ void PRD::command(int narg, char **arg)
time_dephase = time_dynamics = time_quench = time_comm = time_output = 0.0;
bigint clock = 0;
timer->barrier_start(TIME_LOOP);
time_start = timer->array[TIME_LOOP];
timer->barrier_start();
time_start = timer->get_wall(Timer::TOTAL);
int istep = 0;
@ -386,7 +386,7 @@ void PRD::command(int narg, char **arg)
lmp->init();
update->integrate->setup();
timer->barrier_start(TIME_LOOP);
timer->barrier_start();
if (t_corr > 0) replicate(ireplica);
if (temp_flag == 0) {
@ -396,16 +396,16 @@ void PRD::command(int narg, char **arg)
universe->uworld);
}
timer->barrier_stop(TIME_LOOP);
time_comm += timer->array[TIME_LOOP];
timer->barrier_stop();
time_comm += timer->get_wall(Timer::TOTAL);
// write restart file of hot coords
if (restart_flag) {
timer->barrier_start(TIME_LOOP);
timer->barrier_start();
output->write_restart(update->ntimestep);
timer->barrier_stop(TIME_LOOP);
time_output += timer->array[TIME_LOOP];
timer->barrier_stop();
time_output += timer->get_wall(Timer::TOTAL);
}
if (stepmode == 0) istep = update->ntimestep - update->beginstep;
@ -416,14 +416,14 @@ void PRD::command(int narg, char **arg)
// set total timers and counters so Finish() will process them
timer->array[TIME_LOOP] = time_start;
timer->barrier_stop(TIME_LOOP);
timer->set_wall(Timer::TOTAL, time_start);
timer->barrier_stop();
timer->array[TIME_PAIR] = time_dephase;
timer->array[TIME_BOND] = time_dynamics;
timer->array[TIME_KSPACE] = time_quench;
timer->array[TIME_COMM] = time_comm;
timer->array[TIME_OUTPUT] = time_output;
timer->set_wall(Timer::DEPHASE, time_dephase);
timer->set_wall(Timer::DYNAMICS, time_dynamics);
timer->set_wall(Timer::QUENCH, time_quench);
timer->set_wall(Timer::REPCOMM, time_comm);
timer->set_wall(Timer::REPOUT, time_output);
neighbor->ncalls = nbuild;
neighbor->ndanger = ndanger;
@ -431,14 +431,14 @@ void PRD::command(int narg, char **arg)
if (me_universe == 0) {
if (universe->uscreen)
fprintf(universe->uscreen,
"Loop time of %g on %d procs for %d steps with " BIGINT_FORMAT
"Loop time of %g on %d procs for %d steps with " BIGINT_FORMAT
" atoms\n",
timer->array[TIME_LOOP],nprocs_universe,nsteps,atom->natoms);
timer->get_wall(Timer::TOTAL),nprocs_universe,nsteps,atom->natoms);
if (universe->ulogfile)
fprintf(universe->ulogfile,
"Loop time of %g on %d procs for %d steps with " BIGINT_FORMAT
"Loop time of %g on %d procs for %d steps with " BIGINT_FORMAT
" atoms\n",
timer->array[TIME_LOOP],nprocs_universe,nsteps,atom->natoms);
timer->get_wall(Timer::TOTAL),nprocs_universe,nsteps,atom->natoms);
}
if (me == 0) {
@ -541,10 +541,10 @@ void PRD::dynamics(int nsteps, double &time_category)
//modify->addstep_compute_all(update->ntimestep);
bigint ncalls = neighbor->ncalls;
timer->barrier_start(TIME_LOOP);
timer->barrier_start();
update->integrate->run(nsteps);
timer->barrier_stop(TIME_LOOP);
time_category += timer->array[TIME_LOOP];
timer->barrier_stop();
time_category += timer->get_wall(Timer::TOTAL);
nbuild += neighbor->ncalls - ncalls;
ndanger += neighbor->ndanger;
@ -583,10 +583,10 @@ void PRD::quench()
int ncalls = neighbor->ncalls;
timer->barrier_start(TIME_LOOP);
timer->barrier_start();
update->minimize->run(maxiter);
timer->barrier_stop(TIME_LOOP);
time_quench += timer->array[TIME_LOOP];
timer->barrier_stop();
time_quench += timer->get_wall(Timer::TOTAL);
if (neighbor->ncalls == ncalls) quench_reneighbor = 0;
else quench_reneighbor = 1;
@ -619,7 +619,7 @@ int PRD::check_event(int replica_num)
if (compute_event->compute_scalar() > 0.0) worldflag = 1;
if (replica_num >= 0 && replica_num != universe->iworld) worldflag = 0;
timer->barrier_start(TIME_LOOP);
timer->barrier_start();
if (me == 0) MPI_Allreduce(&worldflag,&universeflag,1,
MPI_INT,MPI_SUM,comm_replica);
@ -655,8 +655,8 @@ int PRD::check_event(int replica_num)
MPI_Bcast(&ireplica,1,MPI_INT,0,world);
}
timer->barrier_stop(TIME_LOOP);
time_comm += timer->array[TIME_LOOP];
timer->barrier_stop();
time_comm += timer->get_wall(Timer::TOTAL);
return ireplica;
}
@ -672,14 +672,14 @@ int PRD::check_event(int replica_num)
void PRD::share_event(int ireplica, int flag, int decrement)
{
timer->barrier_start(TIME_LOOP);
timer->barrier_start();
// communicate quenched coords to all replicas and store as event
// decrement event counter if flag = 0 since not really an event
replicate(ireplica);
timer->barrier_stop(TIME_LOOP);
time_comm += timer->array[TIME_LOOP];
timer->barrier_stop();
time_comm += timer->get_wall(Timer::TOTAL);
// adjust time for last correlated event check (not on first event)
@ -717,21 +717,21 @@ void PRD::share_event(int ireplica, int flag, int decrement)
// addstep_compute_all insures eng/virial are calculated if needed
if (output->ndump && universe->iworld == 0) {
timer->barrier_start(TIME_LOOP);
timer->barrier_start();
modify->addstep_compute_all(update->ntimestep);
update->integrate->setup_minimal(1);
output->write_dump(update->ntimestep);
timer->barrier_stop(TIME_LOOP);
time_output += timer->array[TIME_LOOP];
timer->barrier_stop();
time_output += timer->get_wall(Timer::TOTAL);
}
// restore and communicate hot coords to all replicas
fix_event->restore_state_quench();
timer->barrier_start(TIME_LOOP);
timer->barrier_start();
replicate(ireplica);
timer->barrier_stop(TIME_LOOP);
time_comm += timer->array[TIME_LOOP];
timer->barrier_stop();
time_comm += timer->get_wall(Timer::TOTAL);
}
/* ----------------------------------------------------------------------
@ -740,13 +740,13 @@ void PRD::share_event(int ireplica, int flag, int decrement)
void PRD::log_event()
{
timer->array[TIME_LOOP] = time_start;
timer->set_wall(Timer::TOTAL, time_start);
if (universe->me == 0) {
if (universe->uscreen)
fprintf(universe->uscreen,
BIGINT_FORMAT " %.3f " BIGINT_FORMAT " %d %d %d %d\n",
fix_event->event_timestep,
timer->elapsed(TIME_LOOP),
timer->elapsed(Timer::TOTAL),
fix_event->clock,
fix_event->event_number,fix_event->correlated_event,
fix_event->ncoincident,
@ -755,7 +755,7 @@ void PRD::log_event()
fprintf(universe->ulogfile,
BIGINT_FORMAT " %.3f " BIGINT_FORMAT " %d %d %d %d\n",
fix_event->event_timestep,
timer->elapsed(TIME_LOOP),
timer->elapsed(Timer::TOTAL),
fix_event->clock,
fix_event->event_number,fix_event->correlated_event,
fix_event->ncoincident,

