Homec4science
Diffusion akantu (bugfixes/flatten-internal-that-are-missing) Browse

rAKA/src/synchronizer/grid_synchronizer.cc
54e599859086bugfixes/flatten…

Last Change
Raw File

grid_synchronizer.cc
View Options

/**
* @file grid_synchronizer.cc
*
* @author Nicolas Richart <nicolas.richart@epfl.ch>
*
* @date creation: Mon Oct 03 2011
* @date last modification: Thu Jun 05 2014
*
* @brief implementation of the grid synchronizer
*
* @section LICENSE
*
* Copyright (©) 2014 EPFL (Ecole Polytechnique Fédérale de Lausanne)
* Laboratory (LSMS - Laboratoire de Simulation en Mécanique des Solides)
*
* Akantu is free software: you can redistribute it and/or modify it under the
* terms of the GNU Lesser General Public License as published by the Free
* Software Foundation, either version 3 of the License, or (at your option) any
* later version.
*
* Akantu is distributed in the hope that it will be useful, but WITHOUT ANY
* WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR
* A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more
* details.
*
* You should have received a copy of the GNU Lesser General Public License
* along with Akantu. If not, see <http://www.gnu.org/licenses/>.
*
*/
/* -------------------------------------------------------------------------- */
#include "grid_synchronizer.hh"
#include "aka_grid_dynamic.hh"
#include "mesh.hh"
#include "fe_engine.hh"
#include "static_communicator.hh"
#include "mesh_io.hh"
#include <iostream>
/* -------------------------------------------------------------------------- */
__BEGIN_AKANTU__
/* -------------------------------------------------------------------------- */
GridSynchronizer::GridSynchronizer(Mesh & mesh,
const ID & id,
MemoryID memory_id,
const bool register_to_event_manager) :
DistributedSynchronizer(mesh, id, memory_id, register_to_event_manager) {
AKANTU_DEBUG_IN();
AKANTU_DEBUG_OUT();
}
/* -------------------------------------------------------------------------- */
template <class E>
GridSynchronizer * GridSynchronizer::createGridSynchronizer(Mesh & mesh,
const SpatialGrid<E> & grid,
SynchronizerID id,
SynchronizerRegistry * synchronizer_registry,
const std::set<SynchronizationTag> & tags_to_register,
MemoryID memory_id,
const bool register_to_event_manager) {
AKANTU_DEBUG_IN();
StaticCommunicator & comm = StaticCommunicator::getStaticCommunicator();
UInt nb_proc = comm.getNbProc();
UInt my_rank = comm.whoAmI();
GridSynchronizer & communicator = *(new GridSynchronizer(mesh, id, memory_id,
register_to_event_manager));
if(nb_proc == 1) return &communicator;
UInt spatial_dimension = mesh.getSpatialDimension();
Real * bounding_boxes = new Real[2 * spatial_dimension * nb_proc];
Real * my_bounding_box = bounding_boxes + 2 * spatial_dimension * my_rank;
// mesh.getLocalLowerBounds(my_bounding_box);
// mesh.getLocalUpperBounds(my_bounding_box + spatial_dimension);
const Vector<Real> & lower = grid.getLowerBounds();
const Vector<Real> & upper = grid.getUpperBounds();
const Vector<Real> & spacing = grid.getSpacing();
for (UInt i = 0; i < spatial_dimension; ++i) {
my_bounding_box[i ] = lower(i) - spacing(i);
my_bounding_box[spatial_dimension + i] = upper(i) + spacing(i);
}
AKANTU_DEBUG_INFO("Exchange of bounding box to detect the overlapping regions.");
comm.allGather(bounding_boxes, spatial_dimension * 2);
bool * intersects_proc = new bool[nb_proc];
std::fill_n(intersects_proc, nb_proc, true);
Int * first_cells = new Int[3 * nb_proc];
Int * last_cells = new Int[3 * nb_proc];
std::fill_n(first_cells, 3 * nb_proc, 0);
std::fill_n(first_cells, 3 * nb_proc, 0);
ElementTypeMapArray<UInt> ** element_per_proc = new ElementTypeMapArray<UInt>* [nb_proc];
for (UInt p = 0; p < nb_proc; ++p) element_per_proc[p] = NULL;
// check the overlapping between my box and the one from other processors
for (UInt p = 0; p < nb_proc; ++p) {
if(p == my_rank) continue;
Real * proc_bounding_box = bounding_boxes + 2 * spatial_dimension * p;
bool intersects = false;
Int * first_cell_p = first_cells + p * spatial_dimension;
Int * last_cell_p = last_cells + p * spatial_dimension;
for (UInt s = 0; s < spatial_dimension; ++s) {
// check overlapping of grid
intersects = Math::intersects(my_bounding_box[s],
my_bounding_box[spatial_dimension + s],
proc_bounding_box[s],
proc_bounding_box[spatial_dimension + s]);
intersects_proc[p] &= intersects;
if(intersects) {
AKANTU_DEBUG_INFO("I intersects with processor " << p << " in direction " << s);
// is point 1 of proc p in the dimension s in the range ?
