SG++-Doxygen-Documentation
sgpp::datadriven::MPIMethods Class Reference

#include <MPIMethods.hpp>

Static Public Member Functions

static void assignBatch (int workerID, size_t batchOffset, size_t batchSize, bool doCrossValidation)
 Tell a worker to learn from a batch. More...
 
static void assignSystemMatrixUpdate (int workerID, size_t classIndex)
 Assign the task of updating the system matrix to the specified worker for processing. More...
 
static void bcastCommandNoArgs (MPI_COMMAND_ID commandId)
 Broadcast a command that does not require any additional information. More...
 
template<typename Iterator >
static size_t fillBufferWithData (void *buffer, void *bufferEnd, Iterator &iterator, Iterator &listEnd)
 Copy data from a specified iterator type into a message's payload. More...
 
static size_t fillBufferWithLevelIndexData (void *buffer, const void *bufferEnd, std::list< std::vector< LevelIndexPair >>::iterator &iterator, std::list< std::vector< LevelIndexPair >>::const_iterator &listEnd)
 Copy the level index data for as many grid points as possible into the buffer specified. More...
 
static void finalizeMPI ()
 Cancel any remaining requests and shutdown the MPI communicator. More...
 
static size_t getQueueSize ()
 Check how many pending MPI requests have been registered and not completed yet. More...
 
static unsigned int getWorldSize ()
 Get the number of nodes participating in the current problem. More...
 
static bool hasPendingOutgoingRequests ()
 Check whether there are any outgoing open MPI requests that have not completed yet. More...
 
static void initMPI (LearnerSGDEOnOffParallel *learnerInstance)
 Initializes the MPI setup for the specified learner. More...
 
static bool isMaster ()
 Check whether the current role of this node is master. More...
 
static void processCompletedMPIRequests ()
 Process any asynchronous MPI requests that completed but have not been processed yet. More...
 
static void processIncomingMPICommands (PendingMPIRequest &pendingMPIRequest)
 Process a specific completed MPI request. More...
 
static void receiveGridComponentsUpdate (sgpp::datadriven::RefinementResultNetworkMessage *networkMessage)
 Receive potentially segmented message that contains modifications to the grid. More...
 
static void sendCommandNoArgs (int destinationRank, MPI_COMMAND_ID commandId)
 Send an MPI command using point to point communication without any further payload. More...
 
static PendingMPIRequestsendIBcast (MPI_Packet *mpiPacket)
 Send an assembled packet as an MPI Broadcast from the master. More...
 
static PendingMPIRequestsendISend (int destinationRank, MPI_Packet *mpiPacket, size_t packetSize=sizeof(MPI_Packet), bool highPriority=false)
 Send an MPI packet to a target destination asynchronously. More...
 
static size_t sendMergeGridNetworkMessage (size_t classIndex, size_t batchOffset, size_t batchSize, base::DataVector &alphaVector)
 Send the results of training back to the master node for merging. More...
 
static void sendRefinementUpdates (size_t &classIndex, std::list< size_t > &deletedGridPointsIndices, std::list< LevelIndexVector > &addedGridPoints)
 Send the refinement updates for a specified class to all workers over a broadcast. More...
 
static void sendSystemMatrixDecomposition (const size_t &classIndex, sgpp::base::DataMatrix &newSystemMatrixDecomposition, int mpiTarget)
 Send a new system matrix decomposition with either broadcast or normal send. More...
 
static void waitForAnyMPIRequestsToComplete ()
 Wait for any of the currently pending MPI requests to complete and process them. More...
 
static void waitForGridConsistent (size_t classIndex)
 Wait for a specific grid to return to its consistent state before continuing. More...
 
static void waitForIncomingMessageType (MPI_COMMAND_ID commandId, unsigned int numOccurrences=1, std::function< bool(PendingMPIRequest &)> predicate=[](PendingMPIRequest &request) { return true;})
 Wait for an incoming message of the specified type. More...
 