92
src/REPLICA/tad.cpp
View File

@ -255,8 +255,8 @@ void TAD::command(int narg, char **arg)
quench();
timer->init();
timer->barrier_start(TIME_LOOP);
time_start = timer->array[TIME_LOOP];
timer->barrier_start();
time_start = timer->get_wall(Timer::TOTAL);
fix_event->store_event_tad(update->ntimestep);
log_event(0);
fix_event->restore_state_quench();
@ -275,8 +275,8 @@ void TAD::command(int narg, char **arg)
nbuild = ndanger = 0;
time_neb = time_dynamics = time_quench = time_comm = time_output = 0.0;
timer->barrier_start(TIME_LOOP);
time_start = timer->array[TIME_LOOP];
timer->barrier_start();
time_start = timer->get_wall(Timer::TOTAL);
int confident_flag, event_flag;
@ -347,10 +347,10 @@ void TAD::command(int narg, char **arg)
// write restart file of hot coords
if (restart_flag) {
timer->barrier_start(TIME_LOOP);
timer->barrier_start();
output->write_restart(update->ntimestep);
timer->barrier_stop(TIME_LOOP);
time_output += timer->array[TIME_LOOP];
timer->barrier_stop();
time_output += timer->get_wall(Timer::TOTAL);
}
}
@ -381,14 +381,14 @@ void TAD::command(int narg, char **arg)
// set total timers and counters so Finish() will process them
timer->array[TIME_LOOP] = time_start;
timer->barrier_stop(TIME_LOOP);
timer->set_wall(Timer::TOTAL, time_start);
timer->barrier_stop();
timer->array[TIME_PAIR] = time_neb;
timer->array[TIME_BOND] = time_dynamics;
timer->array[TIME_KSPACE] = time_quench;
timer->array[TIME_COMM] = time_comm;
timer->array[TIME_OUTPUT] = time_output;
timer->set_wall(Timer::NEB, time_neb);
timer->set_wall(Timer::DYNAMICS, time_dynamics);
timer->set_wall(Timer::QUENCH, time_quench);
timer->set_wall(Timer::REPCOMM, time_comm);
timer->set_wall(Timer::REPOUT, time_output);
neighbor->ncalls = nbuild;
neighbor->ndanger = ndanger;
@ -396,14 +396,14 @@ void TAD::command(int narg, char **arg)
if (me_universe == 0) {
if (universe->uscreen)
fprintf(universe->uscreen,
"Loop time of %g on %d procs for %d steps with " BIGINT_FORMAT
"Loop time of %g on %d procs for %d steps with " BIGINT_FORMAT
" atoms\n",
timer->array[TIME_LOOP],nprocs_universe,nsteps,atom->natoms);
if (universe->ulogfile)
timer->get_wall(Timer::TOTAL),nprocs_universe,nsteps,atom->natoms);
if (universe->ulogfile)
fprintf(universe->ulogfile,
"Loop time of %g on %d procs for %d steps with " BIGINT_FORMAT
"Loop time of %g on %d procs for %d steps with " BIGINT_FORMAT
" atoms\n",
timer->array[TIME_LOOP],nprocs_universe,nsteps,atom->natoms);
timer->get_wall(Timer::TOTAL),nprocs_universe,nsteps,atom->natoms);
}
if (me_universe == 0) fclose(ulogfile_neb);
@ -451,10 +451,10 @@ void TAD::dynamics()
//modify->addstep_compute_all(update->ntimestep);
int ncalls = neighbor->ncalls;
timer->barrier_start(TIME_LOOP);
timer->barrier_start();
update->integrate->run(t_event);
timer->barrier_stop(TIME_LOOP);
time_dynamics += timer->array[TIME_LOOP];
timer->barrier_stop();
time_dynamics += timer->get_wall(Timer::TOTAL);
nbuild += neighbor->ncalls - ncalls;
ndanger += neighbor->ndanger;
@ -493,10 +493,10 @@ void TAD::quench()
int ncalls = neighbor->ncalls;
timer->barrier_start(TIME_LOOP);
timer->barrier_start();
update->minimize->run(maxiter);
timer->barrier_stop(TIME_LOOP);
time_quench += timer->array[TIME_LOOP];
timer->barrier_stop();
time_quench += timer->get_wall(Timer::TOTAL);
if (neighbor->ncalls == ncalls) quench_reneighbor = 0;
else quench_reneighbor = 1;
@ -535,14 +535,14 @@ int TAD::check_event()
void TAD::log_event(int ievent)
{
timer->array[TIME_LOOP] = time_start;
timer->set_wall(Timer::TOTAL, time_start);
if (universe->me == 0) {
double tfrac = 0.0;
if (universe->uscreen)
fprintf(universe->uscreen,
BIGINT_FORMAT " %.3f %d %d %s %.3f %.3f %.3f %.3f\n",
fix_event->event_timestep,
timer->elapsed(TIME_LOOP),
timer->elapsed(Timer::TOTAL),
fix_event->event_number,ievent,
"E ",
fix_event->ebarrier,tfrac,
@ -551,7 +551,7 @@ void TAD::log_event(int ievent)
fprintf(universe->ulogfile,
BIGINT_FORMAT " %.3f %d %d %s %.3f %.3f %.3f %.3f\n",
fix_event->event_timestep,
timer->elapsed(TIME_LOOP),
timer->elapsed(Timer::TOTAL),
fix_event->event_number,ievent,
"E ",
fix_event->ebarrier,tfrac,
@ -563,12 +563,12 @@ void TAD::log_event(int ievent)
// addstep_compute_all insures eng/virial are calculated if needed
if (output->ndump && universe->iworld == 0) {
timer->barrier_start(TIME_LOOP);
timer->barrier_start();
modify->addstep_compute_all(update->ntimestep);
update->integrate->setup_minimal(1);
output->write_dump(update->ntimestep);
timer->barrier_stop(TIME_LOOP);
time_output += timer->array[TIME_LOOP];
timer->barrier_stop();
time_output += timer->get_wall(Timer::TOTAL);
}
}
@ -604,10 +604,10 @@ void TAD::options(int narg, char **arg)
while (iarg < narg) {
if (strcmp(arg[iarg],"min") == 0) {
if (iarg+5 > narg) error->all(FLERR,"Illegal tad command");
etol = atof(arg[iarg+1]);
ftol = atof(arg[iarg+2]);
maxiter = atoi(arg[iarg+3]);
maxeval = atoi(arg[iarg+4]);
etol = force->numeric(FLERR,arg[iarg+1]);
ftol = force->numeric(FLERR,arg[iarg+2]);
maxiter = force->inumeric(FLERR,arg[iarg+3]);
maxeval = force->inumeric(FLERR,arg[iarg+4]);
if (maxiter < 0 || maxeval < 0 ||
etol < 0.0 || ftol < 0.0 )
error->all(FLERR,"Illegal tad command");
@ -615,11 +615,11 @@ void TAD::options(int narg, char **arg)
} else if (strcmp(arg[iarg],"neb") == 0) {
if (iarg+6 > narg) error->all(FLERR,"Illegal tad command");
etol_neb = atof(arg[iarg+1]);
ftol_neb = atof(arg[iarg+2]);
n1steps_neb = atoi(arg[iarg+3]);
n2steps_neb = atoi(arg[iarg+4]);
nevery_neb = atoi(arg[iarg+5]);
etol_neb = force->numeric(FLERR,arg[iarg+1]);
ftol_neb = force->numeric(FLERR,arg[iarg+2]);
n1steps_neb = force->inumeric(FLERR,arg[iarg+3]);
n2steps_neb = force->inumeric(FLERR,arg[iarg+4]);
nevery_neb = force->inumeric(FLERR,arg[iarg+5]);
if (etol_neb < 0.0 || ftol_neb < 0.0 ||
n1steps_neb < 0 || n2steps_neb < 0 ||
nevery_neb < 0) error->all(FLERR,"Illegal tad command");
@ -753,10 +753,10 @@ void TAD::perform_neb(int ievent)
// had to bypass timer interface
// because timer->array is reset inside neb->run()
// timer->barrier_start(TIME_LOOP);
// timer->barrier_start();
// neb->run();
// timer->barrier_stop(TIME_LOOP);
// time_neb += timer->array[TIME_LOOP];
// timer->barrier_stop();
// time_neb += timer->get_wall(Timer::TOTAL);
MPI_Barrier(world);
double time_tmp = MPI_Wtime();
@ -977,7 +977,7 @@ void TAD::compute_tlo(int ievent)
// first-replica output about each event
timer->array[TIME_LOOP] = time_start;
timer->set_wall(Timer::TOTAL, time_start);
if (universe->me == 0) {
double tfrac = 0.0;
if (ievent > 0) tfrac = delthi/deltstop;
@ -986,7 +986,7 @@ void TAD::compute_tlo(int ievent)
fprintf(universe->uscreen,
BIGINT_FORMAT " %.3f %d %d %s %.3f %.3f %.3f %.3f\n",
fix_event_list[ievent]->event_timestep,
timer->elapsed(TIME_LOOP),
timer->elapsed(Timer::TOTAL),
fix_event->event_number,
ievent,statstr,ebarrier,tfrac,
fix_event->tlo,deltlo);
@ -995,7 +995,7 @@ void TAD::compute_tlo(int ievent)
fprintf(universe->ulogfile,
BIGINT_FORMAT " %.3f %d %d %s %.3f %.3f %.3f %.3f\n",
fix_event_list[ievent]->event_timestep,
timer->elapsed(TIME_LOOP),
timer->elapsed(Timer::TOTAL),
fix_event->event_number,
ievent,statstr,ebarrier,tfrac,
fix_event->tlo,deltlo);

4
src/REPLICA/temper.cpp
View File

@ -204,7 +204,7 @@ void Temper::command(int narg, char **arg)
}
timer->init();
timer->barrier_start(TIME_LOOP);
timer->barrier_start();
for (int iswap = 0; iswap < nswaps; iswap++) {
@ -309,7 +309,7 @@ void Temper::command(int narg, char **arg)
if (me_universe == 0) print_status();
}
timer->barrier_stop(TIME_LOOP);
timer->barrier_stop();
update->integrate->cleanup();

