Namespaces
	deallocation_impl

	detail

	func_ext

Classes
struct	ValueComparator
	Compare two SSA values in a deterministic manner. More...

class	Ownership
	This class is used to track the ownership of values. More...

struct	DeallocationOptions
	Options for BufferDeallocationOpInterface-based buffer deallocation. More...

class	DeallocationState
	This class collects all the state that we need to perform the buffer deallocation pass with associated helper functions such that we have easy access to it in the BufferDeallocationOpInterface implementations and the BufferDeallocation pass. More...

struct	AliasingOpOperand
	A maybe aliasing OpOperand. More...

struct	AliasingValue
	A maybe aliasing Value. More...

class	AliasList

class	OpFilter

struct	BufferizationOptions
	Options for BufferizableOpInterface-based bufferization. More...

struct	TraversalConfig
	Traversal parameters for `findValueInReverseUseDefChain`. More...

class	AnalysisState
	AnalysisState provides a variety of helper functions for dealing with tensor values. More...

struct	DstBufferizableOpInterfaceExternalModel
	Bufferizable ops that implement the DestinationStyleOpInterface can use this external model base class. More...

struct	OpWithUnstructuredControlFlowBufferizableOpInterfaceExternalModel
	A template that provides a default implementation of `getAliasingOpOperands` for ops that support unstructured control flow within their regions. More...

struct	BranchOpBufferizableOpInterfaceExternalModel
	A template that provides a default implementation of `getAliasingValues` for ops that implement the `BranchOpInterface`. More...

struct	BufferDeallocationPipelineOptions
	Options for the buffer deallocation pipeline. More...

struct	BufferizationStatistics
	Bufferization statistics for debugging. More...

class	BufferPlacementAllocs
	A simple analysis that detects allocation operations. More...

class	BufferPlacementTransformationBase
	The base class for all BufferPlacement transformations. More...

struct	OneShotBufferizationOptions
	Options for analysis-enabled bufferization. More...

class	OneShotAnalysisState
	State for analysis-enabled bufferization. More...

struct	BufferResultsToOutParamsOpts

Typedefs
using	AliasingOpOperandList = AliasList< AliasingOpOperand >
	A list of possible aliasing OpOperands. More...

using	AliasingValueList = AliasList< AliasingValue >
	A list of possible aliasing Values. More...

using	RegisterDependenciesFn = std::function< void(ValueRange, ValueRange)>

using	DeallocHelperMap = llvm::DenseMap< Operation *, func::FuncOp >
	Maps from symbol table to its corresponding dealloc helper function. More...

using	ControlBuildSubsetExtractionFn = std::function< Value(RewriterBase &, SubsetInsertionOpInterface, tensor::EmptyOp emptyTensorOp, Operation *user)>
	A function type that defines a callback to control the construction of the subset extraction of the `SubsetInsertionOpInterface`. More...

Enumerations
enum class	BufferRelation { Unknown , Equivalent }
	Specifies a fine-grain relationship between buffers to enable more analysis. More...

Functions
FailureOr< Value >	allocateTensorForShapedValue (OpBuilder &b, Location loc, Value shapedValue, const BufferizationOptions &options, bool copy=true)
	Create an AllocTensorOp for the given shaped value (memref or tensor). More...

FailureOr< Value >	getBuffer (RewriterBase &rewriter, Value value, const BufferizationOptions &options)
	Lookup the buffer for the given value. More...

FailureOr< BaseMemRefType >	getBufferType (Value value, const BufferizationOptions &options)
	Return the buffer type for a given Value (tensor) after bufferization without bufferizing any IR. More...

FailureOr< BaseMemRefType >	getBufferType (Value value, const BufferizationOptions &options, SmallVector< Value > &invocationStack)
	Return the buffer type for a given Value (tensor) after bufferization without bufferizing any IR. More...

bool	hasTensorSemantics (Operation *op)
	Return "true" if the given op has tensor semantics and should be bufferized. More...

void	replaceOpWithBufferizedValues (RewriterBase &rewriter, Operation *op, ValueRange values)
	Replace an op with replacement values. More...

template<typename OpTy , typename... Args>
OpTy	replaceOpWithNewBufferizedOp (RewriterBase &rewriter, Operation *op, Args &&...args)
	Replace an op with a new op. More...

BaseMemRefType	getMemRefType (Value value, const BufferizationOptions &options, MemRefLayoutAttrInterface layout={}, Attribute memorySpace=nullptr)
	Return a MemRefType to which the type of the given value can be bufferized. More...

BaseMemRefType	getMemRefTypeWithFullyDynamicLayout (TensorType tensorType, Attribute memorySpace=nullptr)
	Return a MemRef type with fully dynamic layout. More...

BaseMemRefType	getMemRefTypeWithStaticIdentityLayout (TensorType tensorType, Attribute memorySpace=nullptr)
	Return a MemRef type with a static identity layout (i.e., no layout map). More...

Operation *	getOwnerOfValue (Value value)
	Return the owner of the given value. More...

Region *	getNextEnclosingRepetitiveRegion (Region *region, const BufferizationOptions &options)
	Assuming that the given region is repetitive, find the next enclosing repetitive region. More...

Region *	getParallelRegion (Region *region, const BufferizationOptions &options)
	If `region` is a parallel region, return `region`. More...

void	populateDynamicDimSizes (OpBuilder &b, Location loc, Value shapedValue, SmallVector< Value > &dynamicDims)
	Populate `dynamicDims` with tensor::DimOp / memref::DimOp results for all dynamic dimensions of the given shaped value. More...

FailureOr< Value >	castOrReallocMemRefValue (OpBuilder &b, Value value, MemRefType type, const BufferizationOptions &options)
	Try to cast the given ranked MemRef-typed value to the given ranked MemRef type. More...