Static Protected Member Functions

static size_t calculateTotalPacketSize (size_t containedPacketSize)
 Calculate the size of an MPI_Packet based on the size of its contained payload. More...
 
static PendingMPIRequestcreatePendingMPIRequest (MPI_Packet *mpiPacket, bool isInbound)
 Create a pending MPI request for the specified packet in preparation to sending it. More...
 
static std::list< sgpp::datadriven::MessageTrackRequest >::iterator createTrackRequest (unsigned int numOccurrences, const std::function< bool(PendingMPIRequest &)> &predicate)
 Create a message track request that tests each message against a predicate until the target number of occurences are found. More...
 
static unsigned int executeMPIWaitAny ()
 Wait for any of the requests in the MPI request pool to complete. More...
 
static std::list< sgpp::datadriven::PendingMPIRequest >::iterator findPendingMPIRequest (unsigned int completedRequestIndex)
 Find a PendingMPIRequest from the index of the completed MPI_Request structure in the pool storage. More...
 
static void handleIncomingRequestFromCallback (PendingMPIRequest &request)
 Callback function that processes the request and zeros then memory region afterwards. More...
 
static void processCompletedMPIRequest (const std::list< sgpp::datadriven::PendingMPIRequest >::iterator &pendingMPIRequestIterator)
 Process a specific completed MPI request from its position in the request list. More...
 
static void receiveMergeGridNetworkMessage (MergeGridNetworkMessage &networkMessage)
 Receive a packet of changes from the master and apply them to the grid and refinement result or to the system matrix decomposition. More...
 
static void runBatch (MPI_Packet *assignBatchMessage)
 Learn from a batch based on the instructions in the specified packet. More...
 

Static Protected Attributes

static LearnerSGDEOnOffParallellearnerInstance
 The instance of the learner for which MPI communications are currently being handled. More...
 
static std::list< MessageTrackRequestmessageTrackRequests
 Structure to track any pending tracking requests that need to be checked on every incoming message. More...
 
static MPIRequestPool mpiRequestStorage
 Structure to hold the actual MPI_Request structures in memory sequentially along with the additional information for each request. More...
 
static unsigned int mpiWorldSize = 0
 The number of participating nodes for solving the problem. More...
 
static std::list< PendingMPIRequestpendingMPIRequests
 Structure to track all pending MPI requests that have been registered but not completed. More...
 

Member Function Documentation

◆ assignBatch()

void sgpp::datadriven::MPIMethods::assignBatch ( int  workerID,
size_t  batchOffset,
size_t  batchSize,
bool  doCrossValidation 
)
static

Tell a worker to learn from a batch.

Parameters
workerIDThe MPI rank of the worker to command.
batchOffsetThe offset from the start of the training set to learn from.
batchSizeThe size of the batch to learn from.
doCrossValidationWhether to apply cross-validation

References sgpp::datadriven::ASSIGN_BATCH, sgpp::datadriven::AssignBatchNetworkMessage::batchOffset, calculateTotalPacketSize(), sgpp::datadriven::MPI_Packet::commandID, and sendISend().

Referenced by sgpp::datadriven::LearnerSGDEOnOffParallel::assignBatchToWorker().

◆ assignSystemMatrixUpdate()

void sgpp::datadriven::MPIMethods::assignSystemMatrixUpdate ( int  workerID,
size_t  classIndex 
)
static

Assign the task of updating the system matrix to the specified worker for processing.

Parameters
workerIDThe MPI rank of the worker to assign to.
classIndexThe class index of the system matrix to update.

References calculateTotalPacketSize(), sgpp::datadriven::MPI_Packet::commandID, sgpp::datadriven::COMPUTE_UPDATE_SYSTEM_MATRIX_DECOMPOSITION, sgpp::datadriven::LearnerSGDEOnOffParallel::getLocalGridVersion(), sgpp::datadriven::AssignSystemMatrixUpdateNetworkMessage::gridversion, learnerInstance, and sendISend().

