clang::CodeGen::CGOpenMPRuntimeGPU Class Reference

#include "/home/buildbot/as-worker-4/publish-doxygen-docs/llvm-project/clang/lib/CodeGen/CGOpenMPRuntimeGPU.h"

Inheritance diagram for clang::CodeGen::CGOpenMPRuntimeGPU:

Public Types
enum	ExecutionMode { EM_SPMD , EM_NonSPMD , EM_Unknown }
	Defines the execution mode. More...
enum	DataSharingMode { DS_CUDA , DS_Generic }
	Target codegen is specialized based on two data-sharing modes: CUDA, in which the local variables are actually global threadlocal, and Generic, in which the local variables are placed in global memory if they may escape their declaration context. More...

Public Member Functions
	CGOpenMPRuntimeGPU (CodeGenModule &CGM)
bool	isGPU () const override
	Returns true if the current target is a GPU.
bool	isDelayedVariableLengthDecl (CodeGenFunction &CGF, const VarDecl *VD) const override
	Declare generalized virtual functions which need to be defined by all specializations of OpenMPGPURuntime Targets like AMDGCN and NVPTX.
std::pair< llvm::Value *, llvm::Value * >	getKmpcAllocShared (CodeGenFunction &CGF, const VarDecl *VD) override
	Get call to __kmpc_alloc_shared.
void	getKmpcFreeShared (CodeGenFunction &CGF, const std::pair< llvm::Value *, llvm::Value * > &AddrSizePair) override
	Get call to __kmpc_free_shared.
llvm::Value *	getGPUWarpSize (CodeGenFunction &CGF)
	Get the GPU warp size.
llvm::Value *	getGPUThreadID (CodeGenFunction &CGF)
	Get the id of the current thread on the GPU.
llvm::Value *	getGPUNumThreads (CodeGenFunction &CGF)
	Get the maximum number of threads in a block of the GPU.
void	emitProcBindClause (CodeGenFunction &CGF, llvm::omp::ProcBindKind ProcBind, SourceLocation Loc) override
	Emit call to void __kmpc_push_proc_bind(ident_t *loc, kmp_int32 global_tid, int proc_bind) to generate code for 'proc_bind' clause.
void	emitNumThreadsClause (CodeGenFunction &CGF, llvm::Value *NumThreads, SourceLocation Loc) override
	Emits call to void __kmpc_push_num_threads(ident_t *loc, kmp_int32 global_tid, kmp_int32 num_threads) to generate code for 'num_threads' clause.
void	emitNumTeamsClause (CodeGenFunction &CGF, const Expr *NumTeams, const Expr *ThreadLimit, SourceLocation Loc) override
	This function ought to emit, in the general case, a call to.
llvm::Function *	emitParallelOutlinedFunction (CodeGenFunction &CGF, const OMPExecutableDirective &D, const VarDecl *ThreadIDVar, OpenMPDirectiveKind InnermostKind, const RegionCodeGenTy &CodeGen) override
	Emits inlined function for the specified OpenMP parallel.
llvm::Function *	emitTeamsOutlinedFunction (CodeGenFunction &CGF, const OMPExecutableDirective &D, const VarDecl *ThreadIDVar, OpenMPDirectiveKind InnermostKind, const RegionCodeGenTy &CodeGen) override
	Emits inlined function for the specified OpenMP teams.
void	emitTeamsCall (CodeGenFunction &CGF, const OMPExecutableDirective &D, SourceLocation Loc, llvm::Function *OutlinedFn, ArrayRef< llvm::Value * > CapturedVars) override
	Emits code for teams call of the OutlinedFn with variables captured in a record which address is stored in CapturedStruct.
void	emitParallelCall (CodeGenFunction &CGF, SourceLocation Loc, llvm::Function *OutlinedFn, ArrayRef< llvm::Value * > CapturedVars, const Expr *IfCond, llvm::Value *NumThreads) override
	Emits code for parallel or serial call of the OutlinedFn with variables captured in a record which address is stored in CapturedStruct.
void	emitBarrierCall (CodeGenFunction &CGF, SourceLocation Loc, OpenMPDirectiveKind Kind, bool EmitChecks=true, bool ForceSimpleCall=false) override
	Emit an implicit/explicit barrier for OpenMP threads.
void	emitCriticalRegion (CodeGenFunction &CGF, StringRef CriticalName, const RegionCodeGenTy &CriticalOpGen, SourceLocation Loc, const Expr *Hint=nullptr) override
	Emits a critical region.
void	emitReduction (CodeGenFunction &CGF, SourceLocation Loc, ArrayRef< const Expr * > Privates, ArrayRef< const Expr * > LHSExprs, ArrayRef< const Expr * > RHSExprs, ArrayRef< const Expr * > ReductionOps, ReductionOptionsTy Options) override
	Emit a code for reduction clause.
const VarDecl *	translateParameter (const FieldDecl *FD, const VarDecl *NativeParam) const override
	Translates the native parameter of outlined function if this is required for target.
Address	getParameterAddress (CodeGenFunction &CGF, const VarDecl *NativeParam, const VarDecl *TargetParam) const override
	Gets the address of the native argument basing on the address of the target-specific parameter.
void	emitOutlinedFunctionCall (CodeGenFunction &CGF, SourceLocation Loc, llvm::FunctionCallee OutlinedFn, ArrayRef< llvm::Value * > Args=std::nullopt) const override
	Emits call of the outlined function with the provided arguments, translating these arguments to correct target-specific arguments.
void	emitFunctionProlog (CodeGenFunction &CGF, const Decl *D) override
	Emits OpenMP-specific function prolog.
Address	getAddressOfLocalVariable (CodeGenFunction &CGF, const VarDecl *VD) override
	Gets the OpenMP-specific address of the local variable.
void	functionFinished (CodeGenFunction &CGF) override
	Cleans up references to the objects in finished function.
void	getDefaultDistScheduleAndChunk (CodeGenFunction &CGF, const OMPLoopDirective &S, OpenMPDistScheduleClauseKind &ScheduleKind, llvm::Value *&Chunk) const override
	Choose a default value for the dist_schedule clause.
void	getDefaultScheduleAndChunk (CodeGenFunction &CGF, const OMPLoopDirective &S, OpenMPScheduleClauseKind &ScheduleKind, const Expr *&ChunkExpr) const override
	Choose a default value for the schedule clause.
void	adjustTargetSpecificDataForLambdas (CodeGenFunction &CGF, const OMPExecutableDirective &D) const override
	Adjust some parameters for the target-based directives, like addresses of the variables captured by reference in lambdas.
void	processRequiresDirective (const OMPRequiresDecl *D) override
	Perform check on requires decl to ensure that target architecture supports unified addressing.
bool	hasAllocateAttributeForGlobalVar (const VarDecl *VD, LangAS &AS) override
	Checks if the variable has associated OMPAllocateDeclAttr attribute with the predefined allocator and translates it into the corresponding address space.
Public Member Functions inherited from clang::CodeGen::CGOpenMPRuntime
llvm::OpenMPIRBuilder &	getOMPBuilder ()
	CGOpenMPRuntime (CodeGenModule &CGM)
virtual	~CGOpenMPRuntime ()
virtual void	clear ()
llvm::Value *	emitUpdateLocation (CodeGenFunction &CGF, SourceLocation Loc, unsigned Flags=0, bool EmitLoc=false)
	Emits object of ident_t type with info for source location.
const Expr *	getNumTeamsExprForTargetDirective (CodeGenFunction &CGF, const OMPExecutableDirective &D, int32_t &MinTeamsVal, int32_t &MaxTeamsVal)
	Emit the number of teams for a target directive.
llvm::Value *	emitNumTeamsForTargetDirective (CodeGenFunction &CGF, const OMPExecutableDirective &D)
const Expr *	getNumThreadsExprForTargetDirective (CodeGenFunction &CGF, const OMPExecutableDirective &D, int32_t &UpperBound, bool UpperBoundOnly, llvm::Value **CondExpr=nullptr, const Expr **ThreadLimitExpr=nullptr)
	Check for a number of threads upper bound constant value (stored in UpperBound), or expression (returned).
llvm::Value *	emitNumThreadsForTargetDirective (CodeGenFunction &CGF, const OMPExecutableDirective &D)
	Emit an expression that denotes the number of threads a target region shall use.
llvm::Value *	emitTargetNumIterationsCall (CodeGenFunction &CGF, const OMPExecutableDirective &D, llvm::function_ref< llvm::Value *(CodeGenFunction &CGF, const OMPLoopDirective &D)> SizeEmitter)
	Return the trip count of loops associated with constructs / 'target teams distribute' and 'teams distribute parallel for'.
virtual bool	isGPU () const
	Returns true if the current target is a GPU.
virtual bool	isDelayedVariableLengthDecl (CodeGenFunction &CGF, const VarDecl *VD) const
	Check if the variable length declaration is delayed:
virtual std::pair< llvm::Value *, llvm::Value * >	getKmpcAllocShared (CodeGenFunction &CGF, const VarDecl *VD)
	Get call to __kmpc_alloc_shared.
virtual void	getKmpcFreeShared (CodeGenFunction &CGF, const std::pair< llvm::Value *, llvm::Value * > &AddrSizePair)
	Get call to __kmpc_free_shared.
void	emitIfClause (CodeGenFunction &CGF, const Expr *Cond, const RegionCodeGenTy &ThenGen, const RegionCodeGenTy &ElseGen)
	Emits code for OpenMP 'if' clause using specified CodeGen function.
std::string	getName (ArrayRef< StringRef > Parts) const
	Get the platform-specific name separator.
virtual void	emitUserDefinedReduction (CodeGenFunction *CGF, const OMPDeclareReductionDecl *D)
	Emit code for the specified user defined reduction construct.
virtual std::pair< llvm::Function *, llvm::Function * >	getUserDefinedReduction (const OMPDeclareReductionDecl *D)
	Get combiner/initializer for the specified user-defined reduction, if any.
void	emitUserDefinedMapper (const OMPDeclareMapperDecl *D, CodeGenFunction *CGF=nullptr)
	Emit the function for the user defined mapper construct.
llvm::Function *	getOrCreateUserDefinedMapperFunc (const OMPDeclareMapperDecl *D)
	Get the function for the specified user-defined mapper.
virtual llvm::Function *	emitParallelOutlinedFunction (CodeGenFunction &CGF, const OMPExecutableDirective &D, const VarDecl *ThreadIDVar, OpenMPDirectiveKind InnermostKind, const RegionCodeGenTy &CodeGen)
	Emits outlined function for the specified OpenMP parallel directive D.
virtual llvm::Function *	emitTeamsOutlinedFunction (CodeGenFunction &CGF, const OMPExecutableDirective &D, const VarDecl *ThreadIDVar, OpenMPDirectiveKind InnermostKind, const RegionCodeGenTy &CodeGen)
	Emits outlined function for the specified OpenMP teams directive D.
virtual llvm::Function *	emitTaskOutlinedFunction (const OMPExecutableDirective &D, const VarDecl *ThreadIDVar, const VarDecl *PartIDVar, const VarDecl *TaskTVar, OpenMPDirectiveKind InnermostKind, const RegionCodeGenTy &CodeGen, bool Tied, unsigned &NumberOfParts)
	Emits outlined function for the OpenMP task directive D.
virtual void	functionFinished (CodeGenFunction &CGF)
	Cleans up references to the objects in finished function.
virtual void	emitParallelCall (CodeGenFunction &CGF, SourceLocation Loc, llvm::Function *OutlinedFn, ArrayRef< llvm::Value * > CapturedVars, const Expr *IfCond, llvm::Value *NumThreads)
	Emits code for parallel or serial call of the OutlinedFn with variables captured in a record which address is stored in CapturedStruct.
virtual void	emitCriticalRegion (CodeGenFunction &CGF, StringRef CriticalName, const RegionCodeGenTy &CriticalOpGen, SourceLocation Loc, const Expr *Hint=nullptr)
	Emits a critical region.
virtual void	emitMasterRegion (CodeGenFunction &CGF, const RegionCodeGenTy &MasterOpGen, SourceLocation Loc)
	Emits a master region.
virtual void	emitMaskedRegion (CodeGenFunction &CGF, const RegionCodeGenTy &MaskedOpGen, SourceLocation Loc, const Expr *Filter=nullptr)
	Emits a masked region.
virtual void	emitTaskyieldCall (CodeGenFunction &CGF, SourceLocation Loc)
	Emits code for a taskyield directive.
virtual void	emitErrorCall (CodeGenFunction &CGF, SourceLocation Loc, Expr *ME, bool IsFatal)
	Emit __kmpc_error call for error directive extern void __kmpc_error(ident_t *loc, int severity, const char *message);.
virtual void	emitTaskgroupRegion (CodeGenFunction &CGF, const RegionCodeGenTy &TaskgroupOpGen, SourceLocation Loc)
	Emit a taskgroup region.
virtual void	emitSingleRegion (CodeGenFunction &CGF, const RegionCodeGenTy &SingleOpGen, SourceLocation Loc, ArrayRef< const Expr * > CopyprivateVars, ArrayRef< const Expr * > DestExprs, ArrayRef< const Expr * > SrcExprs, ArrayRef< const Expr * > AssignmentOps)
	Emits a single region.
virtual void	emitOrderedRegion (CodeGenFunction &CGF, const RegionCodeGenTy &OrderedOpGen, SourceLocation Loc, bool IsThreads)
	Emit an ordered region.
virtual void	emitBarrierCall (CodeGenFunction &CGF, SourceLocation Loc, OpenMPDirectiveKind Kind, bool EmitChecks=true, bool ForceSimpleCall=false)
	Emit an implicit/explicit barrier for OpenMP threads.
virtual bool	isStaticNonchunked (OpenMPScheduleClauseKind ScheduleKind, bool Chunked) const
	Check if the specified ScheduleKind is static non-chunked.
virtual bool	isStaticNonchunked (OpenMPDistScheduleClauseKind ScheduleKind, bool Chunked) const
	Check if the specified ScheduleKind is static non-chunked.
virtual bool	isStaticChunked (OpenMPScheduleClauseKind ScheduleKind, bool Chunked) const
	Check if the specified ScheduleKind is static chunked.
virtual bool	isStaticChunked (OpenMPDistScheduleClauseKind ScheduleKind, bool Chunked) const
	Check if the specified ScheduleKind is static non-chunked.
virtual bool	isDynamic (OpenMPScheduleClauseKind ScheduleKind) const
	Check if the specified ScheduleKind is dynamic.
virtual void	emitForDispatchInit (CodeGenFunction &CGF, SourceLocation Loc, const OpenMPScheduleTy &ScheduleKind, unsigned IVSize, bool IVSigned, bool Ordered, const DispatchRTInput &DispatchValues)
	Call the appropriate runtime routine to initialize it before start of loop.
virtual void	emitForStaticInit (CodeGenFunction &CGF, SourceLocation Loc, OpenMPDirectiveKind DKind, const OpenMPScheduleTy &ScheduleKind, const StaticRTInput &Values)
	Call the appropriate runtime routine to initialize it before start of loop.
virtual void	emitDistributeStaticInit (CodeGenFunction &CGF, SourceLocation Loc, OpenMPDistScheduleClauseKind SchedKind, const StaticRTInput &Values)