LogicalResult	foldToBufferToTensorPair (RewriterBase &rewriter, ToBufferOp toBuffer, const BufferizationOptions &options)
	Try to fold to_buffer(to_tensor(x)). More...

void	populateDeallocOpCanonicalizationPatterns (RewritePatternSet &patterns, MLIRContext *context)
	Add the canonicalization patterns for bufferization.dealloc to the given pattern set to make them available to other passes (such as BufferDeallocationSimplification). More...

void	buildBufferDeallocationPipeline (OpPassManager &pm, const BufferDeallocationPipelineOptions &options)
	Adds the buffer deallocation pipeline to the `OpPassManager`. More...

void	registerBufferizationPipelines ()
	Registers all pipelines for the `bufferization` dialect. More...

void	registerTransformDialectExtension (DialectRegistry &registry)

LogicalResult	bufferizeOp (Operation op, const BufferizationOptions &options, BufferizationStatistics statistics=nullptr)
	Bufferize `op` and its nested ops that implement `BufferizableOpInterface`. More...

LogicalResult	bufferizeBlockSignature (Block *block, RewriterBase &rewriter, const BufferizationOptions &options)
	Bufferize the signature of `block` and its callers (i.e., ops that have the given block as a successor). More...

template<typename DominatorT >
Block *	findCommonDominator (Value value, const BufferViewFlowAnalysis::ValueSetT &values, const DominatorT &doms)
	Finds a common dominator for the given value while taking the positions of the values in the value set into account. More...

FailureOr< memref::GlobalOp >	getGlobalFor (arith::ConstantOp constantOp, uint64_t alignment, Attribute memorySpace={})

SmallVector< func::ReturnOp >	getReturnOps (func::FuncOp funcOp)
	Helper function that returns all func.return ops in the given function. More...

LogicalResult	analyzeOp (Operation op, OneShotAnalysisState &state, BufferizationStatistics statistics=nullptr)
	Analyze `op` and its nested ops. More...

LogicalResult	runOneShotBufferize (Operation op, const OneShotBufferizationOptions &options, BufferizationStatistics statistics=nullptr)
	Run One-Shot Bufferize on the given op: Analysis + Bufferization. More...

llvm::LogicalResult	analyzeModuleOp (ModuleOp moduleOp, OneShotAnalysisState &state, BufferizationStatistics *statistics=nullptr)
	Analyze `moduleOp` and its nested ops. More...

llvm::LogicalResult	bufferizeModuleOp (ModuleOp moduleOp, const OneShotBufferizationOptions &options, BufferizationStatistics *statistics=nullptr)
	Bufferize `op` and its nested ops that implement `BufferizableOpInterface`. More...

void	removeBufferizationAttributesInModule (ModuleOp moduleOp)
	Remove bufferization attributes on every FuncOp arguments in the ModuleOp. More...

llvm::LogicalResult	runOneShotModuleBufferize (ModuleOp moduleOp, const bufferization::OneShotBufferizationOptions &options, BufferizationStatistics *statistics=nullptr)
	Run One-Shot Module Bufferization on the given module. More...

void	populateBufferizationDeallocLoweringPattern (RewritePatternSet &patterns, const DeallocHelperMap &deallocHelperFuncMap)
	Adds the conversion pattern of the `bufferization.dealloc` operation to the given pattern set for use in other transformation passes. More...

func::FuncOp	buildDeallocationLibraryFunction (OpBuilder &builder, Location loc, SymbolTable &symbolTable)
	Construct the library function needed for the fully generic `bufferization.dealloc` lowering implemented in the LowerDeallocations pass. More...

LogicalResult	deallocateBuffersOwnershipBased (FunctionOpInterface op, DeallocationOptions options)
	Run the ownership-based buffer deallocation. More...

LogicalResult	promoteBufferResultsToOutParams (ModuleOp module, const BufferResultsToOutParamsOpts &options)
	Replace buffers that are returned from a function with an out parameter. More...

LogicalResult	dropEquivalentBufferResults (ModuleOp module)
	Drop all memref function results that are equivalent to a function argument. More...

std::unique_ptr< Pass >	createPromoteBuffersToStackPass (std::function< bool(Value)> isSmallAlloc)
	Creates a pass that promotes heap-based allocations to stack-based ones. More...

LogicalResult	eliminateEmptyTensors (RewriterBase &rewriter, Operation *op)
	Try to eliminate "tensor.empty" ops inside `op`. More...

Value	buildSubsetExtraction (RewriterBase &rewriter, SubsetInsertionOpInterface op, tensor::EmptyOp emptyTensorOp, Operation *user)
	This method builds and returns a subset extraction value for the destination tensor that the given `op` inserts into. More...

LogicalResult	eliminateEmptyTensors (RewriterBase &rewriter, Operation *op, OneShotAnalysisState &state, ControlBuildSubsetExtractionFn subsetsExtractionFn=buildSubsetExtraction)
	Try to eliminate "tensor.empty" ops inside `op`. More...

void	hoistBuffersFromLoops (Operation *op)
	Within the given operation, hoist buffers from loops where possible. More...

LogicalResult	insertTensorCopies (Operation op, const OneShotBufferizationOptions &options, BufferizationStatistics statistics=nullptr)
	Resolve RaW and other conflicts by inserting bufferization.alloc_tensor ops. More...

LogicalResult	insertTensorCopies (Operation *op, const AnalysisState &state)
	Resolve RaW and other conflicts by inserting bufferization.alloc_tensor ops. More...

void	populateEmptyTensorToAllocTensorPattern (RewritePatternSet &patterns)
	Populate patterns to lower tensor.empty ops to bufferization.alloc_tensor ops. More...

Typedef Documentation

◆ AliasingOpOperandList

using mlir::bufferization::AliasingOpOperandList = typedef AliasList<AliasingOpOperand>

A list of possible aliasing OpOperands.