bool point1 = Math::is_in_range(proc_bounding_box[s],
my_bounding_box[s],
my_bounding_box[s+spatial_dimension]);
// is point 2 of proc p in the dimension s in the range ?
bool point2 = Math::is_in_range(proc_bounding_box[s+spatial_dimension],
my_bounding_box[s],
my_bounding_box[s+spatial_dimension]);
Real start = 0.;
Real end = 0.;
if(point1 && !point2) {
/* |-----------| my_bounding_box(i)
* |-----------| proc_bounding_box(i)
* 1 2
*/
start = proc_bounding_box[s];
end = my_bounding_box[s+spatial_dimension];
AKANTU_DEBUG_INFO("Intersection scheme 1 in direction " << s << " with processor " << p << " [" << start << ", " << end <<"]");
} else if(point1 && point2) {
/* |-----------------| my_bounding_box(i)
* |-----------| proc_bounding_box(i)
* 1 2
*/
start = proc_bounding_box[s];
end = proc_bounding_box[s+spatial_dimension];
AKANTU_DEBUG_INFO("Intersection scheme 2 in direction " << s << " with processor " << p << " [" << start << ", " << end << "]");
} else if(!point1 && point2) {
/* |-----------| my_bounding_box(i)
* |-----------| proc_bounding_box(i)
* 1 2
*/
start = my_bounding_box[s];
end = proc_bounding_box[s+spatial_dimension];
AKANTU_DEBUG_INFO("Intersection scheme 3 in direction " << s << " with processor " << p << " [" << start << ", " << end <<"]");
} else {
/* |-----------| my_bounding_box(i)
* |-----------------| proc_bounding_box(i)
* 1 2
*/
start = my_bounding_box[s];
end = my_bounding_box[s+spatial_dimension];
AKANTU_DEBUG_INFO("Intersection scheme 4 in direction " << s << " with processor " << p << " [" << start << ", " << end <<"]");
}
first_cell_p[s] = grid.getCellID(start, s);
last_cell_p [s] = grid.getCellID(end, s);
}
}
//create the list of cells in the overlapping
typedef typename SpatialGrid<E>::CellID CellID;
std::vector<CellID> * cell_ids = new std::vector<CellID>;
if(intersects_proc[p]) {
AKANTU_DEBUG_INFO("I intersects with processor " << p);
CellID cell_id(spatial_dimension);
// for (UInt i = 0; i < spatial_dimension; ++i) {
// if(first_cell_p[i] != 0) --first_cell_p[i];
// if(last_cell_p[i] != 0) ++last_cell_p[i];
// }
for (Int fd = first_cell_p[0]; fd <= last_cell_p[0]; ++fd) {
cell_id.setID(0, fd);
if(spatial_dimension == 1) {
cell_ids->push_back(cell_id);
}
else {
for (Int sd = first_cell_p[1]; sd <= last_cell_p[1] ; ++sd) {
cell_id.setID(1, sd);
if(spatial_dimension == 2) {
cell_ids->push_back(cell_id);
} else {
for (Int ld = first_cell_p[2]; ld <= last_cell_p[2] ; ++ld) {
cell_id.setID(2, ld);
cell_ids->push_back(cell_id);
}
}
}
}
}
// get the list of elements in the cells of the overlapping
typename std::vector<CellID>::iterator cur_cell_id = cell_ids->begin();
typename std::vector<CellID>::iterator last_cell_id = cell_ids->end();
std::set<Element> * to_send = new std::set<Element>();
for (; cur_cell_id != last_cell_id; ++cur_cell_id) {
typename SpatialGrid<E>::Cell::const_iterator cur_elem = grid.beginCell(*cur_cell_id);
typename SpatialGrid<E>::Cell::const_iterator last_elem = grid.endCell(*cur_cell_id);
for (; cur_elem != last_elem; ++cur_elem) {
to_send->insert(*cur_elem);
}
}
AKANTU_DEBUG_INFO("I have prepared " << to_send->size() << " elements to send to processor " << p);
std::stringstream sstr; sstr << "element_per_proc_" << p;
element_per_proc[p] = new ElementTypeMapArray<UInt>(sstr.str(), id);
