Description

Main data structure for CUDA nonbonded force calculations.

Public Attributes
struct gmx_device_info_t *	dev_info
	CUDA device information.

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

cu_atomdata_t *	atdat
	atom data

cu_nbparam_t *	nbparam
	parameters required for the non-bonded calc. More...

cu_plist_t *	plist [2]
	pair-list data structures (local and non-local)

nb_staging_t	nbst
	staging area where fshift/energies get downloaded

cudaStream_t	stream [2]
	local and non-local GPU streams

cudaEvent_t	nonlocal_done
	events used for synchronization More...

cudaEvent_t	misc_ops_and_local_H2D_done
	event triggered when the tasks issued in the local stream that need to precede the non-local force calculations are done (e.g. More...

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

cu_timers_t *	timers
	CUDA event-based timers. More...

gmx_wallclock_gpu_t *	timings
	Timing data. More...

Member Data Documentation

bool gmx_nbnxn_cuda_t::bDoTime

True if event-based timing is enabled.

cudaEvent_t gmx_nbnxn_cuda_t::misc_ops_and_local_H2D_done

event triggered when the tasks issued in the local stream that need to precede the non-local force calculations are done (e.g.

f buffer 0-ing, local x/q H2D)

cu_nbparam_t* gmx_nbnxn_cuda_t::nbparam

parameters required for the non-bonded calc.

cudaEvent_t gmx_nbnxn_cuda_t::nonlocal_done

events used for synchronization

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

cu_timers_t* gmx_nbnxn_cuda_t::timers

CUDA event-based timers.

gmx_wallclock_gpu_t* gmx_nbnxn_cuda_t::timings

Timing data.

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