This list models the runtime aliasing relationship for a Value.

Definition at line 94 of file BufferizableOpInterface.h.

◆ AliasingValueList

using mlir::bufferization::AliasingValueList = typedef AliasList<AliasingValue>

A list of possible aliasing Values.

This list models the runtime aliasing relationship for an OpOperand.

Definition at line 98 of file BufferizableOpInterface.h.

◆ ControlBuildSubsetExtractionFn

using mlir::bufferization::ControlBuildSubsetExtractionFn = typedef std::function<Value(RewriterBase &, SubsetInsertionOpInterface, tensor::EmptyOp emptyTensorOp, Operation *user)>

A function type that defines a callback to control the construction of the subset extraction of the SubsetInsertionOpInterface.

The subset extraction value can be used as a replacement for the emptyTensorOp value which is being consumed by user, failing of building such a value should be indicated with an empty value. This function should guarantee the legality of the replacement, i.e. the replacement should dominate the user of the emptyTensorOp being eliminated.

Definition at line 47 of file Transforms.h.

◆ DeallocHelperMap

using mlir::bufferization::DeallocHelperMap = typedef llvm::DenseMap<Operation *, func::FuncOp>

Maps from symbol table to its corresponding dealloc helper function.

Definition at line 25 of file Passes.h.

◆ RegisterDependenciesFn

using mlir::bufferization::RegisterDependenciesFn = typedef std::function<void(ValueRange, ValueRange)>

Definition at line 20 of file BufferViewFlowOpInterface.h.

Enumeration Type Documentation

◆ BufferRelation

enum mlir::bufferization::BufferRelation

strong

Specifies a fine-grain relationship between buffers to enable more analysis.

Enumerator
Unknown
Equivalent

Definition at line 33 of file BufferizableOpInterface.h.

Function Documentation

◆ allocateTensorForShapedValue()

FailureOr< Value > mlir::bufferization::allocateTensorForShapedValue	(	OpBuilder &	b,
		Location	loc,
		Value	shapedValue,
		const BufferizationOptions &	options,
		bool	copy = `true`
	)

Create an AllocTensorOp for the given shaped value (memref or tensor).

Create an AllocTensorOp for the given shaped value.

If copy is set, the shaped value is copied. Otherwise, a tensor with undefined contents is allocated.

Definition at line 162 of file BufferizableOpInterface.cpp.

References copy(), mlir::OpBuilder::create(), mlir::Operation::emitError(), mlir::detail::enumerate(), getBufferType(), mlir::Value::getDefiningOp(), getOwnerOfValue(), mlir::Operation::getResult(), mlir::Value::getType(), options, populateDynamicDimSizes(), and mlir::reifyResultShapes().

◆ analyzeModuleOp()

LogicalResult mlir::bufferization::analyzeModuleOp	(	ModuleOp	moduleOp,
		OneShotAnalysisState &	state,
		BufferizationStatistics *	statistics = `nullptr`
	)

Analyze moduleOp and its nested ops.

Bufferization decisions are stored in state.

Definition at line 442 of file OneShotModuleBufferize.cpp.

◆ analyzeOp()

LogicalResult mlir::bufferization::analyzeOp	(	Operation *	op,
		OneShotAnalysisState &	state,
		BufferizationStatistics *	statistics = `nullptr`
	)

Analyze op and its nested ops.

Bufferization decisions are stored in state.

Definition at line 1327 of file OneShotAnalysis.cpp.

◆ bufferizeBlockSignature()

LogicalResult mlir::bufferization::bufferizeBlockSignature	(	Block *	block,
		RewriterBase &	rewriter,
		const BufferizationOptions &	options
	)

Bufferize the signature of block and its callers (i.e., ops that have the given block as a successor).

All block argument types are changed to memref types. All corresponding operands of all callers are wrapped in bufferization.to_buffer ops. All uses of bufferized tensor block arguments are wrapped in bufferization.to_tensor ops.

It is expected that all callers implement the BranchOpInterface. Otherwise, this function will fail. The BranchOpInterface is used to query the range of operands that are forwarded to this block.

It is expected that the parent op of this block implements the BufferizableOpInterface. The buffer types of tensor block arguments are computed with BufferizableOpIntercace::getBufferType.

Definition at line 394 of file Bufferize.cpp.

References mlir::MutableOperandRange::assign(), mlir::OpBuilder::create(), mlir::Operation::emitOpError(), mlir::Block::getArguments(), getBufferType(), mlir::SuccessorOperands::getForwardedOperands(), mlir::SuccessorOperands::getMutableForwardedOperands(), mlir::Block::getParentOp(), mlir::Operation::getSuccessors(), mlir::IRObjectWithUseList< OperandType >::getUsers(), options, mlir::OpBuilder::setInsertionPointAfterValue(), and mlir::OpBuilder::setInsertionPointToStart().

Referenced by mlir::bufferization::func_ext::FuncOpInterface::bufferize().

◆ bufferizeModuleOp()

LogicalResult mlir::bufferization::bufferizeModuleOp	(	ModuleOp	moduleOp,
		const OneShotBufferizationOptions &	options,
		BufferizationStatistics *	statistics = `nullptr`
	)

Bufferize op and its nested ops that implement BufferizableOpInterface.

Note: This function does not run One-Shot Analysis. No buffer copies are inserted except two cases:

options.copyBeforeWrite is set, in which case buffers are copied before every write.
options.copyBeforeWrite is not set and options.noAnalysisFuncFilter is not empty. The FuncOps it contains were not analyzed. Buffer copies will be inserted only to these FuncOps.

Definition at line 513 of file OneShotModuleBufferize.cpp.

References bufferizeOp(), mlir::bufferization::BufferizationOptions::copyBeforeWrite, foldMemRefCasts(), getFuncOpsOrderedByCalls(), mlir::Operation::hasTrait(), options, and removeBufferizationAttributesInModule().

