Description

Main data structure for CUDA nonbonded force calculations.

Main data structure for SYCL nonbonded force calculations.

Main data structure for OpenCL nonbonded force calculations.

Main data structure for HIP nonbonded force calculations.

Public Attributes
const DeviceContext *	deviceContext_
	GPU device context. More...

bool	bUseTwoStreams = false
	true if doing both local/non-local NB work on GPU

bool	bNonLocalStreamDoneMarked = false
	true indicates that the nonlocal_done event was marked

NBAtomDataGpu *	atdat = nullptr
	atom data

int *	atomIndices = nullptr
	array of atom indices

int	atomIndicesSize = 0
	size of atom indices

int	atomIndicesSize_alloc = 0
	size of atom indices allocated in device buffer

int *	cxy_na = nullptr
	x buf ops num of atoms

int	ncxy_na = 0
	number of elements in cxy_na

int	ncxy_na_alloc = 0
	number of elements allocated allocated in device buffer

int *	cxy_ind = nullptr
	x buf ops cell index mapping

int	ncxy_ind = 0
	number of elements in cxy_ind

int	ncxy_ind_alloc = 0
	number of elements allocated allocated in device buffer

NBParamGpu *	nbparam = nullptr
	parameters required for the non-bonded calc.

EnumerationArray < InteractionLocality, std::unique_ptr< GpuPairlist > >	plist = { { nullptr } }
	pair-list data structures (local and non-local)

NBStagingData	nbst
	staging area where fshift/energies get downloaded More...

EnumerationArray < InteractionLocality, const DeviceStream * >	deviceStreams
	local and non-local GPU streams

EnumerationArray < InteractionLocality, bool >	haveWork = { { false } }
	True if there is work for the current domain in the respective locality. More...

bool	bDoTime = false
	True if event-based timing is enabled. More...

GpuTimers *	timers = nullptr
	CUDA event-based timers. More...

gmx_wallclock_gpu_nbnxn_t *	timings = nullptr
	Timing data. TODO: deprecate this and query timers for accumulated data instead. More...

EnumerationArray < InteractionLocality, bool >	didPairlistH2D = { { false } }
	true when a pair-list transfer has been done at this step

EnumerationArray < InteractionLocality, bool >	didPrune = { { false } }
	true when we we did pruning on this step

EnumerationArray < InteractionLocality, bool >	didRollingPrune = { { false } }
	true when we did rolling pruning (at the previous step)

struct gmx_device_runtime_data_t *	dev_rundata = nullptr
	OpenCL runtime data (context, kernels) More...

cl_kernel	kernel_pruneonly [ePruneNR] = { nullptr }
	prune kernels, ePruneKind defined the kernel kinds

bool	bPrefetchLjParam = false
	true if prefetching fg i-atom LJ parameters should be used in the kernels More...

gmx::EnumerationArray < InteractionLocality, std::unique_ptr< GpuPairlist > >	plist = { nullptr }
	pair-list data structures (local and non-local)

DeviceBuffer< int >	atomIndices
	array of atom indices

DeviceBuffer< int >	cxy_na
	x buf ops num of atoms

DeviceBuffer< int >	cxy_ind
	x buf ops cell index mapping

gmx::EnumerationArray < InteractionLocality, const DeviceStream * >	deviceStreams
	local and non-local GPU queues

gmx::EnumerationArray < InteractionLocality, bool >	haveWork
	True if there has been local/nonlocal GPU work, either bonded or nonbonded, scheduled.


GpuEventSynchronizer	nonlocal_done
	Event triggered when the non-local non-bonded kernel is done (and the local transfer can proceed) More...

GpuEventSynchronizer	misc_ops_and_local_H2D_done
	Event triggered when the tasks issued in the local stream that need to precede the non-local force or buffer operation calculations are done (e.g. f buffer 0-ing, local x/q H2D, buffer op initialization in local stream that is required also by nonlocal stream )


cl_kernel	kernel_noener_noprune_ptr [c_numElecTypes][c_numVdwTypes] = { { nullptr } }

cl_kernel	kernel_ener_noprune_ptr [c_numElecTypes][c_numVdwTypes] = { { nullptr } }

cl_kernel	kernel_noener_prune_ptr [c_numElecTypes][c_numVdwTypes] = { { nullptr } }

cl_kernel	kernel_ener_prune_ptr [c_numElecTypes][c_numVdwTypes] = { { nullptr } }


cl_kernel	kernel_memset_f = nullptr

cl_kernel	kernel_memset_f2 = nullptr

cl_kernel	kernel_memset_f3 = nullptr

Member Data Documentation

bool gmx::NbnxmGpu::bDoTime = false

True if event-based timing is enabled.

True if event-based timing is enabled. Always false for SYCL.

bool gmx::NbnxmGpu::bPrefetchLjParam = false

true if prefetching fg i-atom LJ parameters should be used in the kernels

auxiliary kernels implementing memset-like functions

struct gmx_device_runtime_data_t* gmx::NbnxmGpu::dev_rundata = nullptr

OpenCL runtime data (context, kernels)

Pointers to non-bonded kernel functions organized similar with nb_kfunc_xxx arrays in nbnxn_ocl.cpp

const DeviceContext * gmx::NbnxmGpu::deviceContext_

GPU device context.

Todo:: Make it constant reference, once NbnxmGpu is a proper class.

Todo:: Make it constant reference, once NbnxmGpu is a proper class.

Todo:: Make it constant reference, once NbnxmGpu is a proper class.

EnumerationArray< InteractionLocality, bool > gmx::NbnxmGpu::haveWork = { { false } }

True if there is work for the current domain in the respective locality.

This includes local/nonlocal GPU work, either bonded or nonbonded, scheduled to be executed in the current domain. As long as bonded work is not split up into local/nonlocal, if there is bonded GPU work, both flags will be true.

NBStagingData gmx::NbnxmGpu::nbst

staging area where fshift/energies get downloaded

staging area where fshift/energies get downloaded. Will be removed in SYCL.

GpuEventSynchronizer gmx::NbnxmGpu::nonlocal_done

Event triggered when the non-local non-bonded kernel is done (and the local transfer can proceed)

Events used for synchronization.

Event triggered when the non-local non-bonded kernel is done (and the local transfer can proceed)

GpuTimers * gmx::NbnxmGpu::timers = nullptr

CUDA event-based timers.

Dummy timers.

OpenCL event-based timers.

HIP event-based timers.

gmx_wallclock_gpu_nbnxn_t * gmx::NbnxmGpu::timings = nullptr

Timing data. TODO: deprecate this and query timers for accumulated data instead.

Dummy timing data.

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

src/gromacs/nbnxm/cuda/nbnxm_cuda_types.h
src/gromacs/nbnxm/hip/nbnxm_hip_types.h
src/gromacs/nbnxm/opencl/nbnxm_ocl_types.h
src/gromacs/nbnxm/sycl/nbnxm_sycl_types.h

Description

Public Attributes

Member Data Documentation