Open Qmin: multirankSimulation Class Reference

Open Qmin 0.8.0

GPU-accelerated Q-tensor-based liquid crystal simulations

#include <multirankSimulation.h>

Inheritance diagram for multirankSimulation:

[legend]

Collaboration diagram for multirankSimulation:

[legend]

Public Member Functions
	multirankSimulation (int _myRank, int xDiv, int yDiv, int zDiv, bool _edges, bool _corners)

virtual void	moveParticles (GPUArray< dVec > &displacements, scalar scale=1.0)
	move particles, and also communicate halo sites More...

void	finalizeObjects ()
	transfer buffers and make sure sites on the skin of each rank have correct type More...

void	createMultirankBoundaryObject (vector< int3 > &latticeSites, vector< dVec > &qTensors, boundaryType _type, scalar Param1, scalar Param2)
	the flexible base function...given lattice sites composing an object, determine which are on this rank and add the object More...

void	createWall (int xyz, int plane, boundaryObject &bObj)
	make a wall with x, y, or z normal More...

void	createSphericalColloid (scalar3 center, scalar radius, boundaryObject &bObj)
	make a simple sphere, setting all points within radius of center to be the object More...

void	createSphericalCavity (scalar3 center, scalar radius, boundaryObject &bObj)
	make a simple sphere, setting all points farther than radius of center to be the object More...

void	createCylindricalObject (scalar3 cylinderStart, scalar3 cylinderEnd, scalar radius, bool colloidOrCapillary, boundaryObject &bObj)
	make a cylindrical object, with either the inside or outside defined as the object More...

void	createSpherocylinder (scalar3 cylinderStart, scalar3 cylinderEnd, scalar radius, boundaryObject &bObj)
	make a spherocylinder, defined by the start and end of the cylindrical section and the radius More...

virtual void	createBoundaryFromFile (string fname, bool verbose=false)
	import a boundary object from a (carefully prepared) text file More...

void	setDipolarField (scalar3 center, scalar ThetaD, scalar radius, scalar range, scalar S0)
	a function of convenience... make a dipolar field a la Lubensky et al. More...

void	setDipolarField (scalar3 center, scalar3 direction, scalar radius, scalar range, scalar S0)
	a function of convenience... make a dipolar field a la Lubensky et al. Simpler, uses the ravnik expression More...

virtual void	communicateHaloSitesRoutine ()
	handles calls to all necessary halo site transfer More...

virtual void	synchronizeAndTransferBuffers ()
	synchronize mpi and make transfer buffers More...

shared_ptr< multirankSimulation >	getPointer ()
	return a shared pointer to this Simulation More...

void	setConfiguration (MConfigPtr _config)
	Pass in a reference to the configuration. More...

virtual void	computeForces ()
	Call the force computer to compute the forces. More...

void	addUpdater (UpdaterPtr _upd)
	Add an updater. More...

void	addUpdater (UpdaterPtr _upd, MConfigPtr _config)
	Add an updater with a reference to a configuration. More...

virtual void	addForce (ForcePtr _force)
	Add a force computer configuration. More...

virtual void	addForce (ForcePtr _force, MConfigPtr _config)
	Add a force computer configuration. More...

void	clearForceComputers ()
	Clear out the vector of forceComputes. More...

void	clearUpdaters ()
	Clear out the vector of updaters. More...

scalar	getMaxForce ()
	A utility function that just checks the first updater for a max force. More...

void	performTimestep ()
	Call every updater to advance one time step. More...

virtual void	sumUpdaterData (vector< scalar > &data)
	manipulate data from updaters More...

virtual scalar	computePotentialEnergy (bool verbose=false)
	compute the potential energy associated with all of the forces More...

virtual scalar	computeKineticEnergy (bool verbose=false)
	compute the kinetic energy More...

virtual scalar	computeEnergy (bool verbose=false)
	compute the total energy More...

void	setIntegrationTimestep (scalar dt)
	Set the simulation timestep. More...

void	setCPUOperation (bool setcpu)
	turn on CPU-only mode for all components More...

void	setReproducible (bool reproducible)
	Enforce reproducible dynamics. More...

void	saveState (string fname, int latticeSkip=1, int defectType=0)
	save a file for each rank recording the expanded lattice; lattice skip controls the sparsity of saved sites More...

void	setBox (BoxPtr _box)
	This changes the contents of the Box pointed to by Box to match that of _box. More...

void	setSortPeriod (int sp)
	Set the time between spatial sorting operations. More...

virtual void	setCurrentTime (scalar _cTime)
	reset the simulation clock More...

virtual void	setCurrentTimestep (int _cTime)
	reset the simulation clock counter More...