Referenced by mlir::sparse_tensor::SparsificationAndBufferizationPass::runDenseBufferization(), and runOneShotModuleBufferize().

◆ bufferizeOp()

LogicalResult mlir::bufferization::bufferizeOp	(	Operation *	op,
		const BufferizationOptions &	options,
		BufferizationStatistics *	statistics = `nullptr`
	)

Bufferize op and its nested ops that implement BufferizableOpInterface.

Note: This function does not resolve read-after-write conflicts. Use this function only if it is guaranteed that the input IR can bufferize without additional buffer copies or set "options.copyBeforeWrite = true". The general bufferization entry point is runOneShotBufferize.

Check the result of bufferization. Return an error if an op was not bufferized, unless partial bufferization is allowed.

Definition at line 276 of file Bufferize.cpp.

References options.

Referenced by bufferizeModuleOp(), and runOneShotBufferize().

◆ buildBufferDeallocationPipeline()

void mlir::bufferization::buildBufferDeallocationPipeline	(	OpPassManager &	pm,
		const BufferDeallocationPipelineOptions &	options
	)

Adds the buffer deallocation pipeline to the OpPassManager.

This is the standard pipeline for deallocating the MemRefs introduced by the One-Shot bufferization pass.

Definition at line 21 of file BufferizationPipelines.cpp.

References mlir::OpPassManager::addPass(), mlir::createCanonicalizerPass(), mlir::createCSEPass(), and options.

Referenced by registerBufferizationPipelines().

◆ buildDeallocationLibraryFunction()

func::FuncOp mlir::bufferization::buildDeallocationLibraryFunction	(	OpBuilder &	builder,
		Location	loc,
		SymbolTable &	symbolTable
	)

Construct the library function needed for the fully generic bufferization.dealloc lowering implemented in the LowerDeallocations pass.

The function can then be called at bufferization dealloc sites to determine aliasing and ownership.

The generated function takes two memrefs of indices and three memrefs of booleans as arguments:

The first argument A should contain the result of the extract_aligned_pointer_as_index operation applied to the memrefs to be deallocated
The second argument B should contain the result of the extract_aligned_pointer_as_index operation applied to the memrefs to be retained
The third argument C should contain the conditions as passed directly to the deallocation operation.
The fourth argument D is used to pass results to the caller. Those represent the condition under which the memref at the corresponding position in A should be deallocated.
The fifth argument E is used to pass results to the caller. It provides the ownership value corresponding the the memref at the same position in B

This helper function is supposed to be called once for each bufferization.dealloc operation to determine the deallocation need and new ownership indicator for the retained values, but does not perform the deallocation itself.

Generated code:

func.func @dealloc_helper(
    %dyn_dealloc_base_pointer_list: memref<?xindex>,
    %dyn_retain_base_pointer_list: memref<?xindex>,
    %dyn_cond_list: memref<?xi1>,
    %dyn_dealloc_cond_out: memref<?xi1>,
    %dyn_ownership_out: memref<?xi1>) {
  %c0 = arith.constant 0 : index
  %c1 = arith.constant 1 : index
  %true = arith.constant true
  %false = arith.constant false
  %num_dealloc_memrefs = memref.dim %dyn_dealloc_base_pointer_list, %c0
  %num_retain_memrefs = memref.dim %dyn_retain_base_pointer_list, %c0
  // Zero initialize result buffer.
  scf.for %i = %c0 to %num_retain_memrefs step %c1 {
    memref.store %false, %dyn_ownership_out[%i] : memref<?xi1>
  }
  scf.for %i = %c0 to %num_dealloc_memrefs step %c1 {
    %dealloc_bp = memref.load %dyn_dealloc_base_pointer_list[%i]
    %cond = memref.load %dyn_cond_list[%i]
    // Check for aliasing with retained memrefs.
    %does_not_alias_retained = scf.for %j = %c0 to %num_retain_memrefs
        step %c1 iter_args(%does_not_alias_aggregated = %true) -> (i1) {
      %retain_bp = memref.load %dyn_retain_base_pointer_list[%j]
      %does_alias = arith.cmpi eq, %retain_bp, %dealloc_bp : index
      scf.if %does_alias {
        %curr_ownership = memref.load %dyn_ownership_out[%j]
        %updated_ownership = arith.ori %curr_ownership, %cond : i1
        memref.store %updated_ownership, %dyn_ownership_out[%j]
      }
      %does_not_alias = arith.cmpi ne, %retain_bp, %dealloc_bp : index
      %updated_aggregate = arith.andi %does_not_alias_aggregated,
                                      %does_not_alias : i1
      scf.yield %updated_aggregate : i1
    }
    // Check for aliasing with dealloc memrefs in the list before the
    // current one, i.e.,
    // `fix i, forall j < i: check_aliasing(%dyn_dealloc_base_pointer[j],
    // %dyn_dealloc_base_pointer[i])`
    %does_not_alias_any = scf.for %j = %c0 to %i step %c1
       iter_args(%does_not_alias_agg = %does_not_alias_retained) -> (i1) {
      %prev_dealloc_bp = memref.load %dyn_dealloc_base_pointer_list[%j]
      %does_not_alias = arith.cmpi ne, %prev_dealloc_bp, %dealloc_bp
      %updated_alias_agg = arith.andi %does_not_alias_agg, %does_not_alias
      scf.yield %updated_alias_agg : i1
    }
    %dealloc_cond = arith.andi %does_not_alias_any, %cond : i1
    memref.store %dealloc_cond, %dyn_dealloc_cond_out[%i] : memref<?xi1>
  }
  return
}

Definition at line 427 of file LowerDeallocations.cpp.

References mlir::OpBuilder::clearInsertionPoint(), mlir::OpBuilder::create(), mlir::get(), mlir::Builder::getBoolAttr(), mlir::Builder::getFunctionType(), mlir::Builder::getI1Type(), mlir::Builder::getIndexAttr(), mlir::Builder::getIndexType(), mlir::SymbolTable::insert(), mlir::SymbolTable::Private, and mlir::OpBuilder::setInsertionPointToStart().

◆ buildSubsetExtraction()

Value mlir::bufferization::buildSubsetExtraction	(	RewriterBase &	rewriter,
		SubsetInsertionOpInterface	op,
		tensor::EmptyOp	emptyTensorOp,
		Operation *	user
	)

This method builds and returns a subset extraction value for the destination tensor that the given op inserts into.

It returns a value which should replace the emptyTensorOp use that is being consumed by user. If no such a value found it will return an empty Value.

Definition at line 96 of file EmptyTensorElimination.cpp.

References findValidInsertionPoint(), and mlir::OpBuilder::setInsertionPoint().

◆ castOrReallocMemRefValue()

FailureOr< Value > mlir::bufferization::castOrReallocMemRefValue	(	OpBuilder &	b,
		Value	value,
		MemRefType	type,
		const BufferizationOptions &	options
	)

Try to cast the given ranked MemRef-typed value to the given ranked MemRef type.

Insert a reallocation + copy if it cannot be statically guaranteed that a direct cast would be valid.

E.g., when casting from a ranked MemRef type with dynamic layout to a ranked MemRef type with static layout, it is not statically known whether the cast will succeed or not. Such memref.cast ops may fail at runtime. This function never generates such casts and conservatively inserts a copy.

This function returns failure() in case of unsupported casts. E.g., casts with differing element types or memory spaces.

Definition at line 26 of file BufferizationOps.cpp.

References copy(), mlir::OpBuilder::create(), mlir::Value::getLoc(), mlir::Value::getType(), and options.

Referenced by mlir::bufferization::func_ext::CallOpInterface::bufferize(), and foldToBufferToTensorPair().

◆ createPromoteBuffersToStackPass()

std::unique_ptr< Pass > mlir::bufferization::createPromoteBuffersToStackPass ( std::function< bool(Value)> isSmallAlloc )

Creates a pass that promotes heap-based allocations to stack-based ones.

Only buffers smaller with isSmallAlloc(alloc) == true are promoted.

Definition at line 478 of file BufferOptimizations.cpp.

◆ deallocateBuffersOwnershipBased()

LogicalResult mlir::bufferization::deallocateBuffersOwnershipBased	(	FunctionOpInterface	op,
		DeallocationOptions	options
	)

Run the ownership-based buffer deallocation.

Definition at line 1051 of file OwnershipBasedBufferDeallocation.cpp.

References options.

◆ dropEquivalentBufferResults()

LogicalResult mlir::bufferization::dropEquivalentBufferResults ( ModuleOp module )

Drop all memref function results that are equivalent to a function argument.

Definition at line 71 of file DropEquivalentBufferResults.cpp.

References mlir::OpBuilder::create(), mlir::detail::enumerate(), getAssumedUniqueReturnOp(), getCalledFunction(), mlir::Value::getDefiningOp(), mlir::Value::getType(), mlir::RewriterBase::replaceOp(), and mlir::OpBuilder::setInsertionPoint().

◆ eliminateEmptyTensors() [1/2]

LogicalResult mlir::bufferization::eliminateEmptyTensors	(	RewriterBase &	rewriter,
		Operation *	op
	)

Try to eliminate "tensor.empty" ops inside op.

This transformation looks for subset ops that insert a tensor that originates from a "tensor.empty" (as per the reverse use-def chain). Such "tensor.empty" ops are replaced with the destination subset.

E.g.: %0 = tensor.empty() : tensor<10xf32> %1 = linalg.fill ... outs(%0 : tensor<10xf32>) %2 = tensor.insert_slice %0 into t ...

In the above example, the subset op is "tensor.insert_slice". When tracing back the reverse use-def chain of a the source, we end up at a "tensor.empty" op.

Definition at line 203 of file EmptyTensorElimination.cpp.

References options.

◆ eliminateEmptyTensors() [2/2]

LogicalResult mlir::bufferization::eliminateEmptyTensors	(	RewriterBase &	rewriter,
		Operation *	op,
		OneShotAnalysisState &	state,
		ControlBuildSubsetExtractionFn	subsetsExtractionFn = `buildSubsetExtraction`
	)

Try to eliminate "tensor.empty" ops inside op.

This function overload accepts an existing OneShotAnalysisState, which contains in-place bufferization decisions. This overload is useful if an existing analysis should be reused for empty tensor elimination.

Definition at line 118 of file EmptyTensorElimination.cpp.

◆ findCommonDominator()

template<typename DominatorT >

Block* mlir::bufferization::findCommonDominator	(	Value	value,
		const BufferViewFlowAnalysis::ValueSetT &	values,
		const DominatorT &	doms
	)

Finds a common dominator for the given value while taking the positions of the values in the value set into account.

It supports dominator and post-dominator analyses via template arguments. If no common dominator can be found, this function will return "nullptr".

Definition at line 81 of file BufferUtils.h.

References mlir::Value::getParentBlock(), and mlir::Operation::getUsers().

◆ foldToBufferToTensorPair()

LogicalResult mlir::bufferization::foldToBufferToTensorPair	(	RewriterBase &	rewriter,
		ToBufferOp	toBuffer,
		const BufferizationOptions &	options
	)

Try to fold to_buffer(to_tensor(x)).

If x's type and the result type of the to_buffer op are different, a memref.cast is needed.

Definition at line 86 of file BufferizationOps.cpp.

References castOrReallocMemRefValue(), options, mlir::RewriterBase::replaceOp(), and mlir::RewriterBase::replaceOpWithNewOp().