Referenced by sgpp::datadriven::RefinementHandler::updateClassVariablesAfterRefinement().

◆ bcastCommandNoArgs()

void sgpp::datadriven::MPIMethods::bcastCommandNoArgs ( MPI_COMMAND_ID  commandId)
static

Broadcast a command that does not require any additional information.

Parameters
commandIdThe command to broadcast

References sgpp::datadriven::MPI_Packet::commandID, and sendIBcast().

Referenced by sgpp::datadriven::LearnerSGDEOnOffParallel::shutdownMPINodes().

◆ calculateTotalPacketSize()

size_t sgpp::datadriven::MPIMethods::calculateTotalPacketSize ( size_t  containedPacketSize)
staticprotected

Calculate the size of an MPI_Packet based on the size of its contained payload.

Parameters
containedPacketSizeThe size of the contained payload.
Returns
The total MPI_Packet size.

References sgpp::datadriven::MPI_Packet::payload.

Referenced by assignBatch(), and assignSystemMatrixUpdate().

◆ createPendingMPIRequest()

PendingMPIRequest & sgpp::datadriven::MPIMethods::createPendingMPIRequest ( MPI_Packet mpiPacket,
bool  isInbound 
)
staticprotected

Create a pending MPI request for the specified packet in preparation to sending it.

Parameters
mpiPacketThe packet of data for the request.
isInboundWhether the request is an inbound request.
Returns
Reference to the newly created PendingMPIRequest.

References sgpp::datadriven::PendingMPIRequest::buffer, sgpp::datadriven::PendingMPIRequest::callback, D, sgpp::datadriven::PendingMPIRequest::disposeAfterCallback, sgpp::datadriven::PendingMPIRequest::inbound, mpiRequestStorage, and pendingMPIRequests.

Referenced by initMPI(), sendIBcast(), and sendISend().

◆ createTrackRequest()

std::list< sgpp::datadriven::MessageTrackRequest >::iterator sgpp::datadriven::MPIMethods::createTrackRequest ( unsigned int  numOccurrences,
const std::function< bool(PendingMPIRequest &)> &  predicate 
)
staticprotected

Create a message track request that tests each message against a predicate until the target number of occurences are found.

Parameters
numOccurrencesHow many successful matches are required.
predicateThe predicate to test messages against.
Returns
The newly allocated track request's position in the storage list

References messageTrackRequests.

Referenced by waitForIncomingMessageType().

◆ executeMPIWaitAny()

unsigned int sgpp::datadriven::MPIMethods::executeMPIWaitAny ( )
staticprotected

◆ fillBufferWithData()

template<typename Iterator >
size_t sgpp::datadriven::MPIMethods::fillBufferWithData ( void *  buffer,
void *  bufferEnd,
Iterator &  iterator,
Iterator &  listEnd 
)
static

Copy data from a specified iterator type into a message's payload.

Template Parameters
IteratorThe type of iterator used to read values to copy.
Parameters
bufferThe start of the buffer to copy to.
bufferEndThe end of the buffer to copy to.
iteratorThe start of the data to copy from.
listEndThe end of the data to copy from.
Returns
How many values of the iterator's type could be copied.

Referenced by sendMergeGridNetworkMessage(), and sendSystemMatrixDecomposition().

◆ fillBufferWithLevelIndexData()

size_t sgpp::datadriven::MPIMethods::fillBufferWithLevelIndexData ( void *  buffer,
const void *  bufferEnd,
std::list< std::vector< LevelIndexPair >>::iterator &  iterator,
std::list< std::vector< LevelIndexPair >>::const_iterator &  listEnd 
)
static

Copy the level index data for as many grid points as possible into the buffer specified.

Parameters
bufferThe start of the buffer to copy to.
bufferEndThe end of the buffer to copy to.
iteratorThe start of the data to copy.
listEndThe end of the data to copy.
Returns
The number of grid points copied.

Referenced by sendRefinementUpdates().

◆ finalizeMPI()

void sgpp::datadriven::MPIMethods::finalizeMPI ( )
static

Cancel any remaining requests and shutdown the MPI communicator.

May be called only once during a programs runtime. Collectives (for example MPI_Ibcast) cannot be cancelled. These must complete before calling this method.

References pendingMPIRequests.

Referenced by sgpp::datadriven::LearnerSGDEOnOffParallel::~LearnerSGDEOnOffParallel().

◆ findPendingMPIRequest()

std::list< PendingMPIRequest >::iterator sgpp::datadriven::MPIMethods::findPendingMPIRequest ( unsigned int  completedRequestIndex)
staticprotected

Find a PendingMPIRequest from the index of the completed MPI_Request structure in the pool storage.

Parameters
completedRequestIndexThe index of the completed MPI_Request in the MPIRequestPool.
Returns
The located PendingMPIRequest.

References pendingMPIRequests.

Referenced by waitForAnyMPIRequestsToComplete(), and waitForIncomingMessageType().

◆ getQueueSize()

size_t sgpp::datadriven::MPIMethods::getQueueSize ( )
static

Check how many pending MPI requests have been registered and not completed yet.

Returns
The number of pending MPI requests.

References pendingMPIRequests.

Referenced by sgpp::datadriven::LearnerSGDEOnOffParallel::trainParallel().

◆ getWorldSize()

unsigned int sgpp::datadriven::MPIMethods::getWorldSize ( )
static

Get the number of nodes participating in the current problem.

Returns
The number of participating nodes

References mpiWorldSize.

Referenced by sgpp::datadriven::RoundRobinScheduler::assignTaskStaticTaskSize(), and sgpp::datadriven::LearnerSGDEOnOffParallel::shutdownMPINodes().

◆ handleIncomingRequestFromCallback()

void sgpp::datadriven::MPIMethods::handleIncomingRequestFromCallback ( PendingMPIRequest request)
staticprotected

Callback function that processes the request and zeros then memory region afterwards.

References sgpp::datadriven::PendingMPIRequest::buffer, D, and processIncomingMPICommands().

Referenced by initMPI().

◆ hasPendingOutgoingRequests()

bool sgpp::datadriven::MPIMethods::hasPendingOutgoingRequests ( )
static

Check whether there are any outgoing open MPI requests that have not completed yet.

Returns
Whether such a request exists

References pendingMPIRequests.

Referenced by sgpp::datadriven::LearnerSGDEOnOffParallel::trainParallel().

◆ initMPI()

void sgpp::datadriven::MPIMethods::initMPI ( LearnerSGDEOnOffParallel learnerInstance)
static

Initializes the MPI setup for the specified learner.

This may only be called once, as it calls MPI_Init.

Parameters
learnerInstanceThe instance of the learner to handle MPI communication for.

References CHECK_INT_TO_UINT, createPendingMPIRequest(), D, handleIncomingRequestFromCallback(), isMaster(), learnerInstance, MPI_MASTER_RANK, MPI_MAX_PROCESSOR_NAME_LENGTH, MPI_TAG_HIGH_PRIORITY_NO_BLOCK, MPI_TAG_STANDARD_COMMAND, and mpiWorldSize.

Referenced by sgpp::datadriven::LearnerSGDEOnOffParallel::LearnerSGDEOnOffParallel().

◆ isMaster()

bool sgpp::datadriven::MPIMethods::isMaster ( )
static

◆ processCompletedMPIRequest()

void sgpp::datadriven::MPIMethods::processCompletedMPIRequest ( const std::list< sgpp::datadriven::PendingMPIRequest >::iterator &  pendingMPIRequestIterator)
staticprotected

Process a specific completed MPI request from its position in the request list.

Parameters
pendingMPIRequestIteratorThe position of the completed MPI request.

References D, messageTrackRequests, and pendingMPIRequests.

Referenced by processCompletedMPIRequests(), waitForAnyMPIRequestsToComplete(), and waitForIncomingMessageType().

◆ processCompletedMPIRequests()

void sgpp::datadriven::MPIMethods::processCompletedMPIRequests ( )
static

Process any asynchronous MPI requests that completed but have not been processed yet.

References D, pendingMPIRequests, and processCompletedMPIRequest().

Referenced by sgpp::datadriven::LearnerSGDEOnOffParallel::trainParallel().

◆ processIncomingMPICommands()

◆ receiveGridComponentsUpdate()

void sgpp::datadriven::MPIMethods::receiveGridComponentsUpdate ( sgpp::datadriven::RefinementResultNetworkMessage networkMessage)
static

Receive potentially segmented message that contains modifications to the grid.

Extract the data into a refinement result and apply to grid or apply to system matrix. This includes additions, deletions, and system matrix updates.

Parameters
networkMessageThe message containing the modification data.

References sgpp::datadriven::ADDED_GRID_POINTS_LIST, sgpp::datadriven::RefinementResult::addedGridPoints, sgpp::datadriven::ASSIGN_BATCH, sgpp::datadriven::LearnerSGDEOnOffParallel::checkGridStateConsistent(), sgpp::datadriven::RefinementResultNetworkMessage::classIndex, D, sgpp::datadriven::DELETED_GRID_POINTS_LIST, sgpp::datadriven::RefinementResult::deletedGridPointsIndices, sgpp::datadriven::LearnerSGDEOnOffParallel::getDensityFunctions(), sgpp::datadriven::LearnerSGDEOnOffParallel::getDimensionality(), sgpp::datadriven::LearnerSGDEOnOffParallel::getGrid(), sgpp::datadriven::LearnerSGDEOnOffParallel::getLocalGridVersion(), sgpp::datadriven::LearnerSGDEOnOffParallel::getRefinementHandler(), sgpp::datadriven::RefinementHandler::getRefinementResult(), GRID_RECEIVED_ADDED_POINTS, GRID_TEMPORARILY_INCONSISTENT, sgpp::datadriven::RefinementResultNetworkMessage::gridversion, isMaster(), sgpp::datadriven::LearnerSGDEOnOffParallel::isVersionConsistent(), learnerInstance, sgpp::datadriven::RefinementResultNetworkMessage::listLength, sgpp::datadriven::RefinementResultNetworkMessage::payload, sgpp::base::DataMatrix::resizeRowsCols(), sendSystemMatrixDecomposition(), sgpp::datadriven::LearnerSGDEOnOffParallel::setLocalGridVersion(), sgpp::datadriven::SYSTEM_MATRIX_DECOMPOSITION, sgpp::datadriven::RefinementHandler::updateClassVariablesAfterRefinement(), sgpp::datadriven::RefinementResultNetworkMessage::updateType, and waitForIncomingMessageType().

Referenced by processIncomingMPICommands().

◆ receiveMergeGridNetworkMessage()

◆ runBatch()

void sgpp::datadriven::MPIMethods::runBatch ( MPI_Packet assignBatchMessage)
staticprotected

Learn from a batch based on the instructions in the specified packet.

Parameters
assignBatchMessageThe packet specifying learning parameters.

References D, dataset, sgpp::datadriven::LearnerSGDEOnOffParallel::getDimensionality(), learnerInstance, sgpp::datadriven::MPI_Packet::payload, and sgpp::datadriven::LearnerSGDEOnOffParallel::workBatch().

Referenced by processIncomingMPICommands().

◆ sendCommandNoArgs()

void sgpp::datadriven::MPIMethods::sendCommandNoArgs ( int  destinationRank,
MPI_COMMAND_ID  commandId 
)
static

Send an MPI command using point to point communication without any further payload.

Parameters
destinationRankThe MPI rank of the receiving node.
commandIdThe command to send.

References sgpp::datadriven::MPI_Packet::commandID, and sendISend().

Referenced by sgpp::datadriven::LearnerSGDEOnOffParallel::trainParallel().

◆ sendIBcast()

PendingMPIRequest & sgpp::datadriven::MPIMethods::sendIBcast ( MPI_Packet mpiPacket)
static

Send an assembled packet as an MPI Broadcast from the master.

Parameters
mpiPacketThe packet to transmit.
Returns
The now pending MPI transmission request.

References createPendingMPIRequest(), D, sgpp::datadriven::PendingMPIRequest::getMPIRequestFromHandle(), and MPI_MASTER_RANK.

Referenced by bcastCommandNoArgs(), sendRefinementUpdates(), and sendSystemMatrixDecomposition().

◆ sendISend()

PendingMPIRequest & sgpp::datadriven::MPIMethods::sendISend ( int  destinationRank,
MPI_Packet mpiPacket,
size_t  packetSize = sizeof(MPI_Packet),
bool  highPriority = false 
)
static

Send an MPI packet to a target destination asynchronously.

Parameters
destinationRankThe MPI rank of the destination.
mpiPacketThe MPI packet to send.
packetSizeThe size of the MPI packet to send, defaults to the entire packet.
highPriorityWhether to send the packet on the high priority no_wait channel.
Returns
The pending MPI request to track completion.

References CHECK_SIZE_T_TO_INT, createPendingMPIRequest(), D, sgpp::datadriven::PendingMPIRequest::getMPIRequestFromHandle(), MPI_TAG_HIGH_PRIORITY_NO_BLOCK, and MPI_TAG_STANDARD_COMMAND.

Referenced by assignBatch(), assignSystemMatrixUpdate(), sendCommandNoArgs(), sendMergeGridNetworkMessage(), and sendSystemMatrixDecomposition().

◆ sendMergeGridNetworkMessage()

size_t sgpp::datadriven::MPIMethods::sendMergeGridNetworkMessage ( size_t  classIndex,
size_t  batchOffset,
size_t  batchSize,
base::DataVector alphaVector 
)
static

Send the results of training back to the master node for merging.

Parameters
classIndexThe index of the class to send results for.
batchOffsetThe offset from the start of the dataset used in this batch.
batchSizeThe size of the current batch.
alphaVectorThe results vector to transmit.
Returns
The number of successfully sent values.

References sgpp::datadriven::MergeGridNetworkMessage::classIndex, sgpp::datadriven::MPI_Packet::commandID, D, fillBufferWithData(), sgpp::datadriven::LearnerSGDEOnOffParallel::getLocalGridVersion(), learnerInstance, sgpp::datadriven::MERGE_GRID, MPI_MASTER_RANK, and sendISend().

Referenced by sgpp::datadriven::LearnerSGDEOnOffParallel::workBatch().

◆ sendRefinementUpdates()

void sgpp::datadriven::MPIMethods::sendRefinementUpdates ( size_t &  classIndex,
std::list< size_t > &  deletedGridPointsIndices,
std::list< LevelIndexVector > &  addedGridPoints 
)
static

Send the refinement updates for a specified class to all workers over a broadcast.

Parameters
classIndexThe index of the modified class.
deletedGridPointsIndicesThe list containing deleted grid points indices.
addedGridPointsThe coordinates of any added grid points.

References sgpp::datadriven::ADDED_GRID_POINTS_LIST, sgpp::datadriven::RefinementResultNetworkMessage::classIndex, sgpp::datadriven::MPI_Packet::commandID, D, sgpp::datadriven::DELETED_GRID_POINTS_LIST, fillBufferWithLevelIndexData(), GRID_RECEIVED_ADDED_POINTS, GRID_RECEIVED_DELETED_INDEXES, GRID_TEMPORARILY_INCONSISTENT, sendIBcast(), and sgpp::datadriven::UPDATE_GRID.

Referenced by sgpp::datadriven::RefinementHandler::updateClassVariablesAfterRefinement().

◆ sendSystemMatrixDecomposition()

void sgpp::datadriven::MPIMethods::sendSystemMatrixDecomposition ( const size_t &  classIndex,
sgpp::base::DataMatrix newSystemMatrixDecomposition,
int  mpiTarget 
)
static

◆ waitForAnyMPIRequestsToComplete()

void sgpp::datadriven::MPIMethods::waitForAnyMPIRequestsToComplete ( )
static

Wait for any of the currently pending MPI requests to complete and process them.

References executeMPIWaitAny(), findPendingMPIRequest(), and processCompletedMPIRequest().

Referenced by sgpp::datadriven::LearnerSGDEOnOffParallel::trainParallel().

◆ waitForGridConsistent()

void sgpp::datadriven::MPIMethods::waitForGridConsistent ( size_t  classIndex)
static

◆ waitForIncomingMessageType()

void sgpp::datadriven::MPIMethods::waitForIncomingMessageType ( MPI_COMMAND_ID  commandId,
unsigned int  numOccurrences = 1,
std::function< bool(PendingMPIRequest &)>  predicate = [](                                             PendingMPIRequest &request) { return true; } 
)
static

Wait for an incoming message of the specified type.

Can optionally include a number of occurences to wait for and a predicate to test each message against.

Parameters
commandIdThe command id of the target packet.
numOccurrencesThe number of specified messages to receive.
predicateA predicate to test each message against.

References sgpp::datadriven::PendingMPIRequest::buffer, sgpp::datadriven::MPI_Packet::commandID, createTrackRequest(), D, executeMPIWaitAny(), findPendingMPIRequest(), sgpp::datadriven::PendingMPIRequest::inbound, messageTrackRequests, and processCompletedMPIRequest().

Referenced by sgpp::datadriven::LearnerSGDEOnOffParallel::computeNewSystemMatrixDecomposition(), sgpp::datadriven::LearnerSGDEOnOffParallel::doRefinementForAll(), receiveGridComponentsUpdate(), sgpp::datadriven::LearnerSGDEOnOffParallel::shutdownMPINodes(), and waitForGridConsistent().

Member Data Documentation

◆ learnerInstance

LearnerSGDEOnOffParallel * sgpp::datadriven::MPIMethods::learnerInstance
staticprotected

◆ messageTrackRequests

std::list< MessageTrackRequest > sgpp::datadriven::MPIMethods::messageTrackRequests
staticprotected

Structure to track any pending tracking requests that need to be checked on every incoming message.

Referenced by createTrackRequest(), processCompletedMPIRequest(), and waitForIncomingMessageType().

◆ mpiRequestStorage

MPIRequestPool sgpp::datadriven::MPIMethods::mpiRequestStorage
staticprotected

Structure to hold the actual MPI_Request structures in memory sequentially along with the additional information for each request.

Referenced by createPendingMPIRequest(), and executeMPIWaitAny().

◆ mpiWorldSize

unsigned int sgpp::datadriven::MPIMethods::mpiWorldSize = 0
staticprotected

The number of participating nodes for solving the problem.

Referenced by getWorldSize(), and initMPI().

◆ pendingMPIRequests

std::list< PendingMPIRequest > sgpp::datadriven::MPIMethods::pendingMPIRequests
staticprotected

Structure to track all pending MPI requests that have been registered but not completed.

Referenced by createPendingMPIRequest(), executeMPIWaitAny(), finalizeMPI(), findPendingMPIRequest(), getQueueSize(), hasPendingOutgoingRequests(), processCompletedMPIRequest(), and processCompletedMPIRequests().


The documentation for this class was generated from the following files: