Page MenuHomec4science
Files F66935234

filtered_synchronizer.cc
No OneTemporary
Actions

File Metadata

Created
Tue, Jun 18, 23:29

filtered_synchronizer.cc
View Options

/**
* @file filtered_synchronizer.cc
*
* @author David Simon Kammer <david.kammer@epfl.ch>
* @author Mathilde Radiguet <mathilde.radiguet@epfl.ch>
*
* @date creation: Wed Sep 18 2013
* @date last modification: Sat Jul 11 2015
*
* @brief filtered synchronizer
*
* @section LICENSE
*
* Copyright (©) 2014, 2015 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 "filtered_synchronizer.hh"
/* -------------------------------------------------------------------------- */
#include <map>
/* -------------------------------------------------------------------------- */
__BEGIN_AKANTU__
/* -------------------------------------------------------------------------- */
FilteredSynchronizer::FilteredSynchronizer(Mesh & mesh, SynchronizerID id,
MemoryID memory_id)
: ElementSynchronizer(mesh, id, memory_id) {
AKANTU_DEBUG_IN();
AKANTU_DEBUG_OUT();
}
/* -------------------------------------------------------------------------- */
FilteredSynchronizer * FilteredSynchronizer::createFilteredSynchronizer(
const ElementSynchronizer & d_synchronizer, SynchElementFilter & filter) {
AKANTU_DEBUG_IN();
FilteredSynchronizer & f_synchronizer = *(new FilteredSynchronizer(
d_synchronizer.mesh, d_synchronizer.id + ":filtered",
d_synchronizer.memory_id));
f_synchronizer.setupSynchronizer(d_synchronizer, filter);
AKANTU_DEBUG_OUT();
return &f_synchronizer;
}
/* -------------------------------------------------------------------------- */
void FilteredSynchronizer::setupSynchronizer(
const ElementSynchronizer & d_synchronizer, SynchElementFilter & filter) {
AKANTU_DEBUG_IN();
std::vector<CommunicationRequest> isend_requests;
std::map<UInt, Array<UInt> > keep_element_recv;
auto rs_it = d_synchronizer.communications.begin_recv_scheme();
auto rs_end = d_synchronizer.communications.end_recv_scheme();
// Filter the recv list and communicate it to the sender processor
for (; rs_it != rs_end; ++rs_it) {
const auto & scheme = rs_it->second;
auto & proc = rs_it->first;
Array<UInt> & keep_element = keep_element_recv[proc];
auto & new_scheme = this->communications.createRecvScheme(proc);
UInt el_pos = 0;
for (const auto & element : scheme) {
if (filter(element)) {
new_scheme.push_back(element);
keep_element.push_back(el_pos);
}
++el_pos;
}
keep_element.push_back(-1); // just to be sure to send
// something due to some shitty
// MPI implementation who do not
// know what to do with a 0 size
// send
Tag tag = Tag::genTag(this->rank, 0, RECEIVE_LIST_TAG);
AKANTU_DEBUG_INFO("I have " << keep_element.getSize() - 1
<< " elements to still receive from processor "
<< proc << " (communication tag : " << tag
<< ")");
isend_requests.push_back(communicator.asyncSend(keep_element, proc, tag));
}
auto ss_it = d_synchronizer.communications.begin_send_scheme();
auto ss_end = d_synchronizer.communications.end_send_scheme();
// Receive the element to remove from the sender scheme
for (; ss_it != ss_end; ++ss_it) {
const auto & scheme = rs_it->second;
auto & proc = rs_it->first;
auto & new_scheme = this->communications.createSendScheme(proc);
Tag tag = Tag::genTag(proc, 0, RECEIVE_LIST_TAG);
AKANTU_DEBUG_INFO("Waiting list of elements to keep from processor "
<< proc << " (communication tag : " << tag << ")");
CommunicationStatus status;
communicator.probe<UInt>(proc, tag, status);
Array<UInt> keep_element(status.getSize());
AKANTU_DEBUG_INFO("I have "
<< keep_element.getSize() - 1
<< " elements to keep in my send list to processor "
<< proc << " (communication tag : " << tag << ")");
communicator.receive(keep_element, proc, tag);
for (UInt i = 0; i < keep_element.getSize() - 1; ++i) {
new_scheme.push_back(scheme(keep_element(i)));
}
}
communicator.waitAll(isend_requests);
communicator.freeCommunicationRequest(isend_requests);
AKANTU_DEBUG_OUT();
}
/* -------------------------------------------------------------------------- */
void FilteredSynchronizer::updateElementList(
Array<Element> * source_elements, Array<Element> * destination_elements,
SynchElementFilter & filter) {
AKANTU_DEBUG_IN();
// loop over procs
for (UInt p = 0; p < this->nb_proc; ++p) {
if (p == this->rank)
continue;
// access the element for this proc
const Array<Element> & unfiltered_elements = source_elements[p];
Array<Element> & filtered_elements = destination_elements[p];
// iterator to loop over all source elements
Array<Element>::const_iterator<Element> it = unfiltered_elements.begin();
Array<Element>::const_iterator<Element> end = unfiltered_elements.end();
// if filter accepts this element, push it into the destination elements
for (; it != end; ++it) {
const Element & element = *it;
if (filter(element)) {
filtered_elements.push_back(element);
}
}
}
AKANTU_DEBUG_OUT();
}
__END_AKANTU__

Event Timeline

c4science · Help