◆ getBuffer()

FailureOr< Value > mlir::bufferization::getBuffer	(	RewriterBase &	rewriter,
		Value	value,
		const BufferizationOptions &	options
	)

Lookup the buffer for the given value.

If the value was not bufferized yet, wrap it in a ToBufferOp. Otherwise, it is the result of a ToTensorOp, from which the memref operand is returned.

Definition at line 663 of file BufferizableOpInterface.cpp.

References mlir::OpBuilder::create(), ensureToBufferOpIsValid(), getBufferType(), mlir::Value::getDefiningOp(), mlir::Value::getLoc(), mlir::Value::getType(), options, and setInsertionPointAfter().

Referenced by mlir::bufferization::func_ext::CallOpInterface::bufferize().

◆ getBufferType() [1/2]

FailureOr< BaseMemRefType > mlir::bufferization::getBufferType	(	Value	value,
		const BufferizationOptions &	options
	)

Return the buffer type for a given Value (tensor) after bufferization without bufferizing any IR.

Return the buffer type for a given Value (tensor) after bufferization.

Note: It should be sufficient to call getBuffer()->getType() in most cases. However, when a buffer type should be predicted without modifying any IR, this function can be used.

This function is a wrapper around BufferizableOpInterface::getBufferType.

Definition at line 688 of file BufferizableOpInterface.cpp.

References options.

Referenced by allocateTensorForShapedValue(), mlir::bufferization::func_ext::CallOpInterface::bufferize(), bufferizeBlockSignature(), getBuffer(), and mlir::bufferization::OpWithUnstructuredControlFlowBufferizableOpInterfaceExternalModel< ConcreteModel, ConcreteOp >::getBufferType().

◆ getBufferType() [2/2]

FailureOr< BaseMemRefType > mlir::bufferization::getBufferType	(	Value	value,
		const BufferizationOptions &	options,
		SmallVector< Value > &	invocationStack
	)

Return the buffer type for a given Value (tensor) after bufferization without bufferizing any IR.

Return the buffer type for a given Value (tensor) after bufferization.

This function (and not the other overload without invocationStack) can be used from getBufferType implementations of the BufferizableOpInterface.

Note: It should be sufficient to call getBuffer()->getType() in most cases. However, when a buffer type should be predicted without modifying any IR, this function can be used.

This function is a wrapper around BufferizableOpInterface::getBufferType.

Definition at line 695 of file BufferizableOpInterface.cpp.

References getMemRefType(), getOwnerOfValue(), mlir::Value::getType(), and options.

◆ getGlobalFor()

FailureOr< memref::GlobalOp > mlir::bufferization::getGlobalFor	(	arith::ConstantOp	constantOp,
		uint64_t	alignment,
		Attribute	memorySpace = `{}`
	)

Definition at line 106 of file BufferUtils.cpp.

References mlir::OpBuilder::create(), mlir::get(), mlir::Builder::getI64Type(), getMemRefTypeWithStaticIdentityLayout(), mlir::Region::getOps(), mlir::Operation::getRegion(), mlir::Builder::getStringAttr(), mlir::SymbolTable::insert(), and mlir::MemRefType::Builder::setMemorySpace().

◆ getMemRefType()

BaseMemRefType mlir::bufferization::getMemRefType	(	Value	value,
		const BufferizationOptions &	options,
		MemRefLayoutAttrInterface	layout = `{}`,
		Attribute	memorySpace = `nullptr`
	)

Return a MemRefType to which the type of the given value can be bufferized.

If possible, op bufferization implementations should not use this function and instead infer precise memref types for tensor results by themselves.

Unless a layout map was specified, options.unknownTypeConverterFn determines what kind of layout map will be used. For best composability (without copies), the fully dynamic layout map is used by default.

Note: Canonicalization patterns could clean up layout maps and infer more precise layout maps after bufferization. However, many possible canonicalizations are currently not implemented.

Definition at line 788 of file BufferizableOpInterface.cpp.

References mlir::get(), mlir::Value::getType(), and options.

Referenced by mlir::bufferization::detail::defaultGetBufferType(), and getBufferType().

◆ getMemRefTypeWithFullyDynamicLayout()

BaseMemRefType mlir::bufferization::getMemRefTypeWithFullyDynamicLayout	(	TensorType	tensorType,
		Attribute	memorySpace = `nullptr`
	)

Return a MemRef type with fully dynamic layout.

If the given tensor type is unranked, return an unranked MemRef type.

Definition at line 814 of file BufferizableOpInterface.cpp.

References mlir::get(), and mlir::Type::getContext().

Referenced by createMemcpy(), mlir::bufferization::OpWithUnstructuredControlFlowBufferizableOpInterfaceExternalModel< ConcreteModel, ConcreteOp >::getBufferType(), mlir::sparse_tensor::LoopEmitter::initializeLoopEmit(), and mlir::bufferization::BufferizationOptions::setFunctionBoundaryTypeConversion().

◆ getMemRefTypeWithStaticIdentityLayout()

BaseMemRefType mlir::bufferization::getMemRefTypeWithStaticIdentityLayout	(	TensorType	tensorType,
		Attribute	memorySpace = `nullptr`
	)

Return a MemRef type with a static identity layout (i.e., no layout map).

If the given tensor type is unranked, return an unranked MemRef type.

Definition at line 838 of file BufferizableOpInterface.cpp.

References mlir::get().

Referenced by mlir::getBufferizationOptionsForSparsification(), getGlobalFor(), and mlir::bufferization::BufferizationOptions::setFunctionBoundaryTypeConversion().

◆ getNextEnclosingRepetitiveRegion()

Region * mlir::bufferization::getNextEnclosingRepetitiveRegion	(	Region *	region,
		const BufferizationOptions &	options
	)

Assuming that the given region is repetitive, find the next enclosing repetitive region.

