Homec4science
Diffusion akantu (features/python_fragment_manager) Browse

rAKA/src/synchronizer/communicator.hh
2494dc4e9df3features/python_fragmen…

Last Change
Raw File

communicator.hh
View Options

/**
* @file communicator.hh
*
* @author Nicolas Richart <nicolas.richart@epfl.ch>
*
* @date creation: Fri Jun 18 2010
* @date last modification: Wed Nov 15 2017
*
* @brief Class handling the parallel communications
*
*
* Copyright (©) 2010-2018 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 "aka_array.hh"
#include "aka_common.hh"
#include "aka_event_handler_manager.hh"
#include "communication_buffer.hh"
#include "communication_request.hh"
#include "communicator_event_handler.hh"
/* -------------------------------------------------------------------------- */
#ifndef AKANTU_STATIC_COMMUNICATOR_HH_
#define AKANTU_STATIC_COMMUNICATOR_HH_
namespace akantu {
namespace debug {
class CommunicationException : public Exception {
public:
CommunicationException()
: Exception("An exception happen during a communication process.") {}
};
} // namespace debug
/// @enum SynchronizerOperation reduce operation that the synchronizer can
/// perform
enum class SynchronizerOperation {
_sum,
_min,
_max,
_prod,
_land,
_band,
_lor,
_bor,
_lxor,
_bxor,
_min_loc,
_max_loc,
_null
};
enum class CommunicationMode { _auto, _synchronous, _ready };
namespace {
int _any_source = -1;
}
} // namespace akantu
namespace akantu {
struct CommunicatorInternalData {
virtual ~CommunicatorInternalData() = default;
};
/* -------------------------------------------------------------------------- */
/* -------------------------------------------------------------------------- */
class Communicator : public EventHandlerManager<CommunicatorEventHandler> {
struct private_member {};
/* ------------------------------------------------------------------------ */
/* Constructors/Destructors */
/* ------------------------------------------------------------------------ */
public:
Communicator(int & argc, char **& argv, const private_member & /*m*/);
Communicator(const private_member & /*unused*/ = private_member{});
~Communicator() override;
/* ------------------------------------------------------------------------ */
/* Methods */
/* ------------------------------------------------------------------------ */
public:
/* ------------------------------------------------------------------------ */
/* Point to Point */
/* ------------------------------------------------------------------------ */
template <typename T>
void probe(Int sender, Int tag, CommunicationStatus & status) const;
template <typename T>
bool asyncProbe(Int sender, Int tag, CommunicationStatus & status) const;
/* ------------------------------------------------------------------------ */
template <typename T>
inline void receive(Array<T> & values, Int sender, Int tag) const {
return this->receiveImpl(
values.storage(), values.size() * values.getNbComponent(), sender, tag);
}
template <typename T>
inline void receive(std::vector<T> & values, Int sender, Int tag) const {
return this->receiveImpl(values.data(), values.size(), sender, tag);
}
template <typename Tensor>
inline void
receive(Tensor & values, Int sender, Int tag,
std::enable_if_t<aka::is_tensor<Tensor>::value> * /*unused*/ =
nullptr) const {
return this->receiveImpl(values.storage(), values.size(), sender, tag);
}
inline void receive(CommunicationBufferTemplated<true> & values, Int sender,
Int tag) const {
return this->receiveImpl(values.storage(), values.size(), sender, tag);
}
inline void receive(CommunicationBufferTemplated<false> & values, Int sender,
Int tag) const {
CommunicationStatus status;
this->probe<char>(sender, tag, status);
values.reserve(status.size());
return this->receiveImpl(values.storage(), values.size(), sender, tag);
}
template <typename T>
inline void
receive(T & values, Int sender, Int tag,
std::enable_if_t<std::is_arithmetic<T>::value> * /*unused*/ =
nullptr) const {
return this->receiveImpl(&values, 1, sender, tag);
}
/* ------------------------------------------------------------------------ */
template <typename T>
inline void
send(const Array<T> & values, Int receiver, Int tag,
const CommunicationMode & mode = CommunicationMode::_auto) const {
return this->sendImpl(values.storage(),
values.size() * values.getNbComponent(), receiver,
tag, mode);
}
template <typename T>
inline void