virtual void	emitForOrderedIterationEnd (CodeGenFunction &CGF, SourceLocation Loc, unsigned IVSize, bool IVSigned)
	Call the appropriate runtime routine to notify that we finished iteration of the ordered loop with the dynamic scheduling.
virtual void	emitForStaticFinish (CodeGenFunction &CGF, SourceLocation Loc, OpenMPDirectiveKind DKind)
	Call the appropriate runtime routine to notify that we finished all the work with current loop.
virtual llvm::Value *	emitForNext (CodeGenFunction &CGF, SourceLocation Loc, unsigned IVSize, bool IVSigned, Address IL, Address LB, Address UB, Address ST)
	Call __kmpc_dispatch_next( ident_t *loc, kmp_int32 tid, kmp_int32 *p_lastiter, kmp_int[32|64] *p_lower, kmp_int[32|64] *p_upper, kmp_int[32|64] *p_stride);.
virtual void	emitNumThreadsClause (CodeGenFunction &CGF, llvm::Value *NumThreads, SourceLocation Loc)
	Emits call to void __kmpc_push_num_threads(ident_t *loc, kmp_int32 global_tid, kmp_int32 num_threads) to generate code for 'num_threads' clause.
virtual void	emitProcBindClause (CodeGenFunction &CGF, llvm::omp::ProcBindKind ProcBind, SourceLocation Loc)
	Emit call to void __kmpc_push_proc_bind(ident_t *loc, kmp_int32 global_tid, int proc_bind) to generate code for 'proc_bind' clause.
virtual Address	getAddrOfThreadPrivate (CodeGenFunction &CGF, const VarDecl *VD, Address VDAddr, SourceLocation Loc)
	Returns address of the threadprivate variable for the current thread.
virtual ConstantAddress	getAddrOfDeclareTargetVar (const VarDecl *VD)
	Returns the address of the variable marked as declare target with link clause OR as declare target with to clause and unified memory.
virtual llvm::Function *	emitThreadPrivateVarDefinition (const VarDecl *VD, Address VDAddr, SourceLocation Loc, bool PerformInit, CodeGenFunction *CGF=nullptr)
	Emit a code for initialization of threadprivate variable.
virtual void	emitDeclareTargetFunction (const FunctionDecl *FD, llvm::GlobalValue *GV)
	Emit code for handling declare target functions in the runtime.
virtual Address	getAddrOfArtificialThreadPrivate (CodeGenFunction &CGF, QualType VarType, StringRef Name)
	Creates artificial threadprivate variable with name Name and type VarType.
virtual void	emitFlush (CodeGenFunction &CGF, ArrayRef< const Expr * > Vars, SourceLocation Loc, llvm::AtomicOrdering AO)
	Emit flush of the variables specified in 'omp flush' directive.
virtual void	emitTaskCall (CodeGenFunction &CGF, SourceLocation Loc, const OMPExecutableDirective &D, llvm::Function *TaskFunction, QualType SharedsTy, Address Shareds, const Expr *IfCond, const OMPTaskDataTy &Data)
	Emit task region for the task directive.
virtual void	emitTaskLoopCall (CodeGenFunction &CGF, SourceLocation Loc, const OMPLoopDirective &D, llvm::Function *TaskFunction, QualType SharedsTy, Address Shareds, const Expr *IfCond, const OMPTaskDataTy &Data)
	Emit task region for the taskloop directive.
virtual void	emitInlinedDirective (CodeGenFunction &CGF, OpenMPDirectiveKind InnermostKind, const RegionCodeGenTy &CodeGen, bool HasCancel=false)
	Emit code for the directive that does not require outlining.
llvm::Function *	emitReductionFunction (StringRef ReducerName, SourceLocation Loc, llvm::Type *ArgsElemType, ArrayRef< const Expr * > Privates, ArrayRef< const Expr * > LHSExprs, ArrayRef< const Expr * > RHSExprs, ArrayRef< const Expr * > ReductionOps)
	Emits reduction function.
void	emitSingleReductionCombiner (CodeGenFunction &CGF, const Expr *ReductionOp, const Expr *PrivateRef, const DeclRefExpr *LHS, const DeclRefExpr *RHS)
	Emits single reduction combiner.
virtual void	emitReduction (CodeGenFunction &CGF, SourceLocation Loc, ArrayRef< const Expr * > Privates, ArrayRef< const Expr * > LHSExprs, ArrayRef< const Expr * > RHSExprs, ArrayRef< const Expr * > ReductionOps, ReductionOptionsTy Options)
	Emit a code for reduction clause.
virtual llvm::Value *	emitTaskReductionInit (CodeGenFunction &CGF, SourceLocation Loc, ArrayRef< const Expr * > LHSExprs, ArrayRef< const Expr * > RHSExprs, const OMPTaskDataTy &Data)
	Emit a code for initialization of task reduction clause.
virtual void	emitTaskReductionFini (CodeGenFunction &CGF, SourceLocation Loc, bool IsWorksharingReduction)
	Emits the following code for reduction clause with task modifier:
virtual void	emitTaskReductionFixups (CodeGenFunction &CGF, SourceLocation Loc, ReductionCodeGen &RCG, unsigned N)
	Required to resolve existing problems in the runtime.
virtual Address	getTaskReductionItem (CodeGenFunction &CGF, SourceLocation Loc, llvm::Value *ReductionsPtr, LValue SharedLVal)
	Get the address of void * type of the privatue copy of the reduction item specified by the SharedLVal.
virtual void	emitTaskwaitCall (CodeGenFunction &CGF, SourceLocation Loc, const OMPTaskDataTy &Data)
	Emit code for 'taskwait' directive.
virtual void	emitCancellationPointCall (CodeGenFunction &CGF, SourceLocation Loc, OpenMPDirectiveKind CancelRegion)
	Emit code for 'cancellation point' construct.
virtual void	emitCancelCall (CodeGenFunction &CGF, SourceLocation Loc, const Expr *IfCond, OpenMPDirectiveKind CancelRegion)
	Emit code for 'cancel' construct.
virtual void	emitTargetOutlinedFunction (const OMPExecutableDirective &D, StringRef ParentName, llvm::Function *&OutlinedFn, llvm::Constant *&OutlinedFnID, bool IsOffloadEntry, const RegionCodeGenTy &CodeGen)
	Emit outilined function for 'target' directive.
virtual void	emitTargetCall (CodeGenFunction &CGF, const OMPExecutableDirective &D, llvm::Function *OutlinedFn, llvm::Value *OutlinedFnID, const Expr *IfCond, llvm::PointerIntPair< const Expr *, 2, OpenMPDeviceClauseModifier > Device, llvm::function_ref< llvm::Value *(CodeGenFunction &CGF, const OMPLoopDirective &D)> SizeEmitter)
	Emit the target offloading code associated with D.
virtual bool	emitTargetFunctions (GlobalDecl GD)
	Emit the target regions enclosed in GD function definition or the function itself in case it is a valid device function.
virtual bool	emitTargetGlobalVariable (GlobalDecl GD)
	Emit the global variable if it is a valid device global variable.
virtual void	registerTargetGlobalVariable (const VarDecl *VD, llvm::Constant *Addr)
	Checks if the provided global decl GD is a declare target variable and registers it when emitting code for the host.
virtual bool	emitTargetGlobal (GlobalDecl GD)
	Emit the global GD if it is meaningful for the target.
void	createOffloadEntriesAndInfoMetadata ()
	Creates all the offload entries in the current compilation unit along with the associated metadata.
virtual void	emitTeamsCall (CodeGenFunction &CGF, const OMPExecutableDirective &D, SourceLocation Loc, llvm::Function *OutlinedFn, ArrayRef< llvm::Value * > CapturedVars)
	Emits code for teams call of the OutlinedFn with variables captured in a record which address is stored in CapturedStruct.
virtual void	emitNumTeamsClause (CodeGenFunction &CGF, const Expr *NumTeams, const Expr *ThreadLimit, SourceLocation Loc)
	Emits call to void __kmpc_push_num_teams(ident_t *loc, kmp_int32 global_tid, kmp_int32 num_teams, kmp_int32 thread_limit) to generate code for num_teams clause.
virtual void	emitThreadLimitClause (CodeGenFunction &CGF, const Expr *ThreadLimit, SourceLocation Loc)
	Emits call to void __kmpc_set_thread_limit(ident_t *loc, kmp_int32 global_tid, kmp_int32 thread_limit) to generate code for thread_limit clause on target directive.
virtual void	emitTargetDataCalls (CodeGenFunction &CGF, const OMPExecutableDirective &D, const Expr *IfCond, const Expr *Device, const RegionCodeGenTy &CodeGen, CGOpenMPRuntime::TargetDataInfo &Info)
	Emit the target data mapping code associated with D.
virtual void	emitTargetDataStandAloneCall (CodeGenFunction &CGF, const OMPExecutableDirective &D, const Expr *IfCond, const Expr *Device)
	Emit the data mapping/movement code associated with the directive D that should be of the form 'target [{enter|exit} data | update]'.
virtual void	emitDeclareSimdFunction (const FunctionDecl *FD, llvm::Function *Fn)
	Marks function Fn with properly mangled versions of vector functions.
virtual void	emitDoacrossInit (CodeGenFunction &CGF, const OMPLoopDirective &D, ArrayRef< Expr * > NumIterations)
	Emit initialization for doacross loop nesting support.
virtual void	emitDoacrossOrdered (CodeGenFunction &CGF, const OMPDependClause *C)
	Emit code for doacross ordered directive with 'depend' clause.
virtual void	emitDoacrossOrdered (CodeGenFunction &CGF, const OMPDoacrossClause *C)
	Emit code for doacross ordered directive with 'doacross' clause.
virtual const VarDecl *	translateParameter (const FieldDecl *FD, const VarDecl *NativeParam) const
	Translates the native parameter of outlined function if this is required for target.
virtual Address	getParameterAddress (CodeGenFunction &CGF, const VarDecl *NativeParam, const VarDecl *TargetParam) const
	Gets the address of the native argument basing on the address of the target-specific parameter.
virtual void	getDefaultDistScheduleAndChunk (CodeGenFunction &CGF, const OMPLoopDirective &S, OpenMPDistScheduleClauseKind &ScheduleKind, llvm::Value *&Chunk) const
	Choose default schedule type and chunk value for the dist_schedule clause.
virtual void	getDefaultScheduleAndChunk (CodeGenFunction &CGF, const OMPLoopDirective &S, OpenMPScheduleClauseKind &ScheduleKind, const Expr *&ChunkExpr) const
	Choose default schedule type and chunk value for the schedule clause.
virtual void	emitOutlinedFunctionCall (CodeGenFunction &CGF, SourceLocation Loc, llvm::FunctionCallee OutlinedFn, ArrayRef< llvm::Value * > Args=std::nullopt) const
	Emits call of the outlined function with the provided arguments, translating these arguments to correct target-specific arguments.
virtual void	emitFunctionProlog (CodeGenFunction &CGF, const Decl *D)
	Emits OpenMP-specific function prolog.
virtual Address	getAddressOfLocalVariable (CodeGenFunction &CGF, const VarDecl *VD)
	Gets the OpenMP-specific address of the local variable.
bool	markAsGlobalTarget (GlobalDecl GD)
	Marks the declaration as already emitted for the device code and returns true, if it was marked already, and false, otherwise.
void	emitDeferredTargetDecls () const
	Emit deferred declare target variables marked for deferred emission.
virtual void	adjustTargetSpecificDataForLambdas (CodeGenFunction &CGF, const OMPExecutableDirective &D) const
	Adjust some parameters for the target-based directives, like addresses of the variables captured by reference in lambdas.
virtual void	processRequiresDirective (const OMPRequiresDecl *D)
	Perform check on requires decl to ensure that target architecture supports unified addressing.
llvm::AtomicOrdering	getDefaultMemoryOrdering () const
	Gets default memory ordering as specified in requires directive.
virtual bool	hasAllocateAttributeForGlobalVar (const VarDecl *VD, LangAS &AS)
	Checks if the variable has associated OMPAllocateDeclAttr attribute with the predefined allocator and translates it into the corresponding address space.
bool	hasRequiresUnifiedSharedMemory () const
	Return whether the unified_shared_memory has been specified.
bool	isNontemporalDecl (const ValueDecl *VD) const
	Checks if the VD variable is marked as nontemporal declaration in current context.
Address	emitLastprivateConditionalInit (CodeGenFunction &CGF, const VarDecl *VD)
	Create specialized alloca to handle lastprivate conditionals.
virtual void	checkAndEmitLastprivateConditional (CodeGenFunction &CGF, const Expr *LHS)
	Checks if the provided LVal is lastprivate conditional and emits the code to update the value of the original variable.
virtual void	checkAndEmitSharedLastprivateConditional (CodeGenFunction &CGF, const OMPExecutableDirective &D, const llvm::DenseSet< CanonicalDeclPtr< const VarDecl > > &IgnoredDecls)
	Checks if the lastprivate conditional was updated in inner region and writes the value.
virtual void	emitLastprivateConditionalFinalUpdate (CodeGenFunction &CGF, LValue PrivLVal, const VarDecl *VD, SourceLocation Loc)
	Gets the address of the global copy used for lastprivate conditional update, if any.
std::pair< llvm::Value *, Address >	emitDependClause (CodeGenFunction &CGF, ArrayRef< OMPTaskDataTy::DependData > Dependencies, SourceLocation Loc)
	Emits list of dependecies based on the provided data (array of dependence/expression pairs).
Address	emitDepobjDependClause (CodeGenFunction &CGF, const OMPTaskDataTy::DependData &Dependencies, SourceLocation Loc)
	Emits list of dependecies based on the provided data (array of dependence/expression pairs) for depobj construct.
void	emitDestroyClause (CodeGenFunction &CGF, LValue DepobjLVal, SourceLocation Loc)
	Emits the code to destroy the dependency object provided in depobj directive.
void	emitUpdateClause (CodeGenFunction &CGF, LValue DepobjLVal, OpenMPDependClauseKind NewDepKind, SourceLocation Loc)
	Updates the dependency kind in the specified depobj object.
void	emitUsesAllocatorsInit (CodeGenFunction &CGF, const Expr *Allocator, const Expr *AllocatorTraits)
	Initializes user defined allocators specified in the uses_allocators clauses.
void	emitUsesAllocatorsFini (CodeGenFunction &CGF, const Expr *Allocator)
	Destroys user defined allocators specified in the uses_allocators clause.
bool	isLocalVarInUntiedTask (CodeGenFunction &CGF, const VarDecl *VD) const
	Returns true if the variable is a local variable in untied task.

Protected Member Functions
bool	isDefaultLocationConstant () const override
	Check if the default location must be constant.
Protected Member Functions inherited from clang::CodeGen::CGOpenMPRuntime
void	computeMinAndMaxThreadsAndTeams (const OMPExecutableDirective &D, CodeGenFunction &CGF, int32_t &MinThreadsVal, int32_t &MaxThreadsVal, int32_t &MinTeamsVal, int32_t &MaxTeamsVal)
	Helper to determine the min/max number of threads/teams for D.
virtual void	emitTargetOutlinedFunctionHelper (const OMPExecutableDirective &D, StringRef ParentName, llvm::Function *&OutlinedFn, llvm::Constant *&OutlinedFnID, bool IsOffloadEntry, const RegionCodeGenTy &CodeGen)
	Helper to emit outlined function for 'target' directive.
llvm::Type *	getIdentTyPointerTy ()
	Returns pointer to ident_t type.
llvm::Value *	getThreadID (CodeGenFunction &CGF, SourceLocation Loc)
	Gets thread id value for the current thread.
std::string	getOutlinedHelperName (StringRef Name) const
	Get the function name of an outlined region.
std::string	getOutlinedHelperName (CodeGenFunction &CGF) const
std::string	getReductionFuncName (StringRef Name) const
	Get the function name of a reduction function.
void	emitCall (CodeGenFunction &CGF, SourceLocation Loc, llvm::FunctionCallee Callee, ArrayRef< llvm::Value * > Args=std::nullopt) const
	Emits Callee function call with arguments Args with location Loc.
virtual Address	emitThreadIDAddress (CodeGenFunction &CGF, SourceLocation Loc)
	Emits address of the word in a memory where current thread id is stored.
void	setLocThreadIdInsertPt (CodeGenFunction &CGF, bool AtCurrentPoint=false)
void	clearLocThreadIdInsertPt (CodeGenFunction &CGF)
virtual bool	isDefaultLocationConstant () const
	Check if the default location must be constant.
virtual unsigned	getDefaultLocationReserved2Flags () const
	Returns additional flags that can be stored in reserved_2 field of the default location.
llvm::ArrayType *	getKmpCriticalNameTy () const
	Get the LLVM type for the critical name.
llvm::Value *	getCriticalRegionLock (StringRef CriticalName)
	Returns corresponding lock object for the specified critical region name.
void	scanForTargetRegionsFunctions (const Stmt *S, StringRef ParentName)
	Start scanning from statement S and emit all target regions found along the way.
void	emitKmpRoutineEntryT (QualType KmpInt32Ty)
	Build type kmp_routine_entry_t (if not built yet).
llvm::Type *	getKmpc_MicroPointerTy ()
	Returns pointer to kmpc_micro type.
llvm::Constant *	getOrCreateThreadPrivateCache (const VarDecl *VD)
	If the specified mangled name is not in the module, create and return threadprivate cache object.
void	emitThreadPrivateVarInit (CodeGenFunction &CGF, Address VDAddr, llvm::Value *Ctor, llvm::Value *CopyCtor, llvm::Value *Dtor, SourceLocation Loc)
	Emits initialization code for the threadprivate variables.
void	emitUDMapperArrayInitOrDel (CodeGenFunction &MapperCGF, llvm::Value *Handle, llvm::Value *BasePtr, llvm::Value *Ptr, llvm::Value *Size, llvm::Value *MapType, llvm::Value *MapName, CharUnits ElementSize, llvm::BasicBlock *ExitBB, bool IsInit)
	Emit the array initialization or deletion portion for user-defined mapper code generation.
TaskResultTy	emitTaskInit (CodeGenFunction &CGF, SourceLocation Loc, const OMPExecutableDirective &D, llvm::Function *TaskFunction, QualType SharedsTy, Address Shareds, const OMPTaskDataTy &Data)
	Emit task region for the task directive.
void	emitLastprivateConditionalUpdate (CodeGenFunction &CGF, LValue IVLVal, StringRef UniqueDeclName, LValue LVal, SourceLocation Loc)
	Emit update for lastprivate conditional data.
std::pair< llvm::Value *, LValue >	getDepobjElements (CodeGenFunction &CGF, LValue DepobjLVal, SourceLocation Loc)
	Returns the number of the elements and the address of the depobj dependency array.
SmallVector< llvm::Value *, 4 >	emitDepobjElementsSizes (CodeGenFunction &CGF, QualType &KmpDependInfoTy, const OMPTaskDataTy::DependData &Data)
void	emitDepobjElements (CodeGenFunction &CGF, QualType &KmpDependInfoTy, LValue PosLVal, const OMPTaskDataTy::DependData &Data, Address DependenciesArray)

Additional Inherited Members
Static Public Member Functions inherited from clang::CodeGen::CGOpenMPRuntime
static const Stmt *	getSingleCompoundChild (ASTContext &Ctx, const Stmt *Body)
	Checks if the Body is the CompoundStmt and returns its child statement iff there is only one that is not evaluatable at the compile time.
Protected Types inherited from clang::CodeGen::CGOpenMPRuntime
typedef llvm::DenseMap< SourceLocation, llvm::Value * >	OpenMPDebugLocMapTy
	Map for SourceLocation and OpenMP runtime library debug locations.
typedef llvm::DenseMap< llvm::Function *, DebugLocThreadIdTy >	OpenMPLocThreadIDMapTy
	Map of local debug location, ThreadId and functions.
typedef llvm::DenseMap< const OMPDeclareReductionDecl *, std::pair< llvm::Function *, llvm::Function * > >	UDRMapTy
	Map of UDRs and corresponding combiner/initializer.
typedef llvm::DenseMap< llvm::Function *, SmallVector< const OMPDeclareReductionDecl *, 4 > >	FunctionUDRMapTy
	Map of functions and locally defined UDRs.
using	FunctionUDMMapTy = llvm::DenseMap< llvm::Function *, SmallVector< const OMPDeclareMapperDecl *, 4 > >
	Map of functions and their local user-defined mappers.
using	NontemporalDeclsSet = llvm::SmallDenseSet< CanonicalDeclPtr< const Decl > >
using	UntiedLocalVarsAddressesMap = llvm::MapVector< CanonicalDeclPtr< const VarDecl >, std::pair< Address, Address > >
Static Protected Member Functions inherited from clang::CodeGen::CGOpenMPRuntime
static unsigned	getDefaultFlagsForBarriers (OpenMPDirectiveKind Kind)
	Returns default flags for the barriers depending on the directive, for which this barier is going to be emitted.
Protected Attributes inherited from clang::CodeGen::CGOpenMPRuntime
CodeGenModule &	CGM
llvm::OpenMPIRBuilder	OMPBuilder
	An OpenMP-IR-Builder instance.
OpenMPDebugLocMapTy	OpenMPDebugLocMap
llvm::FunctionType *	Kmpc_MicroTy = nullptr
	The type for a microtask which gets passed to __kmpc_fork_call().
OpenMPLocThreadIDMapTy	OpenMPLocThreadIDMap
UDRMapTy	UDRMap
FunctionUDRMapTy	FunctionUDRMap
llvm::DenseMap< const OMPDeclareMapperDecl *, llvm::Function * >	UDMMap
	Map from the user-defined mapper declaration to its corresponding functions.
FunctionUDMMapTy	FunctionUDMMap
llvm::DenseMap< llvm::Function *, llvm::DenseMap< CanonicalDeclPtr< const Decl >, std::tuple< QualType, const FieldDecl *, const FieldDecl *, LValue > > >	LastprivateConditionalToTypes
	Maps local variables marked as lastprivate conditional to their internal types.
llvm::DenseMap< llvm::Function *, unsigned >	FunctionToUntiedTaskStackMap
	Maps function to the position of the untied task locals stack.
llvm::ArrayType *	KmpCriticalNameTy
	Type kmp_critical_name, originally defined as typedef kmp_int32 kmp_critical_name[8];.
llvm::StringMap< llvm::AssertingVH< llvm::GlobalVariable >, llvm::BumpPtrAllocator >	InternalVars
	An ordered map of auto-generated variables to their unique names.
llvm::Type *	KmpRoutineEntryPtrTy = nullptr
	Type typedef kmp_int32 (* kmp_routine_entry_t)(kmp_int32, void *);.
QualType	KmpRoutineEntryPtrQTy
QualType	KmpTaskTQTy
	Type typedef struct kmp_task { void * shareds; /‍**< pointer to block of pointers to shared vars ‍/ kmp_routine_entry_t routine; /‍**< pointer to routine to call for executing task *‍/ kmp_int32 part_id; /‍**< part id for the task *‍/ kmp_routine_entry_t destructors; / pointer to function to invoke deconstructors of firstprivate C++ objects *‍/ } kmp_task_t;.
QualType	SavedKmpTaskTQTy
	Saved kmp_task_t for task directive.
QualType	SavedKmpTaskloopTQTy
	Saved kmp_task_t for taskloop-based directive.
QualType	KmpDependInfoTy
	Type typedef struct kmp_depend_info { kmp_intptr_t base_addr; size_t len; struct { bool in:1; bool out:1; } flags; } kmp_depend_info_t;.
QualType	KmpTaskAffinityInfoTy
	Type typedef struct kmp_task_affinity_info { kmp_intptr_t base_addr; size_t len; struct { bool flag1 : 1; bool flag2 : 1; kmp_int32 reserved : 30; } flags; } kmp_task_affinity_info_t;.
QualType	KmpDimTy
	struct kmp_dim { // loop bounds info casted to kmp_int64 kmp_int64 lo; // lower kmp_int64 up; // upper kmp_int64 st; // stride };
bool	ShouldMarkAsGlobal = true
llvm::DenseSet< CanonicalDeclPtr< const Decl > >	AlreadyEmittedTargetDecls
	List of the emitted declarations.
llvm::StringMap< llvm::WeakTrackingVH >	EmittedNonTargetVariables
	List of the global variables with their addresses that should not be emitted for the target.
llvm::SmallDenseSet< const VarDecl * >	DeferredGlobalVariables
	List of variables that can become declare target implicitly and, thus, must be emitted.
llvm::SmallVector< NontemporalDeclsSet, 4 >	NontemporalDeclsStack
	Stack for list of declarations in current context marked as nontemporal.
llvm::SmallVector< UntiedLocalVarsAddressesMap, 4 >	UntiedLocalVarsStack
llvm::SmallVector< LastprivateConditionalData, 4 >	LastprivateConditionalStack
	Stack for list of addresses of declarations in current context marked as lastprivate conditional.
bool	HasRequiresUnifiedSharedMemory = false
	Flag for keeping track of weather a requires unified_shared_memory directive is present.
llvm::AtomicOrdering	RequiresAtomicOrdering = llvm::AtomicOrdering::Monotonic
	Atomic ordering from the omp requires directive.
bool	HasEmittedTargetRegion = false
	Flag for keeping track of weather a target region has been emitted.
bool	HasEmittedDeclareTargetRegion = false
	Flag for keeping track of weather a device routine has been emitted.
llvm::StringSet	ThreadPrivateWithDefinition
	Set of threadprivate variables with the generated initializer.
llvm::StringSet	DeclareTargetWithDefinition
	Set of declare target variables with the generated initializer.

Detailed Description

Definition at line 24 of file CGOpenMPRuntimeGPU.h.

Member Enumeration Documentation

DataSharingMode

enum clang::CodeGen::CGOpenMPRuntimeGPU::DataSharingMode

Target codegen is specialized based on two data-sharing modes: CUDA, in which the local variables are actually global threadlocal, and Generic, in which the local variables are placed in global memory if they may escape their declaration context.

Enumerator
DS_CUDA	CUDA data sharing mode.
DS_Generic	Generic data-sharing mode.

Definition at line 40 of file CGOpenMPRuntimeGPU.h.

ExecutionMode

enum clang::CodeGen::CGOpenMPRuntimeGPU::ExecutionMode

Defines the execution mode.

Enumerator
EM_SPMD	SPMD execution mode (all threads are worker threads).
EM_NonSPMD	Non-SPMD execution mode (1 master thread, others are workers).
EM_Unknown	Unknown execution mode (orphaned directive).

Definition at line 27 of file CGOpenMPRuntimeGPU.h.

Constructor & Destructor Documentation

CGOpenMPRuntimeGPU()

CGOpenMPRuntimeGPU::CGOpenMPRuntimeGPU ( CodeGenModule &  CGM )

explicit

Definition at line 890 of file CGOpenMPRuntimeGPU.cpp.

References clang::CodeGen::CGOpenMPRuntime::CGM, DS_CUDA, clang::CodeGen::CodeGenModule::getLangOpts(), clang::CodeGen::CGOpenMPRuntime::getOMPBuilder(), clang::CodeGen::CGOpenMPRuntime::hasRequiresUnifiedSharedMemory(), isGPU(), clang::CodeGen::CGOpenMPRuntime::OMPBuilder, and clang::LangOptions::OMPHostIRFile.

Member Function Documentation

adjustTargetSpecificDataForLambdas()

void CGOpenMPRuntimeGPU::adjustTargetSpecificDataForLambdas ( CodeGenFunction &  CGF, const OMPExecutableDirective &  D  ) const

overridevirtual

Adjust some parameters for the target-based directives, like addresses of the variables captured by reference in lambdas.

Reimplemented from clang::CodeGen::CGOpenMPRuntime.

Definition at line 3361 of file CGOpenMPRuntimeGPU.cpp.

References clang::C, clang::CodeGen::CodeGenFunction::CapturedStmtInfo, clang::CapturedStmt::captures(), clang::CapturedStmt::capturesVariable(), clang::CXXThis, clang::CodeGen::CodeGenFunction::EmitLoadOfReferenceLValue(), clang::CodeGen::CodeGenFunction::EmitLValueForFieldInitialization(), clang::CodeGen::Address::emitRawPointer(), clang::CodeGen::CodeGenFunction::EmitStoreOfScalar(), clang::CodeGen::LValue::getAddress(), clang::CodeGen::CodeGenFunction::GetAddrOfLocalVar(), clang::Type::getAsCXXRecordDecl(), clang::QualType::getCanonicalType(), clang::OMPExecutableDirective::getCapturedStmt(), clang::OMPExecutableDirective::getDirectiveKind(), clang::QualType::getNonReferenceType(), clang::ValueDecl::getType(), clang::CodeGen::CodeGenFunction::CGCapturedStmtInfo::isCXXThisExprCaptured(), clang::isOpenMPTargetExecutionDirective(), clang::Type::isReferenceType(), clang::LCK_ByRef, clang::CodeGen::CodeGenFunction::LoadCXXThis(), and clang::CodeGen::CodeGenFunction::MakeAddrLValue().

emitBarrierCall()

void CGOpenMPRuntimeGPU::emitBarrierCall ( CodeGenFunction &  CGF, SourceLocation  Loc, OpenMPDirectiveKind  Kind, bool  EmitChecks = true, bool  ForceSimpleCall = false  )

overridevirtual

Emit an implicit/explicit barrier for OpenMP threads.

Parameters

Kind	Directive for which this implicit barrier call must be generated. Must be OMPD_barrier for explicit barrier generation.
EmitChecks	true if need to emit checks for cancellation barriers.
ForceSimpleCall	true simple barrier call must be emitted, false if runtime class decides which one to emit (simple or with cancellation checks).

Reimplemented from clang::CodeGen::CGOpenMPRuntime.

Definition at line 1325 of file CGOpenMPRuntimeGPU.cpp.

References clang::CodeGen::CGOpenMPRuntime::CGM, clang::CodeGen::CodeGenFunction::EmitRuntimeCall(), clang::CodeGen::CGOpenMPRuntime::emitUpdateLocation(), clang::CodeGen::CGOpenMPRuntime::getDefaultFlagsForBarriers(), clang::CodeGen::CodeGenModule::getModule(), clang::CodeGen::CGOpenMPRuntime::getThreadID(), clang::CodeGen::CodeGenFunction::HaveInsertPoint(), and clang::CodeGen::CGOpenMPRuntime::OMPBuilder.

emitCriticalRegion()

void CGOpenMPRuntimeGPU::emitCriticalRegion ( CodeGenFunction &  CGF, StringRef  CriticalName, const RegionCodeGenTy &  CriticalOpGen, SourceLocation  Loc, const Expr *  Hint = nullptr  )

overridevirtual

Emits a critical region.

Parameters

CriticalName	Name of the critical region.
CriticalOpGen	Generator for the statement associated with the given critical region.
Hint	Value of the 'hint' clause (optional).

Reimplemented from clang::CodeGen::CGOpenMPRuntime.

Definition at line 1342 of file CGOpenMPRuntimeGPU.cpp.

References clang::CodeGen::CodeGenFunction::Builder, clang::CodeGen::CGOpenMPRuntime::CGM, clang::CodeGen::CodeGenFunction::CGM, clang::CodeGen::CodeGenFunction::createBasicBlock(), clang::CodeGen::CodeGenFunction::CreateMemTemp(), clang::CodeGen::CodeGenFunction::EmitBlock(), clang::CodeGen::CodeGenFunction::EmitBranch(), clang::CodeGen::CGOpenMPRuntime::emitCriticalRegion(), clang::CodeGen::CodeGenFunction::EmitLoadOfScalar(), clang::CodeGen::CodeGenFunction::EmitRuntimeCall(), clang::CodeGen::CodeGenFunction::EmitStoreOfScalar(), clang::CodeGen::CodeGenFunction::getContext(), clang::ASTContext::getIntTypeForBitwidth(), clang::CodeGen::CodeGenModule::getModule(), clang::CodeGen::CodeGenModule::getOpenMPRuntime(), clang::CodeGen::CodeGenTypeCache::Int32Ty, clang::CodeGen::CodeGenFunction::MakeAddrLValue(), and clang::CodeGen::CGOpenMPRuntime::OMPBuilder.

emitFunctionProlog()

void CGOpenMPRuntimeGPU::emitFunctionProlog ( CodeGenFunction &  CGF, const Decl *  D  )

overridevirtual

Emits OpenMP-specific function prolog.

Required for device constructs.

Reimplemented from clang::CodeGen::CGOpenMPRuntime.

Definition at line 3191 of file CGOpenMPRuntimeGPU.cpp.

References clang::CodeGen::CodeGenFunction::CapturedStmtInfo, clang::CodeGen::CodeGenFunction::CGM, clang::CR_OpenMP, clang::CodeGen::CodeGenFunction::CurFn, clang::Data, DS_Generic, clang::CodeGen::CodeGenFunction::EHStack, EM_SPMD, clang::Decl::getBeginLoc(), clang::CodeGen::CodeGenFunction::CGCapturedStmtInfo::getKind(), clang::CodeGen::CodeGenModule::getOpenMPRuntime(), clang::Decl::isCanonicalDecl(), and clang::CodeGen::NormalAndEHCleanup.

emitNumTeamsClause()

void CGOpenMPRuntimeGPU::emitNumTeamsClause ( CodeGenFunction &  CGF, const Expr *  NumTeams, const Expr *  ThreadLimit, SourceLocation  Loc  )

overridevirtual

This function ought to emit, in the general case, a call to.

Parameters

NumTeams	An integer expression of teams.
ThreadLimit	An integer expression of threads.

Reimplemented from clang::CodeGen::CGOpenMPRuntime.

Definition at line 933 of file CGOpenMPRuntimeGPU.cpp.

emitNumThreadsClause()

void CGOpenMPRuntimeGPU::emitNumThreadsClause ( CodeGenFunction &  CGF, llvm::Value *  NumThreads, SourceLocation  Loc  )

overridevirtual

Emits call to void __kmpc_push_num_threads(ident_t *loc, kmp_int32 global_tid, kmp_int32 num_threads) to generate code for 'num_threads' clause.

Parameters

NumThreads

An integer value of threads.

Reimplemented from clang::CodeGen::CGOpenMPRuntime.

Definition at line 927 of file CGOpenMPRuntimeGPU.cpp.

emitOutlinedFunctionCall()

void CGOpenMPRuntimeGPU::emitOutlinedFunctionCall ( CodeGenFunction &  CGF, SourceLocation  Loc, llvm::FunctionCallee  OutlinedFn, ArrayRef< llvm::Value * >  Args = std::nullopt  ) const

overridevirtual

Emits call of the outlined function with the provided arguments, translating these arguments to correct target-specific arguments.

Reimplemented from clang::CodeGen::CGOpenMPRuntime.

Definition at line 3037 of file CGOpenMPRuntimeGPU.cpp.

References clang::CodeGen::CodeGenFunction::Builder, clang::CodeGen::CGBuilderTy::CreatePointerBitCastOrAddrSpaceCast(), and clang::CodeGen::CGOpenMPRuntime::emitOutlinedFunctionCall().

Referenced by emitTeamsCall().

emitParallelCall()

void CGOpenMPRuntimeGPU::emitParallelCall ( CodeGenFunction &  CGF, SourceLocation  Loc, llvm::Function *  OutlinedFn, ArrayRef< llvm::Value * >  CapturedVars, const Expr *  IfCond, llvm::Value *  NumThreads  )

overridevirtual

Emits code for parallel or serial call of the OutlinedFn with variables captured in a record which address is stored in CapturedStruct.

Parameters

OutlinedFn	Outlined function to be run in parallel threads. Type of this function is void(*)(kmp_int32 *, kmp_int32, struct context_vars*).
CapturedVars	A pointer to the record with the references to variables used in OutlinedFn function.
IfCond	Condition in the associated 'if' clause, if it was specified, nullptr otherwise.
NumThreads	The value corresponding to the num_threads clause, if any, or nullptr.

Reimplemented from clang::CodeGen::CGOpenMPRuntime.

Definition at line 1230 of file CGOpenMPRuntimeGPU.cpp.

References clang::CodeGen::CodeGenFunction::Builder, clang::CodeGen::CGOpenMPRuntime::CGM, clang::CodeGen::CGBuilderTy::CreateConstArrayGEP(), clang::CodeGen::CodeGenFunction::CreateDefaultAlignTempAlloca(), clang::CodeGen::CGBuilderTy::CreatePointerBitCastOrAddrSpaceCast(), clang::CodeGen::Address::emitRawPointer(), clang::CodeGen::CodeGenFunction::EmitRuntimeCall(), clang::CodeGen::CodeGenFunction::EmitStoreOfScalar(), clang::CodeGen::CGOpenMPRuntime::emitUpdateLocation(), clang::CodeGen::CodeGenFunction::EvaluateExprAsBool(), clang::CodeGen::CodeGenFunction::getContext(), clang::CodeGen::CodeGenModule::getModule(), clang::ASTContext::getPointerType(), clang::CodeGen::CGOpenMPRuntime::getThreadID(), clang::CodeGen::CodeGenFunction::HaveInsertPoint(), clang::CodeGen::CodeGenTypeCache::Int32Ty, clang::CodeGen::CodeGenTypeCache::Int8PtrTy, clang::CodeGen::CGOpenMPRuntime::OMPBuilder, clang::CodeGen::CodeGenTypeCache::SizeTy, V, clang::CodeGen::CodeGenTypeCache::VoidPtrPtrTy, clang::ASTContext::VoidPtrTy, and clang::CodeGen::CodeGenTypeCache::VoidPtrTy.

emitParallelOutlinedFunction()

llvm::Function * CGOpenMPRuntimeGPU::emitParallelOutlinedFunction ( CodeGenFunction &  CGF, const OMPExecutableDirective &  D, const VarDecl *  ThreadIDVar, OpenMPDirectiveKind  InnermostKind, const RegionCodeGenTy &  CodeGen  )

overridevirtual

Emits inlined function for the specified OpenMP parallel.

D. This outlined function has type void(*)(kmp_int32 ThreadID, kmp_int32 BoundID, struct context_vars).

Parameters

CGF	Reference to current CodeGenFunction.
D	OpenMP directive.
ThreadIDVar	Variable for thread id in the current OpenMP region.
InnermostKind	Kind of innermost directive (for simple directives it is a directive itself, for combined - its innermost directive).
CodeGen	Code generation sequence for the D directive.

Reimplemented from clang::CodeGen::CGOpenMPRuntime.

Definition at line 938 of file CGOpenMPRuntimeGPU.cpp.

References EM_SPMD, and clang::CodeGen::CGOpenMPRuntime::emitParallelOutlinedFunction().

emitProcBindClause()

void CGOpenMPRuntimeGPU::emitProcBindClause ( CodeGenFunction &  CGF, llvm::omp::ProcBindKind  ProcBind, SourceLocation  Loc  )

overridevirtual

Emit call to void __kmpc_push_proc_bind(ident_t *loc, kmp_int32 global_tid, int proc_bind) to generate code for 'proc_bind' clause.

Reimplemented from clang::CodeGen::CGOpenMPRuntime.

Definition at line 921 of file CGOpenMPRuntimeGPU.cpp.

emitReduction()

void CGOpenMPRuntimeGPU::emitReduction ( CodeGenFunction &  CGF, SourceLocation  Loc, ArrayRef< const Expr * >  Privates, ArrayRef< const Expr * >  LHSExprs, ArrayRef< const Expr * >  RHSExprs, ArrayRef< const Expr * >  ReductionOps, ReductionOptionsTy  Options  )

overridevirtual

Emit a code for reduction clause.

Design of OpenMP reductions on the GPU.

Parameters

Privates	List of private copies for original reduction arguments.
LHSExprs	List of LHS in ReductionOps reduction operations.
RHSExprs	List of RHS in ReductionOps reduction operations.
ReductionOps	List of reduction operations in form 'LHS binop RHS' or 'operator binop(LHS, RHS)'.
Options	List of options for reduction codegen: WithNowait true if parent directive has also nowait clause, false otherwise. SimpleReduction Emit reduction operation only. Used for omp simd directive on the host. ReductionKind The kind of reduction to perform.

Consider a typical OpenMP program with one or more reduction clauses:

float foo; double bar; #pragma omp target teams distribute parallel for \ reduction(+:foo) reduction(*:bar) for (int i = 0; i < N; i++) { foo += A[i]; bar *= B[i]; }

where 'foo' and 'bar' are reduced across all OpenMP threads in all teams. In our OpenMP implementation on the NVPTX device an OpenMP team is mapped to a CUDA threadblock and OpenMP threads within a team are mapped to CUDA threads within a threadblock. Our goal is to efficiently aggregate values across all OpenMP threads such that:

• the compiler and runtime are logically concise, and
• the reduction is performed efficiently in a hierarchical manner as follows: within OpenMP threads in the same warp, across warps in a threadblock, and finally across teams on the NVPTX device.

Introduction to Decoupling

We would like to decouple the compiler and the runtime so that the latter is ignorant of the reduction variables (number, data types) and the reduction operators. This allows a simpler interface and implementation while still attaining good performance.

Pseudocode for the aforementioned OpenMP program generated by the compiler is as follows:

1. Create private copies of reduction variables on each OpenMP thread: 'foo_private', 'bar_private'
2. Each OpenMP thread reduces the chunk of 'A' and 'B' assigned to it and writes the result in 'foo_private' and 'bar_private' respectively.

3. Call the OpenMP runtime on the GPU to reduce within a team and store the result on the team master:

   __kmpc_nvptx_parallel_reduce_nowait_v2(..., reduceData, shuffleReduceFn, interWarpCpyFn)

   where: struct ReduceData { double *foo; double *bar; } reduceData reduceData.foo = &foo_private reduceData.bar = &bar_private

   'shuffleReduceFn' and 'interWarpCpyFn' are pointers to two auxiliary functions generated by the compiler that operate on variables of type 'ReduceData'. They aid the runtime perform algorithmic steps in a data agnostic manner.

   'shuffleReduceFn' is a pointer to a function that reduces data of type 'ReduceData' across two OpenMP threads (lanes) in the same warp. It takes the following arguments as input:

   a. variable of type 'ReduceData' on the calling lane, b. its lane_id, c. an offset relative to the current lane_id to generate a remote_lane_id. The remote lane contains the second variable of type 'ReduceData' that is to be reduced. d. an algorithm version parameter determining which reduction algorithm to use.

   'shuffleReduceFn' retrieves data from the remote lane using efficient GPU shuffle intrinsics and reduces, using the algorithm specified by the 4th parameter, the two operands element-wise. The result is written to the first operand.

   Different reduction algorithms are implemented in different runtime functions, all calling 'shuffleReduceFn' to perform the essential reduction step. Therefore, based on the 4th parameter, this function behaves slightly differently to cooperate with the runtime to ensure correctness under different circumstances.

   'InterWarpCpyFn' is a pointer to a function that transfers reduced variables across warps. It tunnels, through CUDA shared memory, the thread-private data of type 'ReduceData' from lane 0 of each warp to a lane in the first warp.

4. Call the OpenMP runtime on the GPU to reduce across teams. The last team writes the global reduced value to memory.

   ret = __kmpc_nvptx_teams_reduce_nowait(..., reduceData, shuffleReduceFn, interWarpCpyFn, scratchpadCopyFn, loadAndReduceFn)

   'scratchpadCopyFn' is a helper that stores reduced data from the team master to a scratchpad array in global memory.

   'loadAndReduceFn' is a helper that loads data from the scratchpad array and reduces it with the input operand.

   These compiler generated functions hide address calculation and alignment information from the runtime.

5. if ret == 1: The team master of the last team stores the reduced result to the globals in memory. foo += reduceData.foo; bar *= reduceData.bar

Warp Reduction Algorithms

On the warp level, we have three algorithms implemented in the OpenMP runtime depending on the number of active lanes:

Full Warp Reduction

The reduce algorithm within a warp where all lanes are active is implemented in the runtime as follows:

full_warp_reduce(void *reduce_data, kmp_ShuffleReductFctPtr ShuffleReduceFn) { for (int offset = WARPSIZE/2; offset > 0; offset /= 2) ShuffleReduceFn(reduce_data, 0, offset, 0); }

The algorithm completes in log(2, WARPSIZE) steps.

'ShuffleReduceFn' is used here with lane_id set to 0 because it is not used therefore we save instructions by not retrieving lane_id from the corresponding special registers. The 4th parameter, which represents the version of the algorithm being used, is set to 0 to signify full warp reduction.

In this version, 'ShuffleReduceFn' behaves, per element, as follows:

#reduce_elem refers to an element in the local lane's data structure #remote_elem is retrieved from a remote lane remote_elem = shuffle_down(reduce_elem, offset, WARPSIZE); reduce_elem = reduce_elem REDUCE_OP remote_elem;

Contiguous Partial Warp Reduction

This reduce algorithm is used within a warp where only the first 'n' (n <= WARPSIZE) lanes are active. It is typically used when the number of OpenMP threads in a parallel region is not a multiple of WARPSIZE. The algorithm is implemented in the runtime as follows:

void contiguous_partial_reduce(void *reduce_data, kmp_ShuffleReductFctPtr ShuffleReduceFn, int size, int lane_id) { int curr_size; int offset; curr_size = size; mask = curr_size/2; while (offset>0) { ShuffleReduceFn(reduce_data, lane_id, offset, 1); curr_size = (curr_size+1)/2; offset = curr_size/2; } }

In this version, 'ShuffleReduceFn' behaves, per element, as follows:

remote_elem = shuffle_down(reduce_elem, offset, WARPSIZE); if (lane_id < offset) reduce_elem = reduce_elem REDUCE_OP remote_elem else reduce_elem = remote_elem

This algorithm assumes that the data to be reduced are located in a contiguous subset of lanes starting from the first. When there is an odd number of active lanes, the data in the last lane is not aggregated with any other lane's dat but is instead copied over.

Dispersed Partial Warp Reduction

This algorithm is used within a warp when any discontiguous subset of lanes are active. It is used to implement the reduction operation across lanes in an OpenMP simd region or in a nested parallel region.

void dispersed_partial_reduce(void *reduce_data, kmp_ShuffleReductFctPtr ShuffleReduceFn) { int size, remote_id; int logical_lane_id = number_of_active_lanes_before_me() * 2; do { remote_id = next_active_lane_id_right_after_me();

the above function returns 0 of no active lane

is present right after the current lane.

size = number_of_active_lanes_in_this_warp(); logical_lane_id /= 2; ShuffleReduceFn(reduce_data, logical_lane_id, remote_id-1-threadIdx.x, 2); } while (logical_lane_id % 2 == 0 && size > 1); }

There is no assumption made about the initial state of the reduction. Any number of lanes (>=1) could be active at any position. The reduction result is returned in the first active lane.

In this version, 'ShuffleReduceFn' behaves, per element, as follows:

remote_elem = shuffle_down(reduce_elem, offset, WARPSIZE); if (lane_id % 2 == 0 && offset > 0) reduce_elem = reduce_elem REDUCE_OP remote_elem else reduce_elem = remote_elem

Intra-Team Reduction

This function, as implemented in the runtime call '__kmpc_nvptx_parallel_reduce_nowait_v2', aggregates data across OpenMP threads in a team. It first reduces within a warp using the aforementioned algorithms. We then proceed to gather all such reduced values at the first warp.

The runtime makes use of the function 'InterWarpCpyFn', which copies data from each of the "warp master" (zeroth lane of each warp, where warp-reduced data is held) to the zeroth warp. This step reduces (in a mathematical sense) the problem of reduction across warp masters in a block to the problem of warp reduction.

Inter-Team Reduction

Once a team has reduced its data to a single value, it is stored in a global scratchpad array. Since each team has a distinct slot, this can be done without locking.

The last team to write to the scratchpad array proceeds to reduce the scratchpad array. One or more workers in the last team use the helper 'loadAndReduceDataFn' to load and reduce values from the array, i.e., the k'th worker reduces every k'th element.

Finally, a call is made to '__kmpc_nvptx_parallel_reduce_nowait_v2' to reduce across workers and compute a globally reduced value.

Reimplemented from clang::CodeGen::CGOpenMPRuntime.

Definition at line 2807 of file CGOpenMPRuntimeGPU.cpp.

References clang::CodeGen::CodeGenFunction::Builder, clang::C, clang::CodeGen::CGOpenMPRuntime::CGM, clang::CodeGen::CodeGenFunction::ConvertTypeForMem(), clang::CodeGen::CodeGenFunction::createBasicBlock(), clang::CodeGen::CGBuilderTy::CreateConstArrayGEP(), clang::CodeGen::ApplyDebugLocation::CreateEmpty(), clang::CodeGen::CodeGenFunction::CreateMemTemp(), clang::CodeGen::CGBuilderTy::CreatePointerBitCastOrAddrSpaceCast(), clang::CodeGen::CGBuilderTy::CreateStore(), clang::CodeGen::CodeGenFunction::CurFn, clang::CodeGen::CodeGenFunction::EmitBlock(), emitGlobalToListCopyFunction(), emitGlobalToListReduceFunction(), emitInterWarpCopyFunction(), emitListToGlobalCopyFunction(), emitListToGlobalReduceFunction(), clang::CodeGen::CodeGenFunction::EmitLValue(), clang::CodeGen::Address::emitRawPointer(), clang::CodeGen::CGOpenMPRuntime::emitReduction(), clang::CodeGen::CGOpenMPRuntime::emitReductionFunction(), clang::CodeGen::CodeGenFunction::EmitRuntimeCall(), emitShuffleAndReduceFunction(), clang::CodeGen::CGOpenMPRuntime::emitSingleReductionCombiner(), clang::CodeGen::CGOpenMPRuntime::emitUpdateLocation(), clang::ASTContext::getAsVariableArrayType(), clang::CodeGen::CodeGenFunction::getContext(), clang::CodeGen::CodeGenModule::getContext(), clang::CodeGen::CodeGenModule::getModule(), clang::CodeGen::LValue::getPointer(), clang::Expr::getType(), clang::CodeGen::CodeGenFunction::getTypeSize(), clang::CodeGen::CodeGenFunction::getVLASize(), clang::CodeGen::CodeGenFunction::HaveInsertPoint(), clang::CodeGen::CodeGenTypeCache::Int32Ty, clang::CodeGen::CodeGenTypeCache::Int64Ty, clang::isOpenMPParallelDirective(), clang::isOpenMPTeamsDirective(), clang::Type::isVariablyModifiedType(), clang::Normal, clang::CodeGen::CodeGenFunction::VlaSizePair::NumElts, clang::CodeGen::CGOpenMPRuntime::OMPBuilder, clang::CodeGen::CodeGenTypeCache::SizeTy, and clang::CodeGen::CodeGenTypeCache::VoidPtrTy.

emitTeamsCall()

void CGOpenMPRuntimeGPU::emitTeamsCall ( CodeGenFunction &  CGF, const OMPExecutableDirective &  D, SourceLocation  Loc, llvm::Function *  OutlinedFn, ArrayRef< llvm::Value * >  CapturedVars  )

overridevirtual

Emits code for teams call of the OutlinedFn with variables captured in a record which address is stored in CapturedStruct.

Parameters

OutlinedFn	Outlined function to be run by team masters. Type of this function is void(*)(kmp_int32 *, kmp_int32, struct context_vars*).
CapturedVars	A pointer to the record with the references to variables used in OutlinedFn function.

Reimplemented from clang::CodeGen::CGOpenMPRuntime.

Definition at line 1205 of file CGOpenMPRuntimeGPU.cpp.

References clang::CodeGen::CodeGenFunction::Builder, clang::CodeGen::CGOpenMPRuntime::CGM, clang::CodeGen::CodeGenFunction::CreateDefaultAlignTempAlloca(), clang::CodeGen::CGBuilderTy::CreateStore(), emitOutlinedFunctionCall(), clang::CodeGen::Address::emitRawPointer(), clang::CodeGen::CGOpenMPRuntime::emitThreadIDAddress(), clang::OMPExecutableDirective::getSingleClause(), clang::CodeGen::CodeGenFunction::HaveInsertPoint(), clang::CodeGen::CodeGenTypeCache::Int32Ty, and clang::CodeGen::CodeGenTypeCache::VoidPtrTy.

emitTeamsOutlinedFunction()

llvm::Function * CGOpenMPRuntimeGPU::emitTeamsOutlinedFunction ( CodeGenFunction &  CGF, const OMPExecutableDirective &  D, const VarDecl *  ThreadIDVar, OpenMPDirectiveKind  InnermostKind, const RegionCodeGenTy &  CodeGen  )

overridevirtual

Emits inlined function for the specified OpenMP teams.

D. This outlined function has type void(*)(kmp_int32 ThreadID, kmp_int32 BoundID, struct context_vars).

Parameters

CGF	Reference to current CodeGenFunction.
D	OpenMP directive.
ThreadIDVar	Variable for thread id in the current OpenMP region.
InnermostKind	Kind of innermost directive (for simple directives it is a directive itself, for combined - its innermost directive).
CodeGen	Code generation sequence for the D directive.

Reimplemented from clang::CodeGen::CGOpenMPRuntime.

Definition at line 996 of file CGOpenMPRuntimeGPU.cpp.

References clang::CodeGen::CGOpenMPRuntime::CGM, clang::CodeGen::CodeGenFunction::CGM, clang::CodeGen::CodeGenFunction::CurFn, Data, EM_SPMD, clang::CodeGen::CGOpenMPRuntime::emitTeamsOutlinedFunction(), clang::OMPExecutableDirective::getBeginLoc(), clang::CapturedStmt::getCapturedDecl(), clang::OMPExecutableDirective::getCapturedStmt(), clang::CodeGen::CodeGenModule::getContext(), getDistributeLastprivateVars(), clang::TargetInfo::getGridValue(), clang::CodeGen::CodeGenModule::getOpenMPRuntime(), clang::CodeGen::CodeGenModule::getTarget(), getTeamsReductionVars(), and clang::CodeGen::RegionCodeGenTy::setAction().

functionFinished()

void CGOpenMPRuntimeGPU::functionFinished ( CodeGenFunction &  CGF )

overridevirtual

Cleans up references to the objects in finished function.

Reimplemented from clang::CodeGen::CGOpenMPRuntime.

Definition at line 3327 of file CGOpenMPRuntimeGPU.cpp.

References clang::CodeGen::CodeGenFunction::CurFn, and clang::CodeGen::CGOpenMPRuntime::functionFinished().

getAddressOfLocalVariable()

Address CGOpenMPRuntimeGPU::getAddressOfLocalVariable ( CodeGenFunction &  CGF, const VarDecl *  VD  )

overridevirtual

Gets the OpenMP-specific address of the local variable.

Reimplemented from clang::CodeGen::CGOpenMPRuntime.

Definition at line 3258 of file CGOpenMPRuntimeGPU.cpp.

References clang::Decl::attr_begin(), clang::Decl::attr_end(), clang::CodeGen::CodeGenFunction::Builder, clang::CodeGen::CGOpenMPRuntime::CGM, clang::CodeGen::CodeGenFunction::ConvertTypeForMem(), clang::CodeGen::CGBuilderTy::CreatePointerBitCastOrAddrSpaceCast(), clang::cuda_constant, clang::cuda_shared, clang::CodeGen::CodeGenFunction::CurFn, clang::Default, DS_Generic, clang::QualType::getAddressSpace(), clang::CharUnits::getAsAlign(), clang::Decl::getAttr(), clang::VarDecl::getCanonicalDecl(), clang::CodeGen::CodeGenModule::getContext(), clang::ASTContext::getDeclAlign(), clang::CodeGen::CodeGenModule::getModule(), clang::NamedDecl::getName(), clang::ASTContext::getTargetAddressSpace(), clang::ValueDecl::getType(), clang::Decl::hasAttr(), clang::Decl::hasAttrs(), and clang::CodeGen::Address::invalid().

getDefaultDistScheduleAndChunk()

void CGOpenMPRuntimeGPU::getDefaultDistScheduleAndChunk ( CodeGenFunction &  CGF, const OMPLoopDirective &  S, OpenMPDistScheduleClauseKind &  ScheduleKind, llvm::Value *&  Chunk  ) const

overridevirtual

Choose a default value for the dist_schedule clause.

Reimplemented from clang::CodeGen::CGOpenMPRuntime.

Definition at line 3332 of file CGOpenMPRuntimeGPU.cpp.

References clang::CodeGen::CodeGenFunction::CGM, EM_SPMD, clang::CodeGen::CodeGenFunction::EmitScalarConversion(), clang::CodeGen::CodeGenFunction::getContext(), clang::CodeGen::CGOpenMPRuntime::getDefaultDistScheduleAndChunk(), clang::ASTContext::getIntTypeForBitwidth(), and clang::CodeGen::CodeGenModule::getOpenMPRuntime().

getDefaultScheduleAndChunk()

void CGOpenMPRuntimeGPU::getDefaultScheduleAndChunk ( CodeGenFunction &  CGF, const OMPLoopDirective &  S, OpenMPScheduleClauseKind &  ScheduleKind, const Expr *&  ChunkExpr  ) const

overridevirtual

Choose a default value for the schedule clause.

Reimplemented from clang::CodeGen::CGOpenMPRuntime.

Definition at line 3349 of file CGOpenMPRuntimeGPU.cpp.

References clang::IntegerLiteral::Create(), clang::CodeGen::CodeGenFunction::getContext(), and clang::ASTContext::getIntTypeForBitwidth().

getGPUNumThreads()

llvm::Value * CGOpenMPRuntimeGPU::getGPUNumThreads

(

CodeGenFunction &

CGF

)

Get the maximum number of threads in a block of the GPU.

Definition at line 3550 of file CGOpenMPRuntimeGPU.cpp.

References clang::CodeGen::CodeGenFunction::Builder, clang::CodeGen::CodeGenFunction::CGM, clang::CodeGen::CodeGenModule::getModule(), and clang::CodeGen::CodeGenTypeCache::Int32Ty.

getGPUThreadID()

llvm::Value * CGOpenMPRuntimeGPU::getGPUThreadID

(

CodeGenFunction &

CGF

)

Get the id of the current thread on the GPU.

Definition at line 3563 of file CGOpenMPRuntimeGPU.cpp.

References clang::CodeGen::CGOpenMPRuntime::CGM, clang::CodeGen::CodeGenFunction::EmitRuntimeCall(), clang::CodeGen::CodeGenModule::getModule(), and clang::CodeGen::CGOpenMPRuntime::OMPBuilder.

Referenced by emitInterWarpCopyFunction().

getGPUWarpSize()

llvm::Value * CGOpenMPRuntimeGPU::getGPUWarpSize

(

CodeGenFunction &

CGF

)

Get the GPU warp size.

Definition at line 3571 of file CGOpenMPRuntimeGPU.cpp.

References clang::CodeGen::CGOpenMPRuntime::CGM, clang::CodeGen::CodeGenFunction::EmitRuntimeCall(), clang::CodeGen::CodeGenModule::getModule(), and clang::CodeGen::CGOpenMPRuntime::OMPBuilder.

Referenced by createRuntimeShuffleFunction().

getKmpcAllocShared()

std::pair< llvm::Value *, llvm::Value * > CGOpenMPRuntimeGPU::getKmpcAllocShared ( CodeGenFunction &  CGF, const VarDecl *  VD  )

overridevirtual

Get call to __kmpc_alloc_shared.

Reimplemented from clang::CodeGen::CGOpenMPRuntime.

Definition at line 1142 of file CGOpenMPRuntimeGPU.cpp.

References clang::CodeGen::CodeGenFunction::Builder, clang::CodeGen::CGOpenMPRuntime::CGM, clang::CodeGen::CodeGenFunction::EmitRuntimeCall(), clang::CodeGen::CodeGenModule::getContext(), clang::ASTContext::getDeclAlign(), clang::CodeGen::CodeGenModule::getLLVMContext(), clang::CodeGen::CodeGenModule::getModule(), clang::NamedDecl::getName(), clang::CharUnits::getQuantity(), clang::ValueDecl::getType(), clang::CodeGen::CodeGenFunction::getTypeSize(), clang::CodeGen::CGOpenMPRuntime::OMPBuilder, and clang::CodeGen::CodeGenTypeCache::SizeTy.

getKmpcFreeShared()

void CGOpenMPRuntimeGPU::getKmpcFreeShared ( CodeGenFunction &  CGF, const std::pair< llvm::Value *, llvm::Value * > &  AddrSizePair  )

overridevirtual

Get call to __kmpc_free_shared.

Reimplemented from clang::CodeGen::CGOpenMPRuntime.

Definition at line 1168 of file CGOpenMPRuntimeGPU.cpp.

References clang::CodeGen::CGOpenMPRuntime::CGM, clang::CodeGen::CodeGenFunction::EmitRuntimeCall(), clang::CodeGen::CodeGenModule::getModule(), and clang::CodeGen::CGOpenMPRuntime::OMPBuilder.

getParameterAddress()

Address CGOpenMPRuntimeGPU::getParameterAddress ( CodeGenFunction &  CGF, const VarDecl *  NativeParam, const VarDecl *  TargetParam  ) const

overridevirtual

Gets the address of the native argument basing on the address of the target-specific parameter.

Parameters

NativeParam	Parameter itself.
TargetParam	Corresponding target-specific parameter.

Reimplemented from clang::CodeGen::CGOpenMPRuntime.

Definition at line 3011 of file CGOpenMPRuntimeGPU.cpp.

References clang::CodeGen::CodeGenFunction::Builder, clang::CodeGen::CodeGenFunction::CreateMemTemp(), clang::CodeGen::CGBuilderTy::CreatePointerBitCastOrAddrSpaceCast(), clang::CodeGen::CodeGenFunction::EmitLoadOfScalar(), clang::CodeGen::CodeGenFunction::EmitStoreOfScalar(), clang::CodeGen::CodeGenFunction::GetAddrOfLocalVar(), clang::CodeGen::CodeGenFunction::getLLVMContext(), clang::Type::getPointeeType(), clang::CodeGen::CodeGenTypes::getTargetAddressSpace(), clang::ValueDecl::getType(), clang::CodeGen::CodeGenFunction::getTypes(), clang::Type::isReferenceType(), and clang::QualifierCollector::strip().

hasAllocateAttributeForGlobalVar()

bool CGOpenMPRuntimeGPU::hasAllocateAttributeForGlobalVar ( const VarDecl *  VD, LangAS &  AS  )

overridevirtual

Checks if the variable has associated OMPAllocateDeclAttr attribute with the predefined allocator and translates it into the corresponding address space.

Reimplemented from clang::CodeGen::CGOpenMPRuntime.

Definition at line 3415 of file CGOpenMPRuntimeGPU.cpp.

References clang::cuda_constant, clang::cuda_shared, clang::Default, clang::Decl::getAttr(), and clang::Decl::hasAttr().

isDefaultLocationConstant()

bool clang::CodeGen::CGOpenMPRuntimeGPU::isDefaultLocationConstant ( ) const

inlineoverrideprotectedvirtual

Check if the default location must be constant.

Constant for NVPTX for better optimization.

Reimplemented from clang::CodeGen::CGOpenMPRuntime.

Definition at line 129 of file CGOpenMPRuntimeGPU.h.

isDelayedVariableLengthDecl()

bool CGOpenMPRuntimeGPU::isDelayedVariableLengthDecl ( CodeGenFunction &  CGF, const VarDecl *  VD  ) const

overridevirtual

Declare generalized virtual functions which need to be defined by all specializations of OpenMPGPURuntime Targets like AMDGCN and NVPTX.

Check if the variable length declaration is delayed:

Reimplemented from clang::CodeGen::CGOpenMPRuntime.

Definition at line 1131 of file CGOpenMPRuntimeGPU.cpp.

References clang::CodeGen::CodeGenFunction::CurFn.

isGPU()

bool clang::CodeGen::CGOpenMPRuntimeGPU::isGPU ( ) const

inlineoverridevirtual

Returns true if the current target is a GPU.

Reimplemented from clang::CodeGen::CGOpenMPRuntime.

Definition at line 134 of file CGOpenMPRuntimeGPU.h.

Referenced by CGOpenMPRuntimeGPU().

processRequiresDirective()

void CGOpenMPRuntimeGPU::processRequiresDirective ( const OMPRequiresDecl *  D )

overridevirtual

Perform check on requires decl to ensure that target architecture supports unified addressing.

Check to see if target architecture supports unified addressing which is a restriction for OpenMP requires clause "unified_shared_memory".

Reimplemented from clang::CodeGen::CGOpenMPRuntime.

Definition at line 3460 of file CGOpenMPRuntimeGPU.cpp.

References clang::CodeGen::CGOpenMPRuntime::CGM, clang::OMPRequiresDecl::clauselists(), clang::CudaArchToString(), clang::CodeGen::CodeGenModule::Error(), clang::Generic, getCudaArch(), clang::GFX1010, clang::GFX1011, clang::GFX1012, clang::GFX1013, clang::GFX1030, clang::GFX1031, clang::GFX1032, clang::GFX1033, clang::GFX1034, clang::GFX1035, clang::GFX1036, clang::GFX1100, clang::GFX1101, clang::GFX1102, clang::GFX1103, clang::GFX1150, clang::GFX1151, clang::GFX1200, clang::GFX1201, clang::GFX600, clang::GFX601, clang::GFX602, clang::GFX700, clang::GFX701, clang::GFX702, clang::GFX703, clang::GFX704, clang::GFX705, clang::GFX801, clang::GFX802, clang::GFX803, clang::GFX805, clang::GFX810, clang::GFX900, clang::GFX902, clang::GFX904, clang::GFX906, clang::GFX908, clang::GFX909, clang::GFX90a, clang::GFX90c, clang::GFX940, clang::GFX941, clang::GFX942, clang::LAST, clang::CodeGen::CGOpenMPRuntime::processRequiresDirective(), clang::SM_20, clang::SM_21, clang::SM_30, clang::SM_32_, clang::SM_35, clang::SM_37, clang::SM_50, clang::SM_52, clang::SM_53, clang::SM_60, clang::SM_61, clang::SM_62, clang::SM_70, clang::SM_72, clang::SM_75, clang::SM_80, clang::SM_86, clang::SM_87, clang::SM_89, clang::SM_90, clang::SM_90a, clang::UNKNOWN, and clang::UNUSED.

translateParameter()

const VarDecl * CGOpenMPRuntimeGPU::translateParameter ( const FieldDecl *  FD, const VarDecl *  NativeParam  ) const

overridevirtual

Translates the native parameter of outlined function if this is required for target.

Parameters

FD	Field decl from captured record for the parameter.
NativeParam	Parameter itself.

Reimplemented from clang::CodeGen::CGOpenMPRuntime.

Definition at line 2979 of file CGOpenMPRuntimeGPU.cpp.

References clang::Qualifiers::addAddressSpace(), clang::Qualifiers::addRestrict(), clang::QualifierCollector::apply(), clang::CodeGen::CGOpenMPRuntime::CGM, clang::ImplicitParamDecl::Create(), clang::ParmVarDecl::Create(), clang::ASTContext::getAddrSpaceQualType(), clang::Decl::getAttr(), clang::DeclaratorDecl::getBeginLoc(), clang::CodeGen::CodeGenModule::getContext(), clang::Decl::getDeclContext(), clang::NamedDecl::getIdentifier(), clang::getLangASFromTargetAS(), clang::Decl::getLocation(), clang::Type::getPointeeType(), clang::ASTContext::getPointerType(), clang::ValueDecl::getType(), clang::Type::isReferenceType(), clang::opencl_global, clang::Other, clang::SC_None, and clang::QualifierCollector::strip().

---

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

• lib/CodeGen/CGOpenMPRuntimeGPU.h
• lib/CodeGen/CGOpenMPRuntimeGPU.cpp