Definition at line 128 of file BufferizableOpInterface.cpp.

References mlir::Region::getParentRegion(), isRepetitiveRegion(), and options.

◆ getOwnerOfValue()

Operation * mlir::bufferization::getOwnerOfValue ( Value value )

Return the owner of the given value.

In case of a BlockArgument that is the owner of the block. In case of an OpResult that is the defining op.

Definition at line 153 of file BufferizableOpInterface.cpp.

References mlir::Operation::getParentOp().

Referenced by allocateTensorForShapedValue(), mlir::bufferization::detail::defaultGetBufferType(), mlir::bufferization::BufferizationOptions::dynCastBufferizableOp(), mlir::bufferization::AnalysisState::getAliasingOpOperands(), getBufferType(), and mlir::bufferization::OneShotAnalysisState::isWritable().

◆ getParallelRegion()

Region * mlir::bufferization::getParallelRegion	(	Region *	region,
		const BufferizationOptions &	options
	)

If region is a parallel region, return region.

Otherwise, find the first enclosing parallel region of region. If there is no such region, return "nullptr".

Note: Whether a region is parallel or sequential is queried from the BufferizableOpInterface.

Definition at line 138 of file BufferizableOpInterface.cpp.

References mlir::Region::getParentOp(), mlir::Region::getParentRegion(), mlir::Region::getRegionNumber(), isRepetitiveRegion(), and options.

◆ getReturnOps()

SmallVector< func::ReturnOp > mlir::bufferization::getReturnOps ( func::FuncOp funcOp )

Helper function that returns all func.return ops in the given function.

Return all func.return ops in the given function.

Definition at line 23 of file FuncBufferizableOpInterfaceImpl.cpp.

Referenced by mlir::bufferization::func_ext::FuncOpInterface::bufferize(), and foldMemRefCasts().

◆ hasTensorSemantics()

bool mlir::bufferization::hasTensorSemantics ( Operation * op )

Return "true" if the given op has tensor semantics and should be bufferized.

If the op is bufferizable, the BufferizableOpInterface is queried. Otherwise, an op has tensor semantics if it has tensor operands, tensor op results and/or tensor block arguments.

Definition at line 718 of file BufferizableOpInterface.cpp.

References mlir::bufferization::detail::defaultHasTensorSemantics().

Referenced by mlir::bufferization::OneShotAnalysisState::analyzeOp(), mlir::vector::getMixedSizesXfer(), mlir::linalg::transposeBatchMatmul(), and mlir::linalg::transposeMatmul().

◆ hoistBuffersFromLoops()

void mlir::bufferization::hoistBuffersFromLoops ( Operation * op )

Within the given operation, hoist buffers from loops where possible.

See "BufferLoopHoistingPass" for more information.

Definition at line 473 of file BufferOptimizations.cpp.

◆ insertTensorCopies() [1/2]

LogicalResult mlir::bufferization::insertTensorCopies	(	Operation *	op,
		const AnalysisState &	state
	)

Resolve RaW and other conflicts by inserting bufferization.alloc_tensor ops.

After applying this transform, the IR can be bufferized without inserting additional buffer allocations.

Definition at line 51 of file TensorCopyInsertion.cpp.

◆ insertTensorCopies() [2/2]

LogicalResult mlir::bufferization::insertTensorCopies	(	Operation *	op,
		const OneShotBufferizationOptions &	options,
		BufferizationStatistics *	statistics = `nullptr`
	)

Resolve RaW and other conflicts by inserting bufferization.alloc_tensor ops.

After applying this transform, the IR can be bufferized without inserting additional buffer allocations.

Definition at line 29 of file TensorCopyInsertion.cpp.

Referenced by runOneShotBufferize(), runOneShotModuleBufferize(), and mlir::sparse_tensor::SparsificationAndBufferizationPass::runOnOperation().

◆ populateBufferizationDeallocLoweringPattern()

void mlir::bufferization::populateBufferizationDeallocLoweringPattern	(	RewritePatternSet &	patterns,
		const DeallocHelperMap &	deallocHelperFuncMap
	)

Adds the conversion pattern of the bufferization.dealloc operation to the given pattern set for use in other transformation passes.

Definition at line 543 of file LowerDeallocations.cpp.

References mlir::patterns.

◆ populateDeallocOpCanonicalizationPatterns()

void mlir::bufferization::populateDeallocOpCanonicalizationPatterns	(	RewritePatternSet &	patterns,
		MLIRContext *	context
	)

Add the canonicalization patterns for bufferization.dealloc to the given pattern set to make them available to other passes (such as BufferDeallocationSimplification).

Definition at line 1173 of file BufferizationOps.cpp.

References mlir::patterns.

◆ populateDynamicDimSizes()

void mlir::bufferization::populateDynamicDimSizes	(	OpBuilder &	b,
		Location	loc,
		Value	shapedValue,
		SmallVector< Value > &	dynamicDims
	)

Populate dynamicDims with tensor::DimOp / memref::DimOp results for all dynamic dimensions of the given shaped value.

Definition at line 131 of file BufferizationOps.cpp.

References mlir::OpBuilder::create(), and mlir::Value::getType().

Referenced by allocateTensorForShapedValue().

◆ populateEmptyTensorToAllocTensorPattern()

void mlir::bufferization::populateEmptyTensorToAllocTensorPattern ( RewritePatternSet & patterns )

Populate patterns to lower tensor.empty ops to bufferization.alloc_tensor ops.

Definition at line 52 of file EmptyTensorToAllocTensor.cpp.

References mlir::patterns.

◆ promoteBufferResultsToOutParams()

LogicalResult mlir::bufferization::promoteBufferResultsToOutParams	(	ModuleOp	module,
		const BufferResultsToOutParamsOpts &	options
	)

Replace buffers that are returned from a function with an out parameter.