Public Attributes
profiler	p1 = profiler("total communication time")

profiler	p4 = profiler("MPI recv")

profiler	p3 = profiler("MPI send")

profiler	p2 = profiler("GPU data buffering kernels time")

WeakMConfigPtr	mConfiguration
	The configuration of latticeSites. More...

vector< WeakUpdaterPtr >	updaters
	A vector of updaters that the simulation will loop through. More...

vector< WeakForcePtr >	forceComputers
	A vector of force computes the simulation will loop through. More...

int3	latticeMinPosition
	in multi-rank simulations, this stores the lowest (x,y,z) coordinate controlled by the current rank More...

WeakConfigPtr	configuration
	The configuration of particles. More...

BoxPtr	Box
	The domain of the simulation. More...

int	integerTimestep
	An integer that keeps track of how often performTimestep has been called. More...

scalar	Time
	The current simulation time. More...

scalar	integrationTimestep
	The dt of a time step. More...

bool	useGPU
	A flag controlling whether to use the GPU. More...

int	myRank
	integer for this rank More...

int	nRanks
	total number of ranks More...

int	NActive = 0
	some measure of the number of active degrees of freedom More...

Protected Member Functions
void	setRankTopology (int x, int y, int z)

void	determineCommunicationPattern (bool _edges, bool _corners)

Protected Attributes
vector< int2 >	communicationDirections

vector< bool >	communicationDirectionParity

vector< int >	communicationTargets

int3	rankTopology
	the number of ranks per {x,y,z} axis More...

int3	rankParity
	the local {x,y,z} rank coordinate More...

Index3D	parityTest

bool	edges
	do edges need to be communicated? More...

bool	corners
	do corner sites? More...

bool	transfersUpToDate
	have the halo sites been communicated? More...

MPI_Status	mpiStatus

vector< MPI_Status >	mpiStatuses

vector< MPI_Request >	mpiRequests

vector< scalar >	dataBuffer

int	sortPeriod
	Determines how frequently the spatial sorter be called...once per sortPeriod Timesteps. When sortPeriod < 0 no sorting occurs. More...

bool	spatialSortThisStep
	A flag that determins if a spatial sorting is due to occur this Timestep. More...

Constructor & Destructor Documentation

◆ multirankSimulation()

multirankSimulation::multirankSimulation	(	int	_myRank,
		int	xDiv,
		int	yDiv,
		int	zDiv,
		bool	_edges,
		bool	_corners
	)

inline

References determineCommunicationPattern(), basicSimulation::myRank, and setRankTopology().

Member Function Documentation

◆ moveParticles()

void multirankSimulation::moveParticles	(	GPUArray< dVec > &	displacements,
		scalar	scale = `1.0`
	)

virtual

move particles, and also communicate halo sites

Calls the configuration to displace the degrees of freedom, and communicates halo sites according to the rankTopology and boolean settings

Implements basicSimulation.

References communicateHaloSitesRoutine(), profiler::end(), mConfiguration, p1, profiler::start(), and transfersUpToDate.

◆ finalizeObjects()

void multirankSimulation::finalizeObjects ( )

transfer buffers and make sure sites on the skin of each rank have correct type

References communicateHaloSitesRoutine(), basicSimulation::NActive, and updaters.

◆ createMultirankBoundaryObject()

void multirankSimulation::createMultirankBoundaryObject	(	vector< int3 > &	latticeSites,
		vector< dVec > &	qTensors,
		boundaryType	_type,
		scalar	Param1,
		scalar	Param2
	)

the flexible base function...given lattice sites composing an object, determine which are on this rank and add the object

References latticeMinPosition, mConfiguration, rankParity, rankTopology, and wrap().

Referenced by createBoundaryFromFile(), createCylindricalObject(), createSphericalCavity(), createSphericalColloid(), createSpherocylinder(), and createWall().

◆ createWall()

void multirankSimulation::createWall	(	int	xyz,
		int	plane,
		boundaryObject &	bObj
	)

make a wall with x, y, or z normal

/param xyz int that is 0, 1, or 2 for wall normal x, y, and z, respectively

References boundaryObject::boundary, createMultirankBoundaryObject(), degeneratePlanar, homeotropic, mConfiguration, boundaryObject::P1, boundaryObject::P2, rankTopology, scalar, and UNWRITTENCODE.