ElementTypeMapArray<UInt> & elempproc = *(element_per_proc[p]);
typename std::set<Element>::iterator elem = to_send->begin();
typename std::set<Element>::iterator last_elem = to_send->end();
for (; elem != last_elem; ++elem) {
ElementType type = elem->type;
UInt nb_nodes_per_element = mesh.getNbNodesPerElement(type);
// /!\ this part must be slow due to the access in the ElementTypeMapArray<UInt>
if(!elempproc.exists(type, _not_ghost))
elempproc.alloc(0, nb_nodes_per_element, type, _not_ghost);
UInt global_connect[nb_nodes_per_element];
UInt * local_connect = mesh.getConnectivity(type).storage() + elem->element * nb_nodes_per_element;
for (UInt i = 0; i < nb_nodes_per_element; ++i) {
global_connect[i] = mesh.getNodeGlobalId(local_connect[i]);
AKANTU_DEBUG_ASSERT(global_connect[i] < mesh.getNbGlobalNodes(),
"This global node send in the connectivity does not seem correct "
<< global_connect[i] << " corresponding to "
<< local_connect[i] << " from element " << elem->element);
}
elempproc(type).push_back(global_connect);
communicator.send_element[p].push_back(*elem);
}
delete to_send;
}
delete cell_ids;
}
delete [] first_cells;
delete [] last_cells;
delete [] bounding_boxes;
AKANTU_DEBUG_INFO("I have finished to compute intersection,"
<< " no it's time to communicate with my neighbors");
/**
* Sending loop, sends the connectivity asynchronously to all concerned proc
*/
std::vector<CommunicationRequest *> isend_requests;
for (UInt p = 0; p < nb_proc; ++p) {
if(p == my_rank) continue;
if(intersects_proc[p]) {
ElementTypeMapArray<UInt> & elempproc = *(element_per_proc[p]);
ElementTypeMapArray<UInt>::type_iterator it_type = elempproc.firstType(_all_dimensions, _not_ghost);
ElementTypeMapArray<UInt>::type_iterator last_type = elempproc.lastType (_all_dimensions, _not_ghost);
UInt count = 0;
for (; it_type != last_type; ++it_type) {
Array<UInt> & conn = elempproc(*it_type, _not_ghost);
UInt info[2];
info[0] = (UInt) *it_type;
info[1] = conn.getSize() * conn.getNbComponent();
AKANTU_DEBUG_INFO("I have " << conn.getSize() << " elements of type " << *it_type
<< " to send to processor " << p
<< " (communication tag : " << Tag::genTag(my_rank, count, DATA_TAG) << ")");
isend_requests.push_back(comm.asyncSend(info, 2, p, Tag::genTag(my_rank, count, SIZE_TAG)));
if(info[1] != 0)
isend_requests.push_back(comm.asyncSend<UInt>(conn.storage(),
info[1],
p, Tag::genTag(my_rank, count, DATA_TAG)));
++count;
}
UInt info[2];
info[0] = (UInt) _not_defined;
info[1] = 0;
isend_requests.push_back(comm.asyncSend(info, 2, p, Tag::genTag(my_rank, count, SIZE_TAG)));
}
}
/**
* Receives the connectivity and store them in the ghosts elements
*/
Array<UInt> & global_nodes_ids = const_cast<Array<UInt> &>(mesh.getGlobalNodesIds());
Array<Int> & nodes_type = const_cast<Array<Int> &>(const_cast<const Mesh &>(mesh).getNodesType());
std::vector<CommunicationRequest *> isend_nodes_requests;
UInt nb_nodes_to_recv[nb_proc];
UInt nb_total_nodes_to_recv = 0;
UInt nb_current_nodes = global_nodes_ids.getSize();
NewNodesEvent new_nodes;
NewElementsEvent new_elements;
Array<UInt> * ask_nodes_per_proc = new Array<UInt>[nb_proc];
for (UInt p = 0; p < nb_proc; ++p) {
nb_nodes_to_recv[p] = 0;
if(p == my_rank) continue;
Array<UInt> & ask_nodes = ask_nodes_per_proc[p];
UInt count = 0;
if(intersects_proc[p]) {
ElementType type = _not_defined;
do {
UInt info[2] = { 0 };
comm.receive(info, 2, p, Tag::genTag(p, count, SIZE_TAG));
type = (ElementType) info[0];
if(type != _not_defined) {
UInt nb_nodes_per_element = mesh.getNbNodesPerElement(type);;
UInt nb_element = info[1] / nb_nodes_per_element;
Array<UInt> tmp_conn(nb_element, nb_nodes_per_element);
tmp_conn.clear();
if(info[1] != 0)
comm.receive<UInt>(tmp_conn.storage(), info[1], p, Tag::genTag(p, count, DATA_TAG));
AKANTU_DEBUG_INFO("I will receive " << nb_element << " elements of type " << ElementType(info[0])
<< " from processor " << p
<< " (communication tag : " << Tag::genTag(p, count, DATA_TAG) << ")");
Array<UInt> & ghost_connectivity = const_cast<Array<UInt> &>(mesh.getConnectivity(type, _ghost));
UInt nb_ghost_element = ghost_connectivity.getSize();
Element element(type, 0, _ghost);
UInt conn[nb_nodes_per_element];
for (UInt el = 0; el < nb_element; ++el) {
UInt nb_node_to_ask_for_elem = 0;
for (UInt n = 0; n < nb_nodes_per_element; ++n) {
UInt gn = tmp_conn(el, n);
UInt ln = global_nodes_ids.find(gn);
AKANTU_DEBUG_ASSERT(gn < mesh.getNbGlobalNodes(), "This global node seems not correct " << gn << " from element " << el << " node " << n);
if(ln == UInt(-1)) {
global_nodes_ids.push_back(gn);
nodes_type.push_back(-3); // pure ghost node
ln = nb_current_nodes;
new_nodes.getList().push_back(ln);
++nb_current_nodes;
ask_nodes.push_back(gn);
++nb_node_to_ask_for_elem;
}
conn[n] = ln;
}
// all the nodes are already known locally, the element should already exists
UInt c = UInt(-1);
if(nb_node_to_ask_for_elem == 0) {
c = ghost_connectivity.find(conn);
element.element = c;
}
if(c == UInt(-1)) {
element.element = nb_ghost_element;
++nb_ghost_element;
ghost_connectivity.push_back(conn);
new_elements.getList().push_back(element);
}
communicator.recv_element[p].push_back(element);
}
}
count++;
} while(type != _not_defined);
AKANTU_DEBUG_INFO("I have " << ask_nodes.getSize()
<< " missing nodes for elements coming from processor " << p
<< " (communication tag : " << Tag::genTag(my_rank, 0, ASK_NODES_TAG) << ")");
ask_nodes.push_back(UInt(-1));
isend_nodes_requests.push_back(comm.asyncSend(ask_nodes.storage(), ask_nodes.getSize(),
p, Tag::genTag(my_rank, 0, ASK_NODES_TAG)));
nb_nodes_to_recv[p] = ask_nodes.getSize() - 1;
nb_total_nodes_to_recv += nb_nodes_to_recv[p];
}
}
comm.waitAll(isend_requests);
comm.freeCommunicationRequest(isend_requests);
for (UInt p = 0; p < nb_proc; ++p) {
if(element_per_proc[p]) delete element_per_proc[p];
}
delete [] element_per_proc;
/**
* Sends requested nodes to proc
*/
Array<Real> & nodes = const_cast<Array<Real> &>(mesh.getNodes());
UInt nb_nodes = nodes.getSize();
std::vector<CommunicationRequest *> isend_coordinates_requests;
Array<Real> * nodes_to_send_per_proc = new Array<Real>[nb_proc];
for (UInt p = 0; p < nb_proc; ++p) {
if(p == my_rank || !intersects_proc[p]) continue;
Array<UInt> asked_nodes;
CommunicationStatus status;
AKANTU_DEBUG_INFO("Waiting list of nodes to send to processor " << p
<< "(communication tag : " << Tag::genTag(p, 0, ASK_NODES_TAG) << ")");
comm.probe<UInt>(p, Tag::genTag(p, 0, ASK_NODES_TAG), status);
UInt nb_nodes_to_send = status.getSize();
asked_nodes.resize(nb_nodes_to_send);
AKANTU_DEBUG_INFO("I have " << nb_nodes_to_send - 1
<< " nodes to send to processor " << p
<< " (communication tag : " << Tag::genTag(p, 0, ASK_NODES_TAG) << ")");
AKANTU_DEBUG_INFO("Getting list of nodes to send to processor " << p
<< " (communication tag : " << Tag::genTag(p, 0, ASK_NODES_TAG) << ")");
comm.receive(asked_nodes.storage(), nb_nodes_to_send, p, Tag::genTag(p, 0, ASK_NODES_TAG));
nb_nodes_to_send--;
asked_nodes.resize(nb_nodes_to_send);
Array<Real> & nodes_to_send = nodes_to_send_per_proc[p];
nodes_to_send.extendComponentsInterlaced(spatial_dimension, 1);
for (UInt n = 0; n < nb_nodes_to_send; ++n) {
UInt ln = global_nodes_ids.find(asked_nodes(n));
AKANTU_DEBUG_ASSERT(ln != UInt(-1), "The node [" << asked_nodes(n) << "] requested by proc " << p << " was not found locally!");
nodes_to_send.push_back(nodes.storage() + ln * spatial_dimension);
}
if(nb_nodes_to_send != 0) {
AKANTU_DEBUG_INFO("Sending the " << nb_nodes_to_send << " nodes to processor " << p
<< " (communication tag : " << Tag::genTag(p, 0, SEND_NODES_TAG) << ")");
isend_coordinates_requests.push_back(comm.asyncSend(nodes_to_send.storage(), nb_nodes_to_send * spatial_dimension, p, Tag::genTag(my_rank, 0, SEND_NODES_TAG)));
}
#if not defined (AKANTU_NDEBUG)
else {
AKANTU_DEBUG_INFO("No nodes to send to processor " << p);
}
#endif
}
comm.waitAll(isend_nodes_requests);
comm.freeCommunicationRequest(isend_nodes_requests);
delete [] ask_nodes_per_proc;
nodes.resize(nb_total_nodes_to_recv + nb_nodes);
for (UInt p = 0; p < nb_proc; ++p) {
if((p != my_rank) && (nb_nodes_to_recv[p] > 0)) {
AKANTU_DEBUG_INFO("Receiving the " << nb_nodes_to_recv[p] << " nodes from processor " << p
<< " (communication tag : " << Tag::genTag(p, 0, SEND_NODES_TAG) << ")");
comm.receive(nodes.storage() + nb_nodes * spatial_dimension,
nb_nodes_to_recv[p] * spatial_dimension,
p, Tag::genTag(p, 0, SEND_NODES_TAG));
nb_nodes += nb_nodes_to_recv[p];
}
#if not defined (AKANTU_NDEBUG)
else {
if(p != my_rank) {
AKANTU_DEBUG_INFO("No nodes to receive from processor " << p);
}
}
#endif
}
comm.waitAll(isend_coordinates_requests);
comm.freeCommunicationRequest(isend_coordinates_requests);
delete [] nodes_to_send_per_proc;
// Register the tags if any
if(synchronizer_registry) {
std::set<SynchronizationTag>::const_iterator it = tags_to_register.begin();
std::set<SynchronizationTag>::const_iterator end = tags_to_register.end();
for (; it != end; ++it) {
synchronizer_registry->registerSynchronizer(communicator, *it);
}
}
mesh.sendEvent(new_nodes);
mesh.sendEvent(new_elements);
delete [] intersects_proc;
AKANTU_DEBUG_OUT();
return &communicator;
}
/* -------------------------------------------------------------------------- */
template GridSynchronizer *
GridSynchronizer::createGridSynchronizer<IntegrationPoint>(Mesh & mesh,
const SpatialGrid<IntegrationPoint> & grid,
SynchronizerID id,
SynchronizerRegistry * synchronizer_registry,
const std::set<SynchronizationTag> & tags_to_register,
MemoryID memory_id,
const bool register_to_event_manager = true);
template GridSynchronizer *
GridSynchronizer::createGridSynchronizer<Element>(Mesh & mesh,
const SpatialGrid<Element> & grid,
SynchronizerID id,
SynchronizerRegistry * synchronizer_registry,
const std::set<SynchronizationTag> & tags_to_register,
MemoryID memory_id,
const bool register_to_event_manager = true);
__END_AKANTU__
c4science · Help