FMassProcessorDependencySolver Struct Reference

#include <MassProcessorDependencySolver.h>

Classes
struct	FNode
struct	FResult

Public Member Functions
MASSENTITY_API	FMassProcessorDependencySolver (TArrayView< UMassProcessor *const > InProcessors, const bool bIsGameRuntime=true)
MASSENTITY_API void	ResolveDependencies (TArray< FMassProcessorOrderInfo > &OutResult, TSharedPtr< FMassEntityManager > EntityManager=nullptr, FResult *InOutOptionalResult=nullptr)
bool	IsSolvingForSingleThread () const

Static Public Member Functions
static MASSENTITY_API void	CreateSubGroupNames (FName InGroupName, TArray< FString > &SubGroupNames)
static bool	IsResultUpToDate (const FResult &InResult, TSharedPtr< FMassEntityManager > EntityManager)

Protected Member Functions
bool	PerformSolverStep (FResourceUsage &ResourceUsage, TArray< int32 > &InOutIndicesRemaining, TArray< int32 > &OutNodeIndices)
int32	CreateNodes (UMassProcessor &Processor)
void	BuildDependencies ()
void	Solve (TArray< FMassProcessorOrderInfo > &OutResult)
void	LogNode (const FNode &Node, int Indent=0)
void	GatherSubsystemInformation (const UE::Mass::FTypeManager &TypeManager)

Protected Attributes
TArrayView< UMassProcessor *const >	Processors
const bool	bSingleThreadTarget = bool(!MASS_DO_PARALLEL)
const bool	bGameRuntime = true
FString	DependencyGraphFileName
TArray< FNode >	AllNodes
TMap< FName, int32 >	NodeIndexMap
FMassExternalSubsystemBitSet	MultiThreadedSystemsBitSet

Constructor & Destructor Documentation

FMassProcessorDependencySolver()

FMassProcessorDependencySolver::FMassProcessorDependencySolver	(	TArrayView< UMassProcessor *const >	InProcessors,
		const bool	bIsGameRuntime = true
	)

Member Function Documentation

BuildDependencies()

void FMassProcessorDependencySolver::BuildDependencies ( )

protected

CreateNodes()

int32 FMassProcessorDependencySolver::CreateNodes ( UMassProcessor &  Processor )

protected

CreateSubGroupNames()

void FMassProcessorDependencySolver::CreateSubGroupNames ( FName  InGroupName, TArray< FString > &  SubGroupNames  )

static

GatherSubsystemInformation()

void FMassProcessorDependencySolver::GatherSubsystemInformation ( const UE::Mass::FTypeManager &  TypeManager )

protected

Finds out which subsystems handle multithreaded RW operations, and caches the result in MultiThreadedSystemsBitSet

IsResultUpToDate()

bool FMassProcessorDependencySolver::IsResultUpToDate ( const FResult &  InResult, TSharedPtr< FMassEntityManager >  EntityManager  )

static

Determines whether the dependency solving that produced InResult will produce different results if run with a given EntityManager

IsSolvingForSingleThread()

bool FMassProcessorDependencySolver::IsSolvingForSingleThread ( ) const

inline

LogNode()

void FMassProcessorDependencySolver::LogNode ( const FNode &  Node, int  Indent = 0  )

protected

PerformSolverStep()

bool FMassProcessorDependencySolver::PerformSolverStep ( FResourceUsage &  ResourceUsage, TArray< int32 > &  InOutIndicesRemaining, TArray< int32 > &  OutNodeIndices  )

protected

Traverses InOutIndicesRemaining in search of the first RootNode's node that has no dependencies left. Once found the node's index gets added to OutNodeIndices, removed from dependency lists from all other nodes and the function quits.

Returns

'true' if a dependency-less node has been found and added to OutNodeIndices; 'false' otherwise.

ResolveDependencies()

void FMassProcessorDependencySolver::ResolveDependencies	(	TArray< FMassProcessorOrderInfo > &	OutResult,
		TSharedPtr< FMassEntityManager >	EntityManager = nullptr,
		FMassProcessorDependencySolver::FResult *	InOutOptionalResult = nullptr
	)

Solve()

void FMassProcessorDependencySolver::Solve ( TArray< FMassProcessorOrderInfo > &  OutResult )

protected

Member Data Documentation

AllNodes

TArray<FNode> FMassProcessorDependencySolver::AllNodes

protected

bGameRuntime

const bool FMassProcessorDependencySolver::bGameRuntime = true

protected

bSingleThreadTarget

const bool FMassProcessorDependencySolver::bSingleThreadTarget = bool(!MASS_DO_PARALLEL)

protected

indicates whether we're generating processor order to be run in single-threaded or multithreaded environment (usually this means Dedicated Server vs Any other configuration). In Single-Threaded mode we can skip a bunch of expensive, fine-tuning tests. @Note currently the value depends on MASS_DO_PARALLEL and there's no way to configure it otherwise, but there's nothing inherently stopping us from letting users configure it.

DependencyGraphFileName

FString FMassProcessorDependencySolver::DependencyGraphFileName

protected

MultiThreadedSystemsBitSet

FMassExternalSubsystemBitSet FMassProcessorDependencySolver::MultiThreadedSystemsBitSet

protected

Stores the subsystems we know of that handle multithreaded access well - we filter those out, we don't need to consider them.

NodeIndexMap

TMap<FName, int32> FMassProcessorDependencySolver::NodeIndexMap

protected

Processors

TArrayView<UMassProcessor* const> FMassProcessorDependencySolver::Processors

protected

---

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

• Engine/Source/Runtime/MassEntity/Public/MassProcessorDependencySolver.h
• Engine/Source/Runtime/MassEntity/Private/MassProcessorDependencySolver.cpp