26
src/REPLICA/verlet_split.cpp
View File

@ -279,7 +279,7 @@ void VerletSplit::run(int n)
MPI_Barrier(universe->uworld);
timer->init();
timer->barrier_start(TIME_LOOP);
timer->barrier_start();
// setup initial Rspace <-> Kspace comm params
@ -325,7 +325,7 @@ void VerletSplit::run(int n)
if (nflag == 0) {
timer->stamp();
comm->forward_comm();
timer->stamp(TIME_COMM);
timer->stamp(Timer::COMM);
} else {
if (n_pre_exchange) modify->pre_exchange();
if (triclinic) domain->x2lamda(atom->nlocal);
@ -340,10 +340,10 @@ void VerletSplit::run(int n)
if (sortflag && ntimestep >= atom->nextsort) atom->sort();
comm->borders();
if (triclinic) domain->lamda2x(atom->nlocal+atom->nghost);
timer->stamp(TIME_COMM);
timer->stamp(Timer::COMM);
if (n_pre_neighbor) modify->pre_neighbor();
neighbor->build();
timer->stamp(TIME_NEIGHBOR);
timer->stamp(Timer::NEIGH);
}
}
@ -363,7 +363,7 @@ void VerletSplit::run(int n)
timer->stamp();
if (force->pair) {
force->pair->compute(eflag,vflag);
timer->stamp(TIME_PAIR);
timer->stamp(Timer::PAIR);
}
if (atom->molecular) {
@ -371,12 +371,12 @@ void VerletSplit::run(int n)
if (force->angle) force->angle->compute(eflag,vflag);
if (force->dihedral) force->dihedral->compute(eflag,vflag);
if (force->improper) force->improper->compute(eflag,vflag);
timer->stamp(TIME_BOND);
timer->stamp(Timer::BOND);
}
if (force->newton) {
comm->reverse_comm();
timer->stamp(TIME_COMM);
timer->stamp(Timer::COMM);
}
} else {
@ -388,14 +388,14 @@ void VerletSplit::run(int n)
if (force->kspace) {
timer->stamp();
force->kspace->compute(eflag,vflag);
timer->stamp(TIME_KSPACE);
timer->stamp(Timer::KSPACE);
}
// TIP4P PPPM puts forces on ghost atoms, so must reverse_comm()
if (tip4p_flag && force->newton) {
comm->reverse_comm();
timer->stamp(TIME_COMM);
timer->stamp(Timer::COMM);
}
}
@ -407,14 +407,16 @@ void VerletSplit::run(int n)
// all output
if (master) {
timer->stamp();
if (n_post_force) modify->post_force(vflag);
modify->final_integrate();
if (n_end_of_step) modify->end_of_step();
timer->stamp(Timer::MODIFY);
if (ntimestep == output->next) {
timer->stamp();
output->write(ntimestep);
timer->stamp(TIME_OUTPUT);
timer->stamp(Timer::OUTPUT);
}
}
}
@ -486,7 +488,7 @@ void VerletSplit::rk_setup()
atom->map_clear();
comm->borders();
if (triclinic) domain->lamda2x(atom->nlocal+atom->nghost);
timer->stamp(TIME_COMM);
timer->stamp(Timer::COMM);
}
}
}
@ -535,7 +537,7 @@ void VerletSplit::r2k_comm()
if (tip4p_flag && !master) {
timer->stamp();
comm->forward_comm();
timer->stamp(TIME_COMM);
timer->stamp(Timer::COMM);
}
}

36
src/USER-CUDA/verlet_cuda.cpp
View File

@ -679,7 +679,7 @@ void VerletCuda::run(int n)
my_gettime(CLOCK_REALTIME, &starttime);
timer->stamp();
comm->forward_comm(1);
timer->stamp(TIME_COMM);
timer->stamp(Timer::COMM);
my_gettime(CLOCK_REALTIME, &endtime);
cuda->shared_data.cuda_timings.comm_forward_total +=
endtime.tv_sec - starttime.tv_sec + 1.0 * (endtime.tv_nsec - starttime.tv_nsec) / 1000000000;
@ -699,7 +699,7 @@ void VerletCuda::run(int n)
//start force calculation asynchronus
cuda->shared_data.comm.comm_phase = 1;
force->pair->compute(eflag, vflag);
timer->stamp(TIME_PAIR);
timer->stamp(Timer::PAIR);
//CudaWrapper_Sync();
//download comm buffers from GPU, perform MPI communication and upload buffers again
@ -708,11 +708,11 @@ void VerletCuda::run(int n)
my_gettime(CLOCK_REALTIME, &endtime);
cuda->shared_data.cuda_timings.comm_forward_total +=
endtime.tv_sec - starttime.tv_sec + 1.0 * (endtime.tv_nsec - starttime.tv_nsec) / 1000000000;
timer->stamp(TIME_COMM);
timer->stamp(Timer::COMM);
//wait for force calculation
CudaWrapper_Sync();
timer->stamp(TIME_PAIR);
timer->stamp(Timer::PAIR);
//unpack communication buffers
my_gettime(CLOCK_REALTIME, &starttime);
@ -721,7 +721,7 @@ void VerletCuda::run(int n)
cuda->shared_data.cuda_timings.comm_forward_total +=
endtime.tv_sec - starttime.tv_sec + 1.0 * (endtime.tv_nsec - starttime.tv_nsec) / 1000000000;
timer->stamp(TIME_COMM);
timer->stamp(Timer::COMM);
MYDBG(printf("# CUDA VerletCuda::iterate: communicate done\n");)
cuda->shared_data.cuda_timings.test1 +=
endtotal.tv_sec - starttotal.tv_sec + 1.0 * (endtotal.tv_nsec - starttotal.tv_nsec) / 1000000000;
@ -732,7 +732,7 @@ void VerletCuda::run(int n)
my_gettime(CLOCK_REALTIME, &endtime);
cuda->shared_data.cuda_timings.comm_forward_total +=
endtime.tv_sec - starttime.tv_sec + 1.0 * (endtime.tv_nsec - starttime.tv_nsec) / 1000000000;
timer->stamp(TIME_COMM);
timer->stamp(Timer::COMM);
MYDBG(printf("# CUDA VerletCuda::iterate: communicate done\n");)
}
} else {
@ -822,7 +822,7 @@ void VerletCuda::run(int n)
cuda->shared_data.buffer_new = 2;
MYDBG(printf("# CUDA VerletCuda::iterate: neighbor build\n");)
timer->stamp(TIME_COMM);
timer->stamp(Timer::COMM);
my_gettime(CLOCK_REALTIME, &endtime);
cuda->shared_data.cuda_timings.test2 +=
endtime.tv_sec - starttime.tv_sec + 1.0 * (endtime.tv_nsec - starttime.tv_nsec) / 1000000000;
@ -830,7 +830,7 @@ void VerletCuda::run(int n)
//rebuild neighbor list
test_atom(testatom, "Pre Neighbor");
neighbor->build(0);
timer->stamp(TIME_NEIGHBOR);
timer->stamp(Timer::NEIGH);
MYDBG(printf("# CUDA VerletCuda::iterate: neighbor done\n");)
//if bonded interactions are used (in this case collect_forces_later is true), transfer data which only changes upon exchange/border routines from GPU to CPU
if(cuda->shared_data.pair.collect_forces_later) {
@ -917,7 +917,7 @@ void VerletCuda::run(int n)
if(not cuda->shared_data.pair.collect_forces_later)
CudaWrapper_Sync();
timer->stamp(TIME_PAIR);
timer->stamp(Timer::PAIR);
}
//calculate bonded interactions
@ -927,11 +927,11 @@ void VerletCuda::run(int n)
if(n_pre_force == 0) Verlet::force_clear();
else cuda->cu_f->downloadAsync(2);
timer->stamp(TIME_PAIR);
timer->stamp(Timer::PAIR);
if(neighbor->lastcall == update->ntimestep) {
neighbor->build_topology();
timer->stamp(TIME_NEIGHBOR);
timer->stamp(Timer::NEIGH);
}
test_atom(testatom, "pre bond force");
@ -944,7 +944,7 @@ void VerletCuda::run(int n)
if(force->improper) force->improper->compute(eflag, vflag);
timer->stamp(TIME_BOND);
timer->stamp(Timer::BOND);
}
//collect forces in case pair force and bonded interactions were overlapped, and either no KSPACE or a GPU KSPACE style is used
@ -969,7 +969,7 @@ void VerletCuda::run(int n)
if(vflag) cuda->cu_virial->download();
timer->stamp(TIME_PAIR);
timer->stamp(Timer::PAIR);
my_gettime(CLOCK_REALTIME, &endtime);
cuda->shared_data.cuda_timings.pair_force_collection +=
@ -987,7 +987,7 @@ void VerletCuda::run(int n)
if(n_pre_force == 0) Verlet::force_clear();
else cuda->cu_f->downloadAsync(2);
timer->stamp(TIME_PAIR);
timer->stamp(Timer::PAIR);
}
force->kspace->compute(eflag, vflag);
@ -995,7 +995,7 @@ void VerletCuda::run(int n)
if((not cuda->shared_data.pppm.cudable_force) && (not cuda->shared_data.pair.collect_forces_later))
cuda->uploadAll();
timer->stamp(TIME_KSPACE);
timer->stamp(Timer::KSPACE);
}
//collect forces in case pair forces and kspace was overlaped
@ -1018,7 +1018,7 @@ void VerletCuda::run(int n)
if(vflag) cuda->cu_virial->download();
timer->stamp(TIME_PAIR);
timer->stamp(Timer::PAIR);
my_gettime(CLOCK_REALTIME, &endtime);
cuda->shared_data.cuda_timings.pair_force_collection +=
@ -1028,7 +1028,7 @@ void VerletCuda::run(int n)
//send forces on ghost atoms back to other GPU: THIS SHOULD NEVER HAPPEN
if(force->newton) {
comm->reverse_comm();
timer->stamp(TIME_COMM);
timer->stamp(Timer::COMM);
}
test_atom(testatom, "post force");
@ -1054,7 +1054,7 @@ void VerletCuda::run(int n)
timer->stamp();
output->write(ntimestep);
timer->stamp(TIME_OUTPUT);
timer->stamp(Timer::OUTPUT);
}