Also update all call sites.

Definition at line 214 of file BufferResultsToOutParams.cpp.

References options, updateCalls(), updateFuncOp(), and updateReturnOps().

◆ registerBufferizationPipelines()

void mlir::bufferization::registerBufferizationPipelines ( )

Registers all pipelines for the bufferization dialect.

Currently, this includes only the "buffer-deallocation-pipeline".

Definition at line 42 of file BufferizationPipelines.cpp.

References buildBufferDeallocationPipeline().

Referenced by mlir::registerAllPasses().

◆ registerTransformDialectExtension()

void mlir::bufferization::registerTransformDialectExtension ( DialectRegistry & registry )

Definition at line 175 of file BufferizationTransformOps.cpp.

References mlir::DialectRegistry::addExtensions().

Referenced by mlir::registerAllExtensions().

◆ removeBufferizationAttributesInModule()

void mlir::bufferization::removeBufferizationAttributesInModule ( ModuleOp moduleOp )

Remove bufferization attributes on every FuncOp arguments in the ModuleOp.

Definition at line 505 of file OneShotModuleBufferize.cpp.

References removeBufferizationAttributes().

Referenced by bufferizeModuleOp(), and mlir::sparse_tensor::SparsificationAndBufferizationPass::runDenseBufferization().

◆ replaceOpWithBufferizedValues()

void mlir::bufferization::replaceOpWithBufferizedValues	(	RewriterBase &	rewriter,
		Operation *	op,
		ValueRange	values
	)

Replace an op with replacement values.

The op is deleted. Tensor OpResults must be replaced with memref values.

Definition at line 724 of file BufferizableOpInterface.cpp.

References mlir::OpBuilder::create(), mlir::Value::getLoc(), mlir::Operation::getNumResults(), mlir::Operation::getOpResults(), mlir::Value::getType(), mlir::RewriterBase::replaceOp(), and mlir::OpBuilder::setInsertionPointAfter().

Referenced by mlir::bufferization::func_ext::CallOpInterface::bufferize(), and replaceOpWithNewBufferizedOp().

◆ replaceOpWithNewBufferizedOp()

template<typename OpTy , typename... Args>

OpTy mlir::bufferization::replaceOpWithNewBufferizedOp	(	RewriterBase &	rewriter,
		Operation *	op,
		Args &&...	args
	)

Replace an op with a new op.

The new op must have the same number of results as the replaced op. The new op may not return any tensor values.

Definition at line 634 of file BufferizableOpInterface.h.

References mlir::OpBuilder::create(), mlir::Operation::getLoc(), mlir::Operation::getResults(), and replaceOpWithBufferizedValues().

◆ runOneShotBufferize()

LogicalResult mlir::bufferization::runOneShotBufferize	(	Operation *	op,
		const OneShotBufferizationOptions &	options,
		BufferizationStatistics *	statistics = `nullptr`
	)

Run One-Shot Bufferize on the given op: Analysis + Bufferization.

Definition at line 1369 of file OneShotAnalysis.cpp.

References bufferizeOp(), insertTensorCopies(), and options.

◆ runOneShotModuleBufferize()

LogicalResult mlir::bufferization::runOneShotModuleBufferize	(	ModuleOp	moduleOp,
		const bufferization::OneShotBufferizationOptions &	options,
		BufferizationStatistics *	statistics = `nullptr`
	)

Run One-Shot Module Bufferization on the given module.

Performs a simple function call analysis to determine which function arguments are inplaceable. Then analyzes and bufferizes FuncOps one-by-one with One-Shot Bufferize.

Definition at line 578 of file OneShotModuleBufferize.cpp.

References bufferizeModuleOp(), mlir::bufferization::OpFilter::denyOperation(), mlir::Operation::getParentOfType(), insertTensorCopies(), mlir::bufferization::BufferizationOptions::opFilter, and options.

Namespaces

Classes

Typedefs

Enumerations

Functions

Typedef Documentation

◆ AliasingOpOperandList

◆ AliasingValueList

◆ ControlBuildSubsetExtractionFn

◆ DeallocHelperMap

◆ RegisterDependenciesFn

Enumeration Type Documentation

◆ BufferRelation

Function Documentation

◆ allocateTensorForShapedValue()

◆ analyzeModuleOp()

◆ analyzeOp()

◆ bufferizeBlockSignature()

◆ bufferizeModuleOp()

◆ bufferizeOp()

◆ buildBufferDeallocationPipeline()

◆ buildDeallocationLibraryFunction()

◆ buildSubsetExtraction()

◆ castOrReallocMemRefValue()

◆ createPromoteBuffersToStackPass()

◆ deallocateBuffersOwnershipBased()

◆ dropEquivalentBufferResults()

◆ eliminateEmptyTensors() [1/2]

◆ eliminateEmptyTensors() [2/2]

◆ findCommonDominator()

◆ foldToBufferToTensorPair()

◆ getBuffer()

◆ getBufferType() [1/2]

◆ getBufferType() [2/2]

◆ getGlobalFor()

◆ getMemRefType()

◆ getMemRefTypeWithFullyDynamicLayout()

◆ getMemRefTypeWithStaticIdentityLayout()

◆ getNextEnclosingRepetitiveRegion()

◆ getOwnerOfValue()

◆ getParallelRegion()

◆ getReturnOps()

◆ hasTensorSemantics()

◆ hoistBuffersFromLoops()

◆ insertTensorCopies() [1/2]

◆ insertTensorCopies() [2/2]

◆ populateBufferizationDeallocLoweringPattern()

◆ populateDeallocOpCanonicalizationPatterns()

◆ populateDynamicDimSizes()

◆ populateEmptyTensorToAllocTensorPattern()

◆ promoteBufferResultsToOutParams()

◆ registerBufferizationