Description

Impl class stub.

Public Member Functions
	Impl (GpuEventSynchronizer *pmeForcesReady, MPI_Comm comm, const DeviceContext &deviceContext, gmx::ArrayRef< PpRanks > ppRanks)
	Creates PME GPU Force sender object. More...

void	setForceSendBuffer (DeviceBuffer< Float3 > d_f)
	Sets location of force to be sent to each PP rank. More...

void	sendFToPpPeerToPeer (int ppRank, int numAtoms, bool sendForcesDirectToPpGpu)
	Send force to PP rank (used with Thread-MPI) More...

void	sendFToPpGpuAwareMpi (DeviceBuffer< RVec > sendbuf, int offset, int numBytes, int ppRank, MPI_Request *request)
	Send force to PP rank (used with Lib-MPI) More...

void	waitForEvents ()

Constructor & Destructor Documentation

gmx::PmeForceSenderGpu::Impl::Impl	(	GpuEventSynchronizer *	pmeForcesReady,
		MPI_Comm	comm,
		const DeviceContext &	deviceContext,
		gmx::ArrayRef< PpRanks >	ppRanks
	)

Creates PME GPU Force sender object.

Create PME-PP GPU communication object.

Parameters

[in]	pmeForcesReady	Event synchronizer marked when PME forces are ready on the GPU
[in]	comm	Communicator used for simulation
[in]	deviceContext	GPU context
[in]	ppRanks	List of PP ranks

void gmx::PmeForceSenderGpu::Impl::sendFToPpGpuAwareMpi	(	DeviceBuffer< RVec >	sendbuf,
		int	offset,
		int	numBytes,
		int	ppRank,
		MPI_Request *	request
	)

Send force to PP rank (used with Lib-MPI)

Send PME data directly using GPU-aware MPI.

Parameters

[in]	sendbuf	force buffer in GPU memory
[in]	offset	starting element in buffer
[in]	numBytes	number of bytes to transfer
[in]	ppRank	PP rank to receive data
[in]	request	MPI request to track asynchronous MPI call status

void gmx::PmeForceSenderGpu::Impl::sendFToPpPeerToPeer	(	int	ppRank,
		int	numAtoms,
		bool	sendForcesDirectToPpGpu
	)

Send force to PP rank (used with Thread-MPI)

Send PME synchronizer directly to the peer devices. Not implemented with SYCL.

Send PME synchronizer directly using HIP memory copy.

Parameters

[in]	ppRank	PP rank to receive data
[in]	numAtoms	number of atoms to send
[in]	sendForcesDirectToPpGpu	whether forces are transferred direct to remote GPU memory

void gmx::PmeForceSenderGpu::Impl::setForceSendBuffer ( DeviceBuffer< Float3 > d_f )

Sets location of force to be sent to each PP rank.

Parameters

[in] d_f force buffer in GPU memory

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