◆ createSphericalColloid()

void multirankSimulation::createSphericalColloid	(	scalar3	center,
		scalar	radius,
		boundaryObject &	bObj
	)

make a simple sphere, setting all points within radius of center to be the object

References boundaryObject::boundary, createMultirankBoundaryObject(), degeneratePlanar, homeotropic, norm(), boundaryObject::P1, boundaryObject::P2, qTensorFromDirector(), scalar, scalar3, and UNWRITTENCODE.

◆ createSphericalCavity()

void multirankSimulation::createSphericalCavity	(	scalar3	center,
		scalar	radius,
		boundaryObject &	bObj
	)

make a simple sphere, setting all points farther than radius of center to be the object

References boundaryObject::boundary, createMultirankBoundaryObject(), degeneratePlanar, homeotropic, mConfiguration, norm(), boundaryObject::P1, boundaryObject::P2, qTensorFromDirector(), rankTopology, scalar, scalar3, and UNWRITTENCODE.

◆ createCylindricalObject()

void multirankSimulation::createCylindricalObject	(	scalar3	cylinderStart,
		scalar3	cylinderEnd,
		scalar	radius,
		bool	colloidOrCapillary,
		boundaryObject &	bObj
	)

make a cylindrical object, with either the inside or outside defined as the object

define a cylinder by its two endpoints and its prependicular radius.

Parameters

colloidOrCapillary if true, points inside the radius are the object (i.e., a colloid), else points outside are the object (i.e., a surrounding capillary)

◆ createSpherocylinder()

void multirankSimulation::createSpherocylinder	(	scalar3	cylinderStart,
		scalar3	cylinderEnd,
		scalar	radius,
		boundaryObject &	bObj
	)

make a spherocylinder, defined by the start and end of the cylindrical section and the radius

References boundaryObject::boundary, createMultirankBoundaryObject(), degeneratePlanar, homeotropic, mConfiguration, norm(), boundaryObject::P1, boundaryObject::P2, pointSegmentDistance(), qTensorFromDirector(), rankTopology, scalar, scalar3, and UNWRITTENCODE.

◆ createBoundaryFromFile()

void multirankSimulation::createBoundaryFromFile	(	string	fname,
		bool	verbose = `false`
	)

virtual

import a boundary object from a (carefully prepared) text file

Reads a carefully prepared text file to create a new boundary object... For convenience this information is copied into the main README.md documentation file, so both should be updated together if needed.

The first line must be a single integer specifying the number of objects to be read in. This is meant in the colloquial English sense, not in the zero-indexed counting sense. So, if you want your file to specify one object, make sure the first line is the number 1.

Each subsequent block must be formatted as follows (with no additional lines between the blocks): The first line MUST be formatted as a b c d where a=0 means oriented anchoring (such as homeotropic or oriented planar), a=1 means degenerate Planar b is a scalar setting the anchoring strength c is the preferred value of S0 d is an integer specifying the number of sites, N_B.

Subsequently, there MUST be N_b lines of the form x y z C1 C2 C3 C4 C5, where x, y, and z are the integer lattice sites, and C1, C2, C3, C4, C5 are real numbers corresponding to the desired anchoring conditions: For oriented anchoring, C1, C2, C3, C4, C5 correspond directly to the surface-preferred Q-tensor: C1 = Qxx, C2 = Qxy, C3=Qxz, C4 = Qyy, C5=Qyz, where one often will set the Q-tensor by choosing a locally preferred director, \nu^s, and setting Q^B = 3 S_0/2*(\nu^s \nu^s - \delta_{ab}/3).

For degenerate planar anchoring the five constants should be specified as, C1 = \hat{nu}_x C2 = \hat{nu}_y C3 = \hat{nu}_z C4 = 0.0 C5 = 0.0, where \nu^s = {Cos[[Phi]] Sin[[theta]], Sin[[Phi]] Sin[[theta]], Cos[[theta]]} is the direction to which the LC should try to be orthogonal

References createMultirankBoundaryObject(), degeneratePlanar, homeotropic, and scalar.

◆ setDipolarField() [1/2]

void multirankSimulation::setDipolarField	(	scalar3	center,
		scalar	ThetaD,
		scalar	radius,
		scalar	range,
		scalar	S0
	)

a function of convenience... make a dipolar field a la Lubensky et al.

for lattice sites within range of center, create a z-aligned dipolar field with opening angle ThetaD about an object of radius "radius" Functional form from Lubensky, Pettey, Currier, and Stark. PRE 57, 610, 1998

References Cos, latticeMinPosition, mConfiguration, norm(), PI, qTensorFromDirector(), rankParity, rankTopology, scalar, scalar3, and Sin.

◆ setDipolarField() [2/2]

void multirankSimulation::setDipolarField	(	scalar3	center,
		scalar3	direction,
		scalar	radius,
		scalar	range,
		scalar	S0
	)

a function of convenience... make a dipolar field a la Lubensky et al. Simpler, uses the ravnik expression

for lattice sites within range of center, create a dipolar field in the direction specified about an object of radius "radius"

References latticeMinPosition, mConfiguration, norm(), qTensorFromDirector(), rankParity, rankTopology, scalar, and scalar3.

◆ communicateHaloSitesRoutine()

void multirankSimulation::communicateHaloSitesRoutine ( )

virtual

handles calls to all necessary halo site transfer

References communicationDirectionParity, communicationDirections, communicationTargets, ArrayHandle< T >::data, access_location::device, access_location::host, mConfiguration, MPI_SCALAR, mpiRequests, basicSimulation::nRanks, access_mode::overwrite, access_mode::read, synchronizeAndTransferBuffers(), and basicSimulation::useGPU.

Referenced by finalizeObjects(), moveParticles(), and setConfiguration().

◆ synchronizeAndTransferBuffers()

void multirankSimulation::synchronizeAndTransferBuffers ( )

virtual

synchronize mpi and make transfer buffers

References communicationDirections, mConfiguration, mpiRequests, mpiStatuses, basicSimulation::nRanks, transfersUpToDate, and basicSimulation::useGPU.

Referenced by communicateHaloSitesRoutine(), and computeForces().

◆ getPointer()

shared_ptr<multirankSimulation> multirankSimulation::getPointer ( )

inline

return a shared pointer to this Simulation

Referenced by addUpdater().

◆ setConfiguration()

void multirankSimulation::setConfiguration ( MConfigPtr _config )

Pass in a reference to the configuration.

Set a pointer to the configuration

References basicSimulation::Box, communicateHaloSitesRoutine(), latticeMinPosition, mConfiguration, and rankParity.

◆ computeForces()

void multirankSimulation::computeForces ( )

virtual

Call the force computer to compute the forces.

Calls all force computers, and evaluate the self force calculation if the model demands it

Implements basicSimulation.

References forceComputers, mConfiguration, synchronizeAndTransferBuffers(), and basicSimulation::useGPU.

◆ addUpdater() [1/2]

void multirankSimulation::addUpdater ( UpdaterPtr _upd )

inline

Add an updater.

References updaters.

◆ addUpdater() [2/2]

void multirankSimulation::addUpdater	(	UpdaterPtr	_upd,
		MConfigPtr	_config
	)

Add an updater with a reference to a configuration.

Add a pointer to the list of updaters, and give that updater a reference to the model...

References getPointer(), and updaters.

◆ addForce() [1/2]

virtual void multirankSimulation::addForce ( ForcePtr _force )

inlinevirtual

Add a force computer configuration.

References forceComputers.

◆ addForce() [2/2]

void multirankSimulation::addForce	(	ForcePtr	_force,
		MConfigPtr	_config
	)

virtual

Add a force computer configuration.

Add a pointer to the list of force computers, and give that FC a reference to the model...

References forceComputers.

◆ clearForceComputers()

void multirankSimulation::clearForceComputers ( )

inline

Clear out the vector of forceComputes.

References forceComputers.

◆ clearUpdaters()

void multirankSimulation::clearUpdaters ( )

inline

Clear out the vector of updaters.

References updaters.

◆ getMaxForce()

scalar multirankSimulation::getMaxForce ( )

inline

A utility function that just checks the first updater for a max force.

References updaters.

◆ performTimestep()

void multirankSimulation::performTimestep ( )

Call every updater to advance one time step.

References basicSimulation::integerTimestep, basicSimulation::integrationTimestep, basicSimulation::Time, transfersUpToDate, and updaters.

◆ sumUpdaterData()

void multirankSimulation::sumUpdaterData ( vector< scalar > & data )

virtual

manipulate data from updaters

Reimplemented from basicSimulation.

References data, dataBuffer, profiler::end(), MPI_SCALAR, basicSimulation::nRanks, p1, and profiler::start().

Referenced by computePotentialEnergy().

◆ computePotentialEnergy()

scalar multirankSimulation::computePotentialEnergy ( bool verbose = false )

virtual

compute the potential energy associated with all of the forces

Reimplemented from basicSimulation.

References forceComputers, basicSimulation::NActive, scalar, and sumUpdaterData().

Referenced by computeEnergy().

◆ computeKineticEnergy()

scalar multirankSimulation::computeKineticEnergy ( bool verbose = false )

virtual

compute the kinetic energy

Reimplemented from basicSimulation.

References mConfiguration.

Referenced by computeEnergy().

◆ computeEnergy()

virtual scalar multirankSimulation::computeEnergy ( bool verbose = false )

inlinevirtual

compute the total energy

References computeKineticEnergy(), and computePotentialEnergy().

◆ setIntegrationTimestep()

void multirankSimulation::setIntegrationTimestep ( scalar dt )

Set the simulation timestep.

Postcondition: the cell configuration and e.o.m. timestep is set to the input value

References basicSimulation::integrationTimestep, and updaters.

◆ setCPUOperation()

void multirankSimulation::setCPUOperation ( bool setcpu )

turn on CPU-only mode for all components

Postcondition: the cell configuration and e.o.m. timestep is set to the input value

References forceComputers, mConfiguration, updaters, and basicSimulation::useGPU.

◆ setReproducible()

void multirankSimulation::setReproducible ( bool reproducible )

Enforce reproducible dynamics.

Precondition: the updaters already know if the GPU will be used

Postcondition: the updaters are set to be reproducible if the boolean is true, otherwise the RNG is initialized

References updaters.

◆ saveState()

void multirankSimulation::saveState	(	string	fname,
		int	latticeSkip = `1`,
		int	defectType = `0`
	)

save a file for each rank recording the expanded lattice; lattice skip controls the sparsity of saved sites

Saves a file for each rank recording the current state of the system (each rank will produce an independent file, named StringJoin[fname,"_x(ToString[X])_y(ToString[Y])_z(ToString[Z]).txt"], where X, Y, and Z denote the position in the partitioning that the rank controls.

The latticeSkip parameter thins out the number of lattice sites that are saved (crucial for storing the results of large simulations). In the default setting of latticeSkip=1, every site is saved. Otherwise, only sites (x_i,y_i,j_i) will be saved, where (i mod(latticeSkip) == 0)

The format of the saved state is a simple text file where every line is the following: i j k Qxx Qxy Qxz Qyy Qyz type defectValue where i, j, and k are the coordinates of the lattice site (in the global frame), followed by the five independent components of the Q tensor, the "type" of the site, and the value computed by computeDefectMeasures(defectType). By default, this is just the maximum eigenvalue of the Q-tensor at the indicated site; see the documentation of computeDefectMeasures for more information.

References ArrayHandle< T >::data, access_location::host, idx, mConfiguration, rankParity, and access_mode::read.

◆ setRankTopology()

void multirankSimulation::setRankTopology	(	int	x,
		int	y,
		int	z
	)

protected

References Index3D::inverseIndex(), basicSimulation::myRank, basicSimulation::nRanks, parityTest, rankParity, and rankTopology.

Referenced by multirankSimulation().

◆ determineCommunicationPattern()

void multirankSimulation::determineCommunicationPattern	(	bool	_edges,
		bool	_corners
	)

protected

References communicationDirectionParity, communicationDirections, communicationTargets, corners, edges, mpiRequests, mpiStatuses, parityTest, rankParity, rankTopology, and Index3D::sizes.

Referenced by multirankSimulation().

◆ setBox()

void basicSimulation::setBox ( BoxPtr _box )

inherited

This changes the contents of the Box pointed to by Box to match that of _box.

Set a new Box for the simulation...This is the function that should be called to propagate a change in the box dimensions throughout the simulation...

References basicSimulation::Box, and basicSimulation::configuration.

◆ setSortPeriod()

void basicSimulation::setSortPeriod ( int sp )

inlineinherited

Set the time between spatial sorting operations.

References basicSimulation::sortPeriod.

◆ setCurrentTime()

void basicSimulation::setCurrentTime ( scalar _cTime )

virtualinherited

reset the simulation clock

Postcondition: the cell configuration and e.o.m. timestep is set to the input value

References basicSimulation::Time.

◆ setCurrentTimestep()

virtual void basicSimulation::setCurrentTimestep ( int _cTime )

inlinevirtualinherited

reset the simulation clock counter

References basicSimulation::integerTimestep.

Member Data Documentation

◆ p1

profiler multirankSimulation::p1 = profiler("total communication time")

Referenced by moveParticles(), and sumUpdaterData().

◆ p4

profiler multirankSimulation::p4 = profiler("MPI recv")

◆ p3

profiler multirankSimulation::p3 = profiler("MPI send")

◆ p2

profiler multirankSimulation::p2 = profiler("GPU data buffering kernels time")

◆ mConfiguration

WeakMConfigPtr multirankSimulation::mConfiguration

The configuration of latticeSites.

Referenced by communicateHaloSitesRoutine(), computeForces(), computeKineticEnergy(), createCylindricalObject(), createMultirankBoundaryObject(), createSphericalCavity(), createSpherocylinder(), createWall(), moveParticles(), saveState(), setConfiguration(), setCPUOperation(), setDipolarField(), and synchronizeAndTransferBuffers().

◆ updaters

vector<WeakUpdaterPtr> multirankSimulation::updaters

A vector of updaters that the simulation will loop through.

Referenced by addUpdater(), clearUpdaters(), finalizeObjects(), getMaxForce(), performTimestep(), setCPUOperation(), setIntegrationTimestep(), and setReproducible().

◆ forceComputers

vector<WeakForcePtr> multirankSimulation::forceComputers

A vector of force computes the simulation will loop through.

Referenced by addForce(), clearForceComputers(), computeForces(), computePotentialEnergy(), and setCPUOperation().

◆ latticeMinPosition

int3 multirankSimulation::latticeMinPosition

in multi-rank simulations, this stores the lowest (x,y,z) coordinate controlled by the current rank

Referenced by createMultirankBoundaryObject(), setConfiguration(), and setDipolarField().

◆ communicationDirections

vector<int2> multirankSimulation::communicationDirections

protected

Referenced by communicateHaloSitesRoutine(), determineCommunicationPattern(), and synchronizeAndTransferBuffers().

◆ communicationDirectionParity

vector<bool> multirankSimulation::communicationDirectionParity

protected

Referenced by communicateHaloSitesRoutine(), and determineCommunicationPattern().

◆ communicationTargets

vector<int> multirankSimulation::communicationTargets

protected

Referenced by communicateHaloSitesRoutine(), and determineCommunicationPattern().

◆ rankTopology

int3 multirankSimulation::rankTopology

protected

the number of ranks per {x,y,z} axis

Referenced by createCylindricalObject(), createMultirankBoundaryObject(), createSphericalCavity(), createSpherocylinder(), createWall(), determineCommunicationPattern(), setDipolarField(), and setRankTopology().

◆ rankParity

int3 multirankSimulation::rankParity

protected

the local {x,y,z} rank coordinate

Referenced by createMultirankBoundaryObject(), determineCommunicationPattern(), saveState(), setConfiguration(), setDipolarField(), and setRankTopology().

◆ parityTest

Index3D multirankSimulation::parityTest

protected

Referenced by determineCommunicationPattern(), and setRankTopology().

◆ edges

bool multirankSimulation::edges

protected

do edges need to be communicated?

Referenced by determineCommunicationPattern().

◆ corners

bool multirankSimulation::corners

protected

do corner sites?

Referenced by determineCommunicationPattern().

◆ transfersUpToDate

bool multirankSimulation::transfersUpToDate

protected

have the halo sites been communicated?

Referenced by moveParticles(), performTimestep(), and synchronizeAndTransferBuffers().

◆ mpiStatus

MPI_Status multirankSimulation::mpiStatus

protected

◆ mpiStatuses

vector<MPI_Status> multirankSimulation::mpiStatuses

protected

Referenced by determineCommunicationPattern(), and synchronizeAndTransferBuffers().

◆ mpiRequests

vector<MPI_Request> multirankSimulation::mpiRequests

protected

Referenced by communicateHaloSitesRoutine(), determineCommunicationPattern(), and synchronizeAndTransferBuffers().

◆ dataBuffer

vector<scalar> multirankSimulation::dataBuffer

protected

Referenced by sumUpdaterData().

◆ configuration

WeakConfigPtr basicSimulation::configuration

inherited

The configuration of particles.

Referenced by Simulation::computeForces(), Simulation::computeKineticEnergy(), Simulation::moveParticles(), basicSimulation::setBox(), Simulation::setConfiguration(), Simulation::setCPUOperation(), and Simulation::setNThreads().

◆ Box

BoxPtr basicSimulation::Box

inherited

The domain of the simulation.

Referenced by basicSimulation::basicSimulation(), basicSimulation::setBox(), Simulation::setConfiguration(), and setConfiguration().

◆ integerTimestep

int basicSimulation::integerTimestep

inherited

An integer that keeps track of how often performTimestep has been called.

Referenced by Simulation::performTimestep(), performTimestep(), and basicSimulation::setCurrentTimestep().

◆ Time

scalar basicSimulation::Time

inherited

The current simulation time.

Referenced by Simulation::performTimestep(), performTimestep(), and basicSimulation::setCurrentTime().

◆ integrationTimestep

scalar basicSimulation::integrationTimestep

inherited

The dt of a time step.

Referenced by Simulation::performTimestep(), performTimestep(), Simulation::setIntegrationTimestep(), and setIntegrationTimestep().

◆ useGPU

bool basicSimulation::useGPU

inherited

A flag controlling whether to use the GPU.

Referenced by communicateHaloSitesRoutine(), computeForces(), Simulation::setCPUOperation(), setCPUOperation(), and synchronizeAndTransferBuffers().

◆ myRank

int basicSimulation::myRank

inherited

integer for this rank

Referenced by multirankSimulation(), and setRankTopology().

◆ nRanks

int basicSimulation::nRanks

inherited

total number of ranks

Referenced by communicateHaloSitesRoutine(), setRankTopology(), sumUpdaterData(), and synchronizeAndTransferBuffers().

◆ NActive

int basicSimulation::NActive = 0

inherited

some measure of the number of active degrees of freedom

Referenced by computePotentialEnergy(), and finalizeObjects().

◆ sortPeriod

int basicSimulation::sortPeriod

protectedinherited

Determines how frequently the spatial sorter be called...once per sortPeriod Timesteps. When sortPeriod < 0 no sorting occurs.

Referenced by basicSimulation::setSortPeriod().

◆ spatialSortThisStep

bool basicSimulation::spatialSortThisStep

protectedinherited

A flag that determins if a spatial sorting is due to occur this Timestep.

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

Public Member Functions

Public Attributes

Protected Member Functions

Protected Attributes

Constructor & Destructor Documentation

◆ multirankSimulation()

Member Function Documentation

◆ moveParticles()

◆ finalizeObjects()

◆ createMultirankBoundaryObject()

◆ createWall()

◆ createSphericalColloid()

◆ createSphericalCavity()

◆ createCylindricalObject()

◆ createSpherocylinder()

◆ createBoundaryFromFile()

◆ setDipolarField() [1/2]

◆ setDipolarField() [2/2]

◆ communicateHaloSitesRoutine()

◆ synchronizeAndTransferBuffers()

◆ getPointer()

◆ setConfiguration()

◆ computeForces()

◆ addUpdater() [1/2]

◆ addUpdater() [2/2]

◆ addForce() [1/2]

◆ addForce() [2/2]

◆ clearForceComputers()

◆ clearUpdaters()

◆ getMaxForce()

◆ performTimestep()

◆ sumUpdaterData()

◆ computePotentialEnergy()

◆ computeKineticEnergy()

◆ computeEnergy()

◆ setIntegrationTimestep()

◆ setCPUOperation()

◆ setReproducible()

◆ saveState()

◆ setRankTopology()

◆ determineCommunicationPattern()

◆ setBox()

◆ setSortPeriod()

◆ setCurrentTime()

◆ setCurrentTimestep()

Member Data Documentation

◆ p1

◆ p4

◆ p3

◆ p2

◆ mConfiguration

◆ updaters

◆ forceComputers

◆ latticeMinPosition

◆ communicationDirections

◆ communicationDirectionParity

◆ communicationTargets

◆ rankTopology

◆ rankParity

◆ parityTest

◆ edges

◆ corners

◆ transfersUpToDate

◆ mpiStatus

◆ mpiStatuses

◆ mpiRequests

◆ dataBuffer

◆ configuration

◆ Box

◆ integerTimestep

◆ Time

◆ integrationTimestep

◆ useGPU

◆ myRank

◆ nRanks

◆ NActive

◆ sortPeriod

◆ spatialSortThisStep