send(const std::vector<T> & values, Int receiver, Int tag,
const CommunicationMode & mode = CommunicationMode::_auto) const {
return this->sendImpl(values.data(), values.size(), receiver, tag, mode);
}
template <typename Tensor>
inline void
send(const Tensor & values, Int receiver, Int tag,
const CommunicationMode & mode = CommunicationMode::_auto,
std::enable_if_t<aka::is_tensor<Tensor>::value> * /*unused*/ =
nullptr) const {
return this->sendImpl(values.storage(), values.size(), receiver, tag, mode);
}
template <bool is_static>
inline void
send(const CommunicationBufferTemplated<is_static> & values, Int receiver,
Int tag,
const CommunicationMode & mode = CommunicationMode::_auto) const {
return this->sendImpl(values.storage(), values.size(), receiver, tag, mode);
}
template <typename T>
inline void send(const T & values, Int receiver, Int tag,
const CommunicationMode & mode = CommunicationMode::_auto,
std::enable_if_t<std::is_arithmetic<T>::value> * /*unused*/ =
nullptr) const {
return this->sendImpl(&values, 1, receiver, tag, mode);
}
/* ------------------------------------------------------------------------ */
template <typename T>
inline CommunicationRequest
asyncSend(const Array<T> & values, Int receiver, Int tag,
const CommunicationMode & mode = CommunicationMode::_auto) const {
return this->asyncSendImpl(values.storage(),
values.size() * values.getNbComponent(),
receiver, tag, mode);
}
template <typename T>
inline CommunicationRequest
asyncSend(const std::vector<T> & values, Int receiver, Int tag,
const CommunicationMode & mode = CommunicationMode::_auto) const {
return this->asyncSendImpl(values.data(), values.size(), receiver, tag,
mode);
}
template <typename Tensor>
inline CommunicationRequest
asyncSend(const Tensor & values, Int receiver, Int tag,
const CommunicationMode & mode = CommunicationMode::_auto,
std::enable_if_t<aka::is_tensor<Tensor>::value> * /*unused*/ =
nullptr) const {
return this->asyncSendImpl(values.storage(), values.size(), receiver, tag,
mode);
}
template <bool is_static>
inline CommunicationRequest
asyncSend(const CommunicationBufferTemplated<is_static> & values,
Int receiver, Int tag,
const CommunicationMode & mode = CommunicationMode::_auto) const {
return this->asyncSendImpl(values.storage(), values.size(), receiver, tag,
mode);
}
template <typename T>
inline CommunicationRequest
asyncSend(const T & values, Int receiver, Int tag,
const CommunicationMode & mode = CommunicationMode::_auto,
std::enable_if_t<std::is_arithmetic<T>::value> * /*unused*/ =
nullptr) const {
return this->asyncSendImpl(&values, 1, receiver, tag, mode);
}
/* ------------------------------------------------------------------------ */
template <typename T>
inline CommunicationRequest asyncReceive(Array<T> & values, Int sender,
Int tag) const {
return this->asyncReceiveImpl(
values.storage(), values.size() * values.getNbComponent(), sender, tag);
}
template <typename T>
inline CommunicationRequest asyncReceive(std::vector<T> & values, Int sender,
Int tag) const {
return this->asyncReceiveImpl(values.data(), values.size(), sender, tag);
}
template <typename Tensor,
typename = std::enable_if_t<aka::is_tensor<Tensor>::value>>
inline CommunicationRequest asyncReceive(Tensor & values, Int sender,
Int tag) const {
return this->asyncReceiveImpl(values.storage(), values.size(), sender, tag);
}
template <bool is_static>
inline CommunicationRequest
asyncReceive(CommunicationBufferTemplated<is_static> & values, Int sender,
Int tag) const {
return this->asyncReceiveImpl(values.storage(), values.size(), sender, tag);
}
/* ------------------------------------------------------------------------ */
/* Collectives */
/* ------------------------------------------------------------------------ */
template <typename T>
inline void
allReduce(Array<T> & values,
SynchronizerOperation op = SynchronizerOperation::_sum) const {
this->allReduceImpl(values.storage(),
values.size() * values.getNbComponent(), op);
}
template <typename Tensor>
inline void
allReduce(Tensor & values,
SynchronizerOperation op = SynchronizerOperation::_sum,
std::enable_if_t<aka::is_tensor<Tensor>::value> * /*unused*/ =
nullptr) const {
this->allReduceImpl(values.storage(), values.size(), op);
}
template <typename T>
inline void
allReduce(T & values, SynchronizerOperation op = SynchronizerOperation::_sum,
std::enable_if_t<std::is_arithmetic<T>::value> * /*unused*/ =
nullptr) const {