8
src/USER-FEP/compute_fep.cpp
View File

@ -305,11 +305,11 @@ void ComputeFEP::compute_vector()
timer->stamp();
if (force->pair && force->pair->compute_flag) {
force->pair->compute(eflag,vflag);
timer->stamp(TIME_PAIR);
timer->stamp(Timer::PAIR);
}
if (chgflag && force->kspace && force->kspace->compute_flag) {
force->kspace->compute(eflag,vflag);
timer->stamp(TIME_KSPACE);
timer->stamp(Timer::KSPACE);
}
// accumulate force/energy/virial from /gpu pair styles
@ -322,11 +322,11 @@ void ComputeFEP::compute_vector()
timer->stamp();
if (force->pair && force->pair->compute_flag) {
force->pair->compute(eflag,vflag);
timer->stamp(TIME_PAIR);
timer->stamp(Timer::PAIR);
}
if (chgflag && force->kspace && force->kspace->compute_flag) {
force->kspace->compute(eflag,vflag);
timer->stamp(TIME_KSPACE);
timer->stamp(Timer::KSPACE);
}
// accumulate force/energy/virial from /gpu pair styles

41
src/USER-INTEL/verlet_intel.cpp
View File

@ -272,8 +272,10 @@ void VerletIntel::run(int n)
// initial time integration
timer->stamp();
modify->initial_integrate(vflag);
if (n_post_integrate) modify->post_integrate();
timer->stamp(Timer::MODIFY);
// regular communication vs neighbor list rebuild
@ -282,9 +284,13 @@ void VerletIntel::run(int n)
if (nflag == 0) {
timer->stamp();
comm->forward_comm();
timer->stamp(TIME_COMM);
timer->stamp(Timer::COMM);
} else {
if (n_pre_exchange) modify->pre_exchange();
if (n_pre_exchange) {
timer->stamp();
modify->pre_exchange();
timer->stamp(Timer::MODIFY);
}
if (triclinic) domain->x2lamda(atom->nlocal);
domain->pbc();
if (domain->box_change) {
@ -297,10 +303,13 @@ void VerletIntel::run(int n)
if (sortflag && ntimestep >= atom->nextsort) atom->sort();
comm->borders();
if (triclinic) domain->lamda2x(atom->nlocal+atom->nghost);
timer->stamp(TIME_COMM);
if (n_pre_neighbor) modify->pre_neighbor();
timer->stamp(Timer::COMM);
if (n_pre_neighbor) {
modify->pre_neighbor();
timer->stamp(Timer::MODIFY);
}
neighbor->build();
timer->stamp(TIME_NEIGHBOR);
timer->stamp(Timer::NEIGH);
}
// force computations
@ -309,13 +318,18 @@ void VerletIntel::run(int n)
// and Pair:ev_tally() needs to be called before any tallying
force_clear();
if (n_pre_force) modify->pre_force(vflag);
timer->stamp();
if (n_pre_force) {
modify->pre_force(vflag);
timer->stamp(Timer::MODIFY);
}
if (pair_compute_flag) {
force->pair->compute(eflag,vflag);
timer->stamp(TIME_PAIR);
timer->stamp(Timer::PAIR);
}
if (atom->molecular) {
@ -323,18 +337,18 @@ void VerletIntel::run(int n)
if (force->angle) force->angle->compute(eflag,vflag);
if (force->dihedral) force->dihedral->compute(eflag,vflag);
if (force->improper) force->improper->compute(eflag,vflag);
timer->stamp(TIME_BOND);
timer->stamp(Timer::BOND);
}
if (kspace_compute_flag) {
force->kspace->compute(eflag,vflag);
timer->stamp(TIME_KSPACE);
timer->stamp(Timer::KSPACE);
}
#ifdef _LMP_INTEL_OFFLOAD
if (sync_mode == 1) {
fix_intel->sync_coprocessor();
timer->stamp(TIME_PAIR);
timer->stamp(Timer::PAIR);
}
#endif
@ -342,13 +356,13 @@ void VerletIntel::run(int n)
if (force->newton) {
comm->reverse_comm();
timer->stamp(TIME_COMM);
timer->stamp(Timer::COMM);
}
#ifdef _LMP_INTEL_OFFLOAD
if (sync_mode == 2) {
fix_intel->sync_coprocessor();
timer->stamp(TIME_PAIR);
timer->stamp(Timer::PAIR);
}
#endif
@ -357,13 +371,14 @@ void VerletIntel::run(int n)
if (n_post_force) modify->post_force(vflag);
modify->final_integrate();
if (n_end_of_step) modify->end_of_step();
timer->stamp(Timer::MODIFY);
// all output
if (ntimestep == output->next) {
timer->stamp();
output->write(ntimestep);
timer->stamp(TIME_OUTPUT);
timer->stamp(Timer::OUTPUT);
}
}
}

30
src/USER-INTEL/verlet_split_intel.cpp
View File

@ -283,7 +283,7 @@ void VerletSplitIntel::run(int n)
MPI_Barrier(universe->uworld);
timer->init();
timer->barrier_start(TIME_LOOP);
timer->barrier_start();
// setup initial Rspace <-> Kspace comm params
@ -329,7 +329,7 @@ void VerletSplitIntel::run(int n)
if (nflag == 0) {
timer->stamp();
comm->forward_comm();
timer->stamp(TIME_COMM);
timer->stamp(Timer::COMM);
} else {
if (n_pre_exchange) modify->pre_exchange();
if (triclinic) domain->x2lamda(atom->nlocal);
@ -344,10 +344,10 @@ void VerletSplitIntel::run(int n)
if (sortflag && ntimestep >= atom->nextsort) atom->sort();
comm->borders();
if (triclinic) domain->lamda2x(atom->nlocal+atom->nghost);
timer->stamp(TIME_COMM);
timer->stamp(Timer::COMM);
if (n_pre_neighbor) modify->pre_neighbor();
neighbor->build();
timer->stamp(TIME_NEIGHBOR);
timer->stamp(Timer::NEIGH);
}
}
@ -367,7 +367,7 @@ void VerletSplitIntel::run(int n)
timer->stamp();
if (force->pair) {
force->pair->compute(eflag,vflag);
timer->stamp(TIME_PAIR);
timer->stamp(Timer::PAIR);
}
if (atom->molecular) {
@ -375,25 +375,25 @@ void VerletSplitIntel::run(int n)
if (force->angle) force->angle->compute(eflag,vflag);
if (force->dihedral) force->dihedral->compute(eflag,vflag);
if (force->improper) force->improper->compute(eflag,vflag);
timer->stamp(TIME_BOND);
timer->stamp(Timer::BOND);
}
#ifdef _LMP_INTEL_OFFLOAD
if (sync_mode == 1) {
fix_intel->sync_coprocessor();
timer->stamp(TIME_PAIR);
timer->stamp(Timer::PAIR);
}
#endif
if (force->newton) {
comm->reverse_comm();
timer->stamp(TIME_COMM);
timer->stamp(Timer::COMM);
}
#ifdef _LMP_INTEL_OFFLOAD
if (sync_mode == 2) {
fix_intel->sync_coprocessor();
timer->stamp(TIME_PAIR);
timer->stamp(Timer::PAIR);
}
#endif
@ -406,14 +406,14 @@ void VerletSplitIntel::run(int n)
if (force->kspace) {
timer->stamp();
force->kspace->compute(eflag,vflag);
timer->stamp(TIME_KSPACE);
timer->stamp(Timer::KSPACE);
}
// TIP4P PPPM puts forces on ghost atoms, so must reverse_comm()
if (tip4p_flag && force->newton) {
comm->reverse_comm();
timer->stamp(TIME_COMM);
timer->stamp(Timer::COMM);
}
}
@ -425,14 +425,16 @@ void VerletSplitIntel::run(int n)
// all output
if (master) {
timer->stamp();
if (n_post_force) modify->post_force(vflag);
modify->final_integrate();
if (n_end_of_step) modify->end_of_step();
timer->stamp(Timer::MODIFY);
if (ntimestep == output->next) {
timer->stamp();
output->write(ntimestep);
timer->stamp(TIME_OUTPUT);
timer->stamp(Timer::OUTPUT);
}
}
}
@ -504,7 +506,7 @@ void VerletSplitIntel::rk_setup()
atom->map_clear();
comm->borders();
if (triclinic) domain->lamda2x(atom->nlocal+atom->nghost);
timer->stamp(TIME_COMM);
timer->stamp(Timer::COMM);
}
}
}
@ -553,7 +555,7 @@ void VerletSplitIntel::r2k_comm()
if (tip4p_flag && !master) {
timer->stamp();
comm->forward_comm();
timer->stamp(TIME_COMM);
timer->stamp(Timer::COMM);
}
}

2
src/USER-OMP/angle_charmm_omp.cpp
View File

@ -61,6 +61,7 @@ void AngleCharmmOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (inum > 0) {
@ -77,6 +78,7 @@ void AngleCharmmOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
}
thr->timer(Timer::BOND);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/angle_class2_omp.cpp
View File

@ -61,6 +61,7 @@ void AngleClass2OMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (inum > 0) {
@ -77,6 +78,7 @@ void AngleClass2OMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
}
thr->timer(Timer::BOND);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/angle_cosine_delta_omp.cpp
View File

@ -61,6 +61,7 @@ void AngleCosineDeltaOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (inum > 0) {
@ -77,6 +78,7 @@ void AngleCosineDeltaOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
}
thr->timer(Timer::BOND);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/angle_cosine_omp.cpp
View File

@ -61,6 +61,7 @@ void AngleCosineOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (inum > 0) {
@ -77,6 +78,7 @@ void AngleCosineOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
}
thr->timer(Timer::BOND);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/angle_cosine_periodic_omp.cpp
View File

@ -63,6 +63,7 @@ void AngleCosinePeriodicOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (inum > 0) {
@ -79,6 +80,7 @@ void AngleCosinePeriodicOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
}
thr->timer(Timer::BOND);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/angle_cosine_shift_exp_omp.cpp
View File

@ -61,6 +61,7 @@ void AngleCosineShiftExpOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (inum > 0) {
@ -77,6 +78,7 @@ void AngleCosineShiftExpOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
}
thr->timer(Timer::BOND);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/angle_cosine_shift_omp.cpp
View File

@ -61,6 +61,7 @@ void AngleCosineShiftOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (inum > 0) {
@ -77,6 +78,7 @@ void AngleCosineShiftOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
}
thr->timer(Timer::BOND);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/angle_cosine_squared_omp.cpp
View File

@ -61,6 +61,7 @@ void AngleCosineSquaredOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (inum > 0) {
@ -77,6 +78,7 @@ void AngleCosineSquaredOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
}
thr->timer(Timer::BOND);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/angle_dipole_omp.cpp
View File

@ -65,6 +65,7 @@ void AngleDipoleOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (inum > 0) {
@ -73,6 +74,7 @@ void AngleDipoleOMP::compute(int eflag, int vflag)
else
eval<0>(ifrom, ito, thr);
}
thr->timer(Timer::BOND);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region

2
src/USER-OMP/angle_fourier_omp.cpp
View File

@ -61,6 +61,7 @@ void AngleFourierOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (inum > 0) {
@ -77,6 +78,7 @@ void AngleFourierOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
}
thr->timer(Timer::BOND);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/angle_fourier_simple_omp.cpp
View File

@ -61,6 +61,7 @@ void AngleFourierSimpleOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (inum > 0) {
@ -77,6 +78,7 @@ void AngleFourierSimpleOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
}
thr->timer(Timer::BOND);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/angle_harmonic_omp.cpp
View File

@ -61,6 +61,7 @@ void AngleHarmonicOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (inum > 0) {
@ -77,6 +78,7 @@ void AngleHarmonicOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
}
thr->timer(Timer::BOND);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/angle_quartic_omp.cpp
View File

@ -61,6 +61,7 @@ void AngleQuarticOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (inum > 0) {
@ -77,6 +78,7 @@ void AngleQuarticOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
}
thr->timer(Timer::BOND);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/angle_sdk_omp.cpp
View File

@ -63,6 +63,7 @@ void AngleSDKOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (inum > 0) {
@ -79,6 +80,7 @@ void AngleSDKOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
}
thr->timer(Timer::BOND);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/angle_table_omp.cpp
View File

@ -61,6 +61,7 @@ void AngleTableOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (inum > 0) {
@ -77,6 +78,7 @@ void AngleTableOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
}
thr->timer(Timer::BOND);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/bond_class2_omp.cpp
View File

@ -57,6 +57,7 @@ void BondClass2OMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (inum > 0) {
@ -73,6 +74,7 @@ void BondClass2OMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
}
thr->timer(Timer::BOND);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/bond_fene_expand_omp.cpp
View File

@ -58,6 +58,7 @@ void BondFENEExpandOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (inum > 0) {
@ -74,6 +75,7 @@ void BondFENEExpandOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
}
thr->timer(Timer::BOND);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/bond_fene_omp.cpp
View File

@ -58,6 +58,7 @@ void BondFENEOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (inum > 0) {
@ -74,6 +75,7 @@ void BondFENEOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
}
thr->timer(Timer::BOND);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/bond_harmonic_omp.cpp
View File

@ -56,6 +56,7 @@ void BondHarmonicOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (inum > 0) {
@ -72,6 +73,7 @@ void BondHarmonicOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
}
thr->timer(Timer::BOND);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/bond_harmonic_shift_cut_omp.cpp
View File

@ -56,6 +56,7 @@ void BondHarmonicShiftCutOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (inum > 0) {
@ -72,6 +73,7 @@ void BondHarmonicShiftCutOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
}
thr->timer(Timer::BOND);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/bond_harmonic_shift_omp.cpp
View File

@ -56,6 +56,7 @@ void BondHarmonicShiftOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (inum > 0) {
@ -72,6 +73,7 @@ void BondHarmonicShiftOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
}
thr->timer(Timer::BOND);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/bond_morse_omp.cpp
View File

@ -56,6 +56,7 @@ void BondMorseOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (inum > 0) {
@ -72,6 +73,7 @@ void BondMorseOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
}
thr->timer(Timer::BOND);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/bond_nonlinear_omp.cpp
View File

@ -56,6 +56,7 @@ void BondNonlinearOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (inum > 0) {
@ -72,6 +73,7 @@ void BondNonlinearOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
}
thr->timer(Timer::BOND);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/bond_quartic_omp.cpp
View File

@ -62,6 +62,7 @@ void BondQuarticOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (inum > 0) {
@ -78,6 +79,7 @@ void BondQuarticOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
}
thr->timer(Timer::BOND);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/bond_table_omp.cpp
View File

@ -56,6 +56,7 @@ void BondTableOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (inum > 0) {
@ -72,6 +73,7 @@ void BondTableOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
}
thr->timer(Timer::BOND);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/dihedral_charmm_omp.cpp
View File

@ -67,6 +67,7 @@ void DihedralCharmmOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (inum > 0) {
@ -83,6 +84,7 @@ void DihedralCharmmOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
}
thr->timer(Timer::BOND);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/dihedral_class2_omp.cpp
View File

@ -60,6 +60,7 @@ void DihedralClass2OMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (inum > 0) {
@ -76,6 +77,7 @@ void DihedralClass2OMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
}
thr->timer(Timer::BOND);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/dihedral_cosine_shift_exp_omp.cpp
View File

@ -60,6 +60,7 @@ void DihedralCosineShiftExpOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (inum > 0) {
@ -76,6 +77,7 @@ void DihedralCosineShiftExpOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
}
thr->timer(Timer::BOND);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/dihedral_fourier_omp.cpp
View File

@ -61,6 +61,7 @@ void DihedralFourierOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (inum > 0) {
@ -77,6 +78,7 @@ void DihedralFourierOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
}
thr->timer(Timer::BOND);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/dihedral_harmonic_omp.cpp
View File

@ -60,6 +60,7 @@ void DihedralHarmonicOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (inum > 0) {
@ -76,6 +77,7 @@ void DihedralHarmonicOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
}
thr->timer(Timer::BOND);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/dihedral_helix_omp.cpp
View File

@ -63,6 +63,7 @@ void DihedralHelixOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (inum > 0) {
@ -79,6 +80,7 @@ void DihedralHelixOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
}
thr->timer(Timer::BOND);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/dihedral_multi_harmonic_omp.cpp
View File

@ -60,6 +60,7 @@ void DihedralMultiHarmonicOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (inum > 0) {
@ -76,6 +77,7 @@ void DihedralMultiHarmonicOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
}
thr->timer(Timer::BOND);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/dihedral_nharmonic_omp.cpp
View File

@ -60,6 +60,7 @@ void DihedralNHarmonicOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (inum > 0) {
@ -76,6 +77,7 @@ void DihedralNHarmonicOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
}
thr->timer(Timer::BOND);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/dihedral_opls_omp.cpp
View File

@ -61,6 +61,7 @@ void DihedralOPLSOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (inum > 0) {
@ -77,6 +78,7 @@ void DihedralOPLSOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
}
thr->timer(Timer::BOND);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/dihedral_quadratic_omp.cpp
View File

@ -61,6 +61,7 @@ void DihedralQuadraticOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (inum > 0) {
@ -77,6 +78,7 @@ void DihedralQuadraticOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
}
thr->timer(Timer::BOND);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/dihedral_table_omp.cpp
View File

@ -128,6 +128,7 @@ void DihedralTableOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (inum > 0) {
@ -144,6 +145,7 @@ void DihedralTableOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
}
thr->timer(Timer::BOND);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/ewald_omp.cpp
View File

@ -118,6 +118,7 @@ void EwaldOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, nlocal, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, 0, NULL, NULL, thr);
for (i = ifrom; i < ito; i++) {
@ -205,6 +206,7 @@ void EwaldOMP::compute(int eflag, int vflag)
for (j = 0; j < 6; j++) vatom[i][j] *= q[i]*qscale;
}
thr->timer(Timer::KSPACE);
reduce_thr(this, eflag,vflag,thr);
} // end of omp parallel region

31
src/USER-OMP/fix_omp.cpp
View File

@ -26,6 +26,7 @@
#include "update.h"
#include "integrate.h"
#include "min.h"
#include "timer.h"
#include "fix_omp.h"
#include "thr_data.h"
@ -65,7 +66,7 @@ static int get_tid()
/* ---------------------------------------------------------------------- */
FixOMP::FixOMP(LAMMPS *lmp, int narg, char **arg)
FixOMP::FixOMP(LAMMPS *lmp, int narg, char **arg)
: Fix(lmp, narg, arg),
thr(NULL), last_omp_style(NULL), last_pair_hybrid(NULL),
_nthr(-1), _neighbor(true), _mixed(false), _reduced(true)
@ -130,7 +131,7 @@ FixOMP::FixOMP(LAMMPS *lmp, int narg, char **arg)
fprintf(screen,"using %s neighbor list subroutines\n", nmode);
fprintf(screen,"prefer %s precision OpenMP force kernels\n", kmode);
}
if (logfile) {
if (reset_thr)
fprintf(logfile,"set %d OpenMP thread(s) per MPI task\n", nthreads);
@ -146,11 +147,12 @@ FixOMP::FixOMP(LAMMPS *lmp, int narg, char **arg)
thr = new ThrData *[nthreads];
_nthr = nthreads;
#if defined(_OPENMP)
#pragma omp parallel default(none)
#pragma omp parallel default(none) shared(lmp)
#endif
{
const int tid = get_tid();
thr[tid] = new ThrData(tid);
Timer *t = new Timer(lmp);
thr[tid] = new ThrData(tid,t);
}
}
@ -196,11 +198,17 @@ int FixOMP::setmask()
void FixOMP::init()
{
// USER-OMP package cannot be used with atom_style template
if (atom->molecular == 2)
if (atom->molecular == 2)
error->all(FLERR,"USER-OMP package does not (yet) work with "
"atom_style template");
// reset per thread timer
for (int i=0; i < comm->nthreads; ++i) {
thr[i]->_timer_active=1;
thr[i]->timer(Timer::RESET);
thr[i]->_timer_active=-1;
}
if ((strstr(update->integrate_style,"respa") != NULL)
&& (strstr(update->integrate_style,"respa/omp") == NULL))
error->all(FLERR,"Need to use respa/omp for r-RESPA with /omp styles");
@ -272,7 +280,7 @@ void FixOMP::init()
CheckStyleForOMP(improper);
CheckHybridForOMP(improper,Improper);
}
if (kspace_split >= 0) {
CheckStyleForOMP(kspace);
}
@ -325,6 +333,15 @@ void FixOMP::set_neighbor_omp()
/* ---------------------------------------------------------------------- */
void FixOMP::setup(int)
{
// we are post the force compute in setup. turn on timers
for (int i=0; i < comm->nthreads; ++i)
thr[i]->_timer_active=0;
}
/* ---------------------------------------------------------------------- */
// adjust size and clear out per thread accumulator arrays
void FixOMP::pre_force(int)
{

2
src/USER-OMP/fix_omp.h
View File

@ -36,6 +36,8 @@ class FixOMP : public Fix {
virtual ~FixOMP();
virtual int setmask();
virtual void init();
virtual void setup(int);
virtual void min_setup(int flag) { setup(flag); }
virtual void pre_force(int);
virtual void setup_pre_force(int vflag) { pre_force(vflag); }

2
src/USER-OMP/improper_class2_omp.cpp
View File

@ -60,6 +60,7 @@ void ImproperClass2OMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (inum > 0) {
@ -76,6 +77,7 @@ void ImproperClass2OMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
}
thr->timer(Timer::BOND);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/improper_cossq_omp.cpp
View File

@ -60,6 +60,7 @@ void ImproperCossqOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (inum > 0) {
@ -76,6 +77,7 @@ void ImproperCossqOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
}
thr->timer(Timer::BOND);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/improper_cvff_omp.cpp
View File

@ -60,6 +60,7 @@ void ImproperCvffOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (inum > 0) {
@ -76,6 +77,7 @@ void ImproperCvffOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
}
thr->timer(Timer::BOND);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

14
src/USER-OMP/improper_fourier_omp.cpp
View File

@ -60,6 +60,7 @@ void ImproperFourierOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (inum > 0) {
@ -76,6 +77,7 @@ void ImproperFourierOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
}
thr->timer(Timer::BOND);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}
@ -115,18 +117,18 @@ void ImproperFourierOMP::eval(int nfrom, int nto, ThrData * const thr)
vb3z = x[i4][2] - x[i1][2];
add1_thr<EVFLAG,EFLAG,NEWTON_BOND>(i1,i2,i3,i4,type,
vb1x,vb1y,vb1z,
vb2x,vb2y,vb2z,
vb1x,vb1y,vb1z,
vb2x,vb2y,vb2z,
vb3x,vb3y,vb3z,thr);
if ( all[type] ) {
add1_thr<EVFLAG,EFLAG,NEWTON_BOND>(i1,i4,i2,i3,type,
vb3x,vb3y,vb3z,
vb1x,vb1y,vb1z,
vb2x,vb2y,vb2z,thr);
vb1x,vb1y,vb1z,
vb2x,vb2y,vb2z,thr);
add1_thr<EVFLAG,EFLAG,NEWTON_BOND>(i1,i3,i4,i2,type,
vb2x,vb2y,vb2z,
vb2x,vb2y,vb2z,
vb3x,vb3y,vb3z,
vb1x,vb1y,vb1z,thr);
vb1x,vb1y,vb1z,thr);
}
}
}

2
src/USER-OMP/improper_harmonic_omp.cpp
View File

@ -60,6 +60,7 @@ void ImproperHarmonicOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (inum > 0) {
@ -76,6 +77,7 @@ void ImproperHarmonicOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
}
thr->timer(Timer::BOND);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/improper_ring_omp.cpp
View File

@ -62,6 +62,7 @@ void ImproperRingOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (inum > 0) {
@ -77,6 +78,7 @@ void ImproperRingOMP::compute(int eflag, int vflag)
if (force->newton_bond) eval<0,0,1>(ifrom, ito, thr);
else eval<0,0,0>(ifrom, ito, thr);
}
thr->timer(Timer::BOND);
reduce_thr(this, eflag, vflag, thr);
}
} // end of omp parallel region

2
src/USER-OMP/improper_umbrella_omp.cpp
View File

@ -60,6 +60,7 @@ void ImproperUmbrellaOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (inum > 0) {
@ -76,6 +77,7 @@ void ImproperUmbrellaOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
}
thr->timer(Timer::BOND);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

1
src/USER-OMP/msm_cg_omp.cpp
View File

@ -312,6 +312,7 @@ void MSMCGOMP::compute(int eflag, int vflag)
const int tid = 0;
#endif
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

8
src/USER-OMP/msm_omp.cpp
View File

@ -66,6 +66,7 @@ void MSMOMP::compute(int eflag, int vflag)
const int tid = 0;
#endif
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}
@ -74,12 +75,12 @@ void MSMOMP::compute(int eflag, int vflag)
MSM direct part procedure for intermediate grid levels
------------------------------------------------------------------------- */
void MSMOMP::direct(int n)
void MSMOMP::direct(int n)
{
// zero out electric potential
memset(&(egrid[n][nzlo_out[n]][nylo_out[n]][nxlo_out[n]]),0,ngrid[n]*sizeof(double));
// zero out virial
if (vflag_atom) {
@ -169,6 +170,8 @@ void MSMOMP::direct_eval(const int nn)
int i,ifrom,ito,tid,icx,icy,icz,ix,iy,iz,k;
loop_setup_thr(ifrom, ito, tid, inum, comm->nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
for (i = ifrom; i < ito; ++i) {
@ -298,6 +301,7 @@ void MSMOMP::direct_eval(const int nn)
}
}
}
thr->timer(Timer::KSPACE);
} // end of omp parallel region
if (EFLAG_GLOBAL || VFLAG_GLOBAL) {

22
src/USER-OMP/neighbor_omp.h
View File

@ -18,14 +18,20 @@
#include <omp.h>
#endif
#include "modify.h"
#include "timer.h"
#include "fix_omp.h"
#include "thr_data.h"
namespace LAMMPS_NS {
// these macros hide some ugly and redundant OpenMP related stuff
#if defined(_OPENMP)
// make sure we have at least one page for each thread
#define NEIGH_OMP_INIT \
const int nthreads = comm->nthreads;
#define NEIGH_OMP_INIT \
const int nthreads = comm->nthreads; \
const int ifix = modify->find_fix("package_omp")
// get thread id and then assign each thread a fixed chunk of atoms
#define NEIGH_OMP_SETUP(num) \
@ -34,14 +40,18 @@ namespace LAMMPS_NS {
const int idelta = 1 + num/nthreads; \
const int ifrom = tid*idelta; \
const int ito = ((ifrom + idelta) > num) \
? num : (ifrom+idelta);
? num : (ifrom+idelta); \
FixOMP *fix = static_cast<FixOMP *>(modify->fix[ifix]); \
ThrData *thr = fix->get_thr(tid); \
thr->timer(Timer::START);
#define NEIGH_OMP_CLOSE }
#define NEIGH_OMP_CLOSE \
thr->timer(Timer::NEIGH); \
}
#else /* !defined(_OPENMP) */
#define NEIGH_OMP_INIT \
const int nthreads = comm->nthreads;
#define NEIGH_OMP_INIT
#define NEIGH_OMP_SETUP(num) \
const int tid = 0; \

4
src/USER-OMP/pair_adp_omp.cpp
View File

@ -71,6 +71,7 @@ void PairADPOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (force->newton_pair)
@ -91,6 +92,7 @@ void PairADPOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
thr->timer(Timer::PAIR);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}
@ -200,6 +202,7 @@ void PairADPOMP::eval(int iifrom, int iito, ThrData * const thr)
if (NEWTON_PAIR) {
// reduce per thread density
thr->timer(Timer::PAIR);
data_reduce_thr(&(rho[0]), nall, comm->nthreads, 1, tid);
data_reduce_thr(&(mu[0][0]), nall, comm->nthreads, 3, tid);
data_reduce_thr(&(lambda[0][0]), nall, comm->nthreads, 6, tid);
@ -217,6 +220,7 @@ void PairADPOMP::eval(int iifrom, int iito, ThrData * const thr)
} else {
// reduce per thread density
thr->timer(Timer::PAIR);
data_reduce_thr(&(rho[0]), nlocal, comm->nthreads, 1, tid);
data_reduce_thr(&(mu[0][0]), nlocal, comm->nthreads, 3, tid);
data_reduce_thr(&(lambda[0][0]), nlocal, comm->nthreads, 6, tid);

2
src/USER-OMP/pair_airebo_omp.cpp
View File

@ -65,12 +65,14 @@ void PairAIREBOOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
FREBO_thr(ifrom,ito,evflag,eflag,vflag_atom,thr);
if (ljflag) FLJ_thr(ifrom,ito,evflag,eflag,vflag_atom,thr);
if (torflag) TORSION_thr(ifrom,ito,evflag,eflag,thr);
thr->timer(Timer::PAIR);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/pair_beck_omp.cpp
View File

@ -54,6 +54,7 @@ void PairBeckOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (evflag) {
@ -69,6 +70,7 @@ void PairBeckOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
thr->timer(Timer::PAIR);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/pair_born_coul_long_omp.cpp
View File

@ -60,6 +60,7 @@ void PairBornCoulLongOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (evflag) {
@ -75,6 +76,7 @@ void PairBornCoulLongOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
thr->timer(Timer::PAIR);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/pair_born_coul_msm_omp.cpp
View File

@ -57,6 +57,7 @@ void PairBornCoulMSMOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (evflag) {
@ -72,6 +73,7 @@ void PairBornCoulMSMOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
thr->timer(Timer::PAIR);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/pair_born_coul_wolf_omp.cpp
View File

@ -54,6 +54,7 @@ void PairBornCoulWolfOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (evflag) {
@ -69,6 +70,7 @@ void PairBornCoulWolfOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
thr->timer(Timer::PAIR);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/pair_born_omp.cpp
View File

@ -52,6 +52,7 @@ void PairBornOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (evflag) {
@ -67,6 +68,7 @@ void PairBornOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
thr->timer(Timer::PAIR);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/pair_brownian_omp.cpp
View File

@ -132,6 +132,7 @@ void PairBrownianOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
// generate a random number generator instance for
@ -157,6 +158,7 @@ void PairBrownianOMP::compute(int eflag, int vflag)
}
}
thr->timer(Timer::PAIR);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/pair_brownian_poly_omp.cpp
View File

@ -132,6 +132,7 @@ void PairBrownianPolyOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
// generate a random number generator instance for
@ -150,6 +151,7 @@ void PairBrownianPolyOMP::compute(int eflag, int vflag)
else eval<0,0>(ifrom, ito, thr);
}
thr->timer(Timer::PAIR);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/pair_buck_coul_cut_omp.cpp
View File

@ -52,6 +52,7 @@ void PairBuckCoulCutOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (evflag) {
@ -66,6 +67,7 @@ void PairBuckCoulCutOMP::compute(int eflag, int vflag)
if (force->newton_pair) eval<0,0,1>(ifrom, ito, thr);
else eval<0,0,0>(ifrom, ito, thr);
}
thr->timer(Timer::PAIR);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/pair_buck_coul_long_omp.cpp
View File

@ -60,6 +60,7 @@ void PairBuckCoulLongOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (evflag) {
@ -76,6 +77,7 @@ void PairBuckCoulLongOMP::compute(int eflag, int vflag)
}
// reduce per thread forces into global force array.
thr->timer(Timer::PAIR);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/pair_buck_coul_msm_omp.cpp
View File

@ -57,6 +57,7 @@ void PairBuckCoulMSMOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (evflag) {
@ -73,6 +74,7 @@ void PairBuckCoulMSMOMP::compute(int eflag, int vflag)
}
// reduce per thread forces into global force array.
thr->timer(Timer::PAIR);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

12
src/USER-OMP/pair_buck_long_coul_long_omp.cpp
View File

@ -64,6 +64,7 @@ void PairBuckLongCoulLongOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (order6) {
@ -302,8 +303,9 @@ void PairBuckLongCoulLongOMP::compute(int eflag, int vflag)
}
}
}
}
}
thr->timer(Timer::PAIR);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}
@ -326,8 +328,10 @@ void PairBuckLongCoulLongOMP::compute_inner()
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(0, 0, nall, 0, 0, thr);
eval_inner(ifrom, ito, thr);
thr->timer(Timer::PAIR);
} // end of omp parallel region
}
@ -349,8 +353,10 @@ void PairBuckLongCoulLongOMP::compute_middle()
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(0, 0, nall, 0, 0, thr);
eval_middle(ifrom, ito, thr);
thr->timer(Timer::PAIR);
} // end of omp parallel region
}
@ -377,6 +383,7 @@ void PairBuckLongCoulLongOMP::compute_outer(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (order6) {
@ -615,8 +622,9 @@ void PairBuckLongCoulLongOMP::compute_outer(int eflag, int vflag)
}
}
}
}
}
thr->timer(Timer::PAIR);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/pair_buck_omp.cpp
View File

@ -52,6 +52,7 @@ void PairBuckOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (evflag) {
@ -67,6 +68,7 @@ void PairBuckOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
thr->timer(Timer::PAIR);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

24
src/USER-OMP/pair_cdeam_omp.cpp
View File

@ -86,6 +86,7 @@ void PairCDEAMOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (force->newton_pair)
@ -128,12 +129,15 @@ void PairCDEAMOMP::compute(int eflag, int vflag)
break;
default:
{
#if defined(_OPENMP)
#pragma omp master
#endif
error->all(FLERR,"unsupported eam/cd pair style variant");
error->all(FLERR,"unsupported eam/cd pair style variant");
}
}
thr->timer(Timer::PAIR);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}
@ -227,6 +231,7 @@ void PairCDEAMOMP::eval(int iifrom, int iito, ThrData * const thr)
if (NEWTON_PAIR) {
// reduce per thread density
thr->timer(Timer::PAIR);
data_reduce_thr(rho, nall, nthreads, 1, tid);
data_reduce_thr(rhoB, nall, nthreads, 1, tid);
if (CDEAMVERSION==1)
@ -246,6 +251,7 @@ void PairCDEAMOMP::eval(int iifrom, int iito, ThrData * const thr)
} else {
// reduce per thread density
thr->timer(Timer::PAIR);
data_reduce_thr(rho, nlocal, nthreads, 1, tid);
data_reduce_thr(rhoB, nlocal, nthreads, 1, tid);
if (CDEAMVERSION==1)
@ -346,6 +352,7 @@ void PairCDEAMOMP::eval(int iifrom, int iito, ThrData * const thr)
}
if (NEWTON_PAIR) {
thr->timer(Timer::PAIR);
data_reduce_thr(D_values, nall, nthreads, 1, tid);
// wait until reduction is complete
@ -361,6 +368,7 @@ void PairCDEAMOMP::eval(int iifrom, int iito, ThrData * const thr)
sync_threads();
} else {
thr->timer(Timer::PAIR);
data_reduce_thr(D_values, nlocal, nthreads, 1, tid);
// wait until reduction is complete
@ -410,7 +418,9 @@ void PairCDEAMOMP::eval(int iifrom, int iito, ThrData * const thr)
D_i = D_values[i] * h_prime_i / (2.0 * rho[i] * rho[i]);
} else if(CDEAMVERSION == 2) {
D_i = D_values[i];
} else ASSERT(false);
} else {
ASSERT(false);
}
}
for(jj = 0; jj < jnum; jj++) {
@ -456,8 +466,9 @@ void PairCDEAMOMP::eval(int iifrom, int iito, ThrData * const thr)
D_j = D_values[j] * h_prime_j / (2.0 * rho[j] * rho[j]);
} else if(CDEAMVERSION == 2) {
D_j = D_values[j];
} else ASSERT(false);
} else {
ASSERT(false);
}
double t2 = -rhoB[j];
if(itype == speciesB) t2 += rho[j];
fpair += D_j * rhoip * t2;
@ -491,8 +502,9 @@ void PairCDEAMOMP::eval(int iifrom, int iito, ThrData * const thr)
double x_ij = 0.5 * (x_i + x_j);
// Calculate h(x_ij) polynomial function.
h = evalH(x_ij);
} else ASSERT(false);
} else {
ASSERT(false);
}
fpair += h * phip;
phi *= h;
}

2
src/USER-OMP/pair_colloid_omp.cpp
View File

@ -55,6 +55,7 @@ void PairColloidOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (evflag) {
@ -70,6 +71,7 @@ void PairColloidOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
thr->timer(Timer::PAIR);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/pair_comb_omp.cpp
View File

@ -61,6 +61,7 @@ void PairCombOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (evflag) {
@ -73,6 +74,7 @@ void PairCombOMP::compute(int eflag, int vflag)
}
} else eval<0,0,0>(ifrom, ito, thr);
thr->timer(Timer::PAIR);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/pair_coul_cut_omp.cpp
View File

@ -52,6 +52,7 @@ void PairCoulCutOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (evflag) {
@ -67,6 +68,7 @@ void PairCoulCutOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
thr->timer(Timer::PAIR);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/pair_coul_cut_soft_omp.cpp
View File

@ -52,6 +52,7 @@ void PairCoulCutSoftOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (evflag) {
@ -67,6 +68,7 @@ void PairCoulCutSoftOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
thr->timer(Timer::PAIR);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/pair_coul_debye_omp.cpp
View File

@ -52,6 +52,7 @@ void PairCoulDebyeOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (evflag) {
@ -67,6 +68,7 @@ void PairCoulDebyeOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
thr->timer(Timer::PAIR);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/pair_coul_diel_omp.cpp
View File

@ -52,6 +52,7 @@ void PairCoulDielOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (evflag) {
@ -67,6 +68,7 @@ void PairCoulDielOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
thr->timer(Timer::PAIR);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/pair_coul_dsf_omp.cpp
View File

@ -62,6 +62,7 @@ void PairCoulDSFOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (evflag) {
@ -77,6 +78,7 @@ void PairCoulDSFOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
thr->timer(Timer::PAIR);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/pair_coul_long_omp.cpp
View File

@ -61,6 +61,7 @@ void PairCoulLongOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (evflag) {
@ -76,6 +77,7 @@ void PairCoulLongOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
thr->timer(Timer::PAIR);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/pair_coul_long_soft_omp.cpp
View File

@ -60,6 +60,7 @@ void PairCoulLongSoftOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (evflag) {
@ -75,6 +76,7 @@ void PairCoulLongSoftOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
thr->timer(Timer::PAIR);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/pair_coul_msm_omp.cpp
View File

@ -58,6 +58,7 @@ void PairCoulMSMOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (evflag) {
@ -73,6 +74,7 @@ void PairCoulMSMOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
thr->timer(Timer::PAIR);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/pair_coul_wolf_omp.cpp
View File

@ -54,6 +54,7 @@ void PairCoulWolfOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (evflag) {
@ -69,6 +70,7 @@ void PairCoulWolfOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
thr->timer(Timer::PAIR);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/pair_dpd_omp.cpp
View File

@ -77,6 +77,7 @@ void PairDPDOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
// generate a random number generator instance for
@ -98,6 +99,7 @@ void PairDPDOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
thr->timer(Timer::PAIR);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/pair_dpd_tstat_omp.cpp
View File

@ -77,6 +77,7 @@ void PairDPDTstatOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
// generate a random number generator instance for
@ -98,6 +99,7 @@ void PairDPDTstatOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
thr->timer(Timer::PAIR);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

4
src/USER-OMP/pair_eam_omp.cpp
View File

@ -66,6 +66,7 @@ void PairEAMOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (force->newton_pair)
@ -86,6 +87,7 @@ void PairEAMOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
thr->timer(Timer::PAIR);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}
@ -162,6 +164,7 @@ void PairEAMOMP::eval(int iifrom, int iito, ThrData * const thr)
if (NEWTON_PAIR) {
// reduce per thread density
thr->timer(Timer::PAIR);
data_reduce_thr(rho, nall, nthreads, 1, tid);
// wait until reduction is complete
@ -176,6 +179,7 @@ void PairEAMOMP::eval(int iifrom, int iito, ThrData * const thr)
sync_threads();
} else {
thr->timer(Timer::PAIR);
data_reduce_thr(rho, nlocal, nthreads, 1, tid);
// wait until reduction is complete

2
src/USER-OMP/pair_edip_omp.cpp
View File

@ -59,6 +59,7 @@ void PairEDIPOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (evflag) {
@ -71,6 +72,7 @@ void PairEDIPOMP::compute(int eflag, int vflag)
}
} else eval<0,0,0>(ifrom, ito, thr);
thr->timer(Timer::PAIR);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

6
src/USER-OMP/pair_eim_omp.cpp
View File

@ -66,6 +66,7 @@ void PairEIMOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (force->newton_pair)
@ -86,6 +87,7 @@ void PairEIMOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
thr->timer(Timer::PAIR);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}
@ -162,6 +164,7 @@ void PairEIMOMP::eval(int iifrom, int iito, ThrData * const thr)
// communicate and sum densities
if (NEWTON_PAIR) {
// reduce per thread density
thr->timer(Timer::PAIR);
data_reduce_thr(rho, nall, nthreads, 1, tid);
// wait until reduction is complete
@ -176,6 +179,7 @@ void PairEIMOMP::eval(int iifrom, int iito, ThrData * const thr)
}
} else {
thr->timer(Timer::PAIR);
data_reduce_thr(rho, nlocal, nthreads, 1, tid);
// wait until reduction is complete
@ -234,6 +238,7 @@ void PairEIMOMP::eval(int iifrom, int iito, ThrData * const thr)
// communicate and sum modified densities
if (NEWTON_PAIR) {
// reduce per thread density
thr->timer(Timer::PAIR);
data_reduce_thr(fp, nall, nthreads, 1, tid);
// wait until reduction is complete
@ -248,6 +253,7 @@ void PairEIMOMP::eval(int iifrom, int iito, ThrData * const thr)
}
} else {
thr->timer(Timer::PAIR);
data_reduce_thr(fp, nlocal, nthreads, 1, tid);
// wait until reduction is complete

2
src/USER-OMP/pair_gauss_cut_omp.cpp
View File

@ -52,6 +52,7 @@ void PairGaussCutOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (evflag) {
@ -67,6 +68,7 @@ void PairGaussCutOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
thr->timer(Timer::PAIR);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/pair_gauss_omp.cpp
View File

@ -54,6 +54,7 @@ void PairGaussOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (evflag) {
@ -69,6 +70,7 @@ void PairGaussOMP::compute(int eflag, int vflag)
else occ = eval<0,0,0>(ifrom, ito, thr);
}
thr->timer(Timer::PAIR);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region

2
src/USER-OMP/pair_gayberne_omp.cpp
View File

@ -54,6 +54,7 @@ void PairGayBerneOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (evflag) {
@ -69,6 +70,7 @@ void PairGayBerneOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
thr->timer(Timer::PAIR);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

2
src/USER-OMP/pair_gran_hertz_history_omp.cpp
View File

@ -77,6 +77,7 @@ void PairGranHertzHistoryOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (evflag)
@ -86,6 +87,7 @@ void PairGranHertzHistoryOMP::compute(int eflag, int vflag)
if (shearupdate) eval<0,1>(ifrom, ito, thr);
else eval<0,0>(ifrom, ito, thr);
thr->timer(Timer::PAIR);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}

4
src/USER-OMP/pair_gran_hooke_history_omp.cpp
View File

@ -79,6 +79,7 @@ void PairGranHookeHistoryOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (evflag)
@ -88,6 +89,7 @@ void PairGranHookeHistoryOMP::compute(int eflag, int vflag)
if (shearupdate) eval<0,1>(ifrom, ito, thr);
else eval<0,0>(ifrom, ito, thr);
thr->timer(Timer::PAIR);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}
@ -95,7 +97,7 @@ void PairGranHookeHistoryOMP::compute(int eflag, int vflag)
template <int EVFLAG, int SHEARUPDATE>
void PairGranHookeHistoryOMP::eval(int iifrom, int iito, ThrData * const thr)
{
int i,j,ii,jj,jnum,itype,jtype;
int i,j,ii,jj,jnum;
double xtmp,ytmp,ztmp,delx,dely,delz,fx,fy,fz;
double myshear[3];
double radi,radj,radsum,rsq,r,rinv,rsqinv;

4
src/USER-OMP/pair_gran_hooke_omp.cpp
View File

@ -74,6 +74,7 @@ void PairGranHookeOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (evflag)
@ -83,6 +84,7 @@ void PairGranHookeOMP::compute(int eflag, int vflag)
if (force->newton_pair) eval<0,1>(ifrom, ito, thr);
else eval<0,0>(ifrom, ito, thr);
thr->timer(Timer::PAIR);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region
}
@ -90,7 +92,7 @@ void PairGranHookeOMP::compute(int eflag, int vflag)
template <int EVFLAG, int NEWTON_PAIR>
void PairGranHookeOMP::eval(int iifrom, int iito, ThrData * const thr)
{
int i,j,ii,jj,jnum,itype,jtype;
int i,j,ii,jj,jnum;
double xtmp,ytmp,ztmp,delx,dely,delz,fx,fy,fz;
double radi,radj,radsum,rsq,r,rinv,rsqinv;
double vr1,vr2,vr3,vnnr,vn1,vn2,vn3,vt1,vt2,vt3;

2
src/USER-OMP/pair_hbond_dreiding_lj_omp.cpp
View File

@ -83,6 +83,7 @@ void PairHbondDreidingLJOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (evflag) {
@ -98,6 +99,7 @@ void PairHbondDreidingLJOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
thr->timer(Timer::PAIR);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region

2
src/USER-OMP/pair_hbond_dreiding_morse_omp.cpp
View File

@ -83,6 +83,7 @@ void PairHbondDreidingMorseOMP::compute(int eflag, int vflag)
loop_setup_thr(ifrom, ito, tid, inum, nthreads);
ThrData *thr = fix->get_thr(tid);
thr->timer(Timer::START);
ev_setup_thr(eflag, vflag, nall, eatom, vatom, thr);
if (evflag) {
@ -98,6 +99,7 @@ void PairHbondDreidingMorseOMP::compute(int eflag, int vflag)
else eval<0,0,0>(ifrom, ito, thr);
}
thr->timer(Timer::PAIR);
reduce_thr(this, eflag, vflag, thr);
} // end of omp parallel region

Some files were not shown because too many files have changed in this diff Show More