this->allReduceImpl(&values, 1, op);
}
template <typename T>
inline void
scan(Array<T> & values,
SynchronizerOperation op = SynchronizerOperation::_sum) const {
this->scanImpl(values.storage(), values.storage(),
values.size() * values.getNbComponent(), op);
}
template <typename Tensor>
inline void
scan(Tensor & values, SynchronizerOperation op,
std::enable_if_t<aka::is_tensor<Tensor>::value> * /*unused*/ =
nullptr) const {
this->scanImpl(values.storage(), values.storage(), values.size(), op);
}
template <typename T>
inline void scan(T & values,
SynchronizerOperation op = SynchronizerOperation::_sum,
std::enable_if_t<std::is_arithmetic<T>::value> * /*unused*/ =
nullptr) const {
this->scanImpl(&values, &values, 1, op);
}
template <typename T>
inline void
exclusiveScan(Array<T> & values,
SynchronizerOperation op = SynchronizerOperation::_sum) const {
this->exclusiveScanImpl(values.storage(), values.storage(),
values.size() * values.getNbComponent(), op);
}
template <typename Tensor>
inline void
exclusiveScan(Tensor & values,
SynchronizerOperation op = SynchronizerOperation::_sum,
std::enable_if_t<aka::is_tensor<Tensor>::value> * /*unused*/ =
nullptr) const {
this->exclusiveScanImpl(values.storage(), values.storage(), values.size(),
op);
}
template <typename T>
inline void
exclusiveScan(T & values,
SynchronizerOperation op = SynchronizerOperation::_sum,
std::enable_if_t<std::is_arithmetic<T>::value> * /*unused*/ =
nullptr) const {
this->exclusiveScanImpl(&values, &values, 1, op);
}
template <typename T>
inline void
exclusiveScan(T & values, T & result,
SynchronizerOperation op = SynchronizerOperation::_sum,
std::enable_if_t<std::is_arithmetic<T>::value> * /*unused*/ =
nullptr) const {
this->exclusiveScanImpl(&values, &result, 1, op);
}
/* ------------------------------------------------------------------------ */
template <typename T> inline void allGather(Array<T> & values) const {
AKANTU_DEBUG_ASSERT(UInt(getNbProc()) == values.size(),
"The array size is not correct");
this->allGatherImpl(values.storage(), values.getNbComponent());
}
template <typename Tensor,
typename = std::enable_if_t<aka::is_tensor<Tensor>::value>>
inline void allGather(Tensor & values) const {
AKANTU_DEBUG_ASSERT(values.size() / getNbProc() > 0,
"The vector size is not correct");
this->allGatherImpl(values.storage(), values.size() / getNbProc());
}
/* ------------------------------------------------------------------------ */
template <typename T>
inline void allGatherV(Array<T> & values, const Array<Int> & sizes) const {
this->allGatherVImpl(values.storage(), sizes.storage());
}
/* ------------------------------------------------------------------------ */
template <typename T>
inline void reduce(Array<T> & values, SynchronizerOperation op,
int root = 0) const {
this->reduceImpl(values.storage(), values.size() * values.getNbComponent(),
op, root);
}
/* ------------------------------------------------------------------------ */
template <typename Tensor>
inline void
gather(Tensor & values, int root = 0,
std::enable_if_t<aka::is_tensor<Tensor>::value> * /*unused*/ =
nullptr) const {
this->gatherImpl(values.storage(), values.getNbComponent(), root);
}
template <typename T>
inline void
gather(T values, int root = 0,
std::enable_if_t<std::is_arithmetic<T>::value> * /*unused*/ =
nullptr) const {
this->gatherImpl(&values, 1, root);
}
/* ------------------------------------------------------------------------ */
template <typename Tensor, typename T>
inline void
gather(Tensor & values, Array<T> & gathered,
std::enable_if_t<aka::is_tensor<Tensor>::value> * /*unused*/ =
nullptr) const {
AKANTU_DEBUG_ASSERT(values.size() == gathered.getNbComponent(),
"The array size is not correct");
gathered.resize(getNbProc());
this->gatherImpl(values.data(), values.size(), gathered.storage(),
gathered.getNbComponent());
}
template <typename T>
inline void
gather(T values, Array<T> & gathered,
std::enable_if_t<std::is_arithmetic<T>::value> * /*unused*/ =
nullptr) const {
this->gatherImpl(&values, 1, gathered.storage(), 1);
}
/* ------------------------------------------------------------------------ */
template <typename T>
inline void gatherV(Array<T> & values, const Array<Int> & sizes,
int root = 0) const {
this->gatherVImpl(values.storage(), sizes.storage(), root);
}
/* ------------------------------------------------------------------------ */
template <typename T>
inline void broadcast(Array<T> & values, int root = 0) const {
this->broadcastImpl(values.storage(),
values.size() * values.getNbComponent(), root);
}
template <typename T>
inline void broadcast(std::vector<T> & values, int root = 0) const {
this->broadcastImpl(values.data(), values.size(), root);
}
inline void broadcast(CommunicationBufferTemplated<true> & buffer,
int root = 0) const {
this->broadcastImpl(buffer.storage(), buffer.size(), root);
}
inline void broadcast(CommunicationBufferTemplated<false> & buffer,
int root = 0) const {
UInt buffer_size = buffer.size();
this->broadcastImpl(&buffer_size, 1, root);
if (whoAmI() != root) {
buffer.reserve(buffer_size);
}
if (buffer_size == 0) {
return;
}
this->broadcastImpl(buffer.storage(), buffer.size(), root);
}
template <typename T> inline void broadcast(T & values, int root = 0) const {
this->broadcastImpl(&values, 1, root);
}
/* ------------------------------------------------------------------------ */
void barrier() const;
CommunicationRequest asyncBarrier() const;
/* ------------------------------------------------------------------------ */
/* Request handling */
/* ------------------------------------------------------------------------ */
static bool test(CommunicationRequest & request);
static bool testAll(std::vector<CommunicationRequest> & request);
static void wait(CommunicationRequest & request);
static void waitAll(std::vector<CommunicationRequest> & requests);
static UInt waitAny(std::vector<CommunicationRequest> & requests);
static inline void freeCommunicationRequest(CommunicationRequest & request);
static inline void
freeCommunicationRequest(std::vector<CommunicationRequest> & requests);
template <typename T, typename MsgProcessor>
inline void
receiveAnyNumber(std::vector<CommunicationRequest> & send_requests,
MsgProcessor && processor, Int tag) const;
protected:
template <typename T>
void
sendImpl(const T * buffer, Int size, Int receiver, Int tag,
const CommunicationMode & mode = CommunicationMode::_auto) const;
template <typename T>
void receiveImpl(T * buffer, Int size, Int sender, Int tag) const;
template <typename T>
CommunicationRequest asyncSendImpl(
const T * buffer, Int size, Int receiver, Int tag,
const CommunicationMode & mode = CommunicationMode::_auto) const;
template <typename T>
CommunicationRequest asyncReceiveImpl(T * buffer, Int size, Int sender,
Int tag) const;
template <typename T>
void allReduceImpl(T * values, int nb_values, SynchronizerOperation op) const;
template <typename T>
void scanImpl(T * values, T * results, int nb_values,
SynchronizerOperation op) const;
template <typename T>
void exclusiveScanImpl(T * values, T * results, int nb_values,
SynchronizerOperation op) const;
template <typename T> void allGatherImpl(T * values, int nb_values) const;
template <typename T> void allGatherVImpl(T * values, int * nb_values) const;
template <typename T>
void reduceImpl(T * values, int nb_values, SynchronizerOperation op,
int root = 0) const;
template <typename T>
void gatherImpl(T * values, int nb_values, int root = 0) const;
template <typename T>
void gatherImpl(T * values, int nb_values, T * gathered,
int nb_gathered = 0) const;
template <typename T>
void gatherVImpl(T * values, int * nb_values, int root = 0) const;
template <typename T>
void broadcastImpl(T * values, int nb_values, int root = 0) const;
/* ------------------------------------------------------------------------ */
/* Accessors */
/* ------------------------------------------------------------------------ */
public:
Int getNbProc() const;
Int whoAmI() const;
static Communicator & getStaticCommunicator();
static Communicator & getStaticCommunicator(int & argc, char **& argv);
int getMaxTag() const;
int getMinTag() const;
AKANTU_GET_MACRO(CommunicatorData, (*communicator_data), decltype(auto));
/* ------------------------------------------------------------------------ */
/* Class Members */
/* ------------------------------------------------------------------------ */
private:
static std::unique_ptr<Communicator> static_communicator;
protected:
std::unique_ptr<CommunicatorInternalData> communicator_data;
};
inline std::ostream & operator<<(std::ostream & stream,
const CommunicationRequest & _this) {
_this.printself(stream);
return stream;
}
} // namespace akantu
#include "communicator_inline_impl.hh"
#endif /* AKANTU_STATIC_COMMUNICATOR_HH_ */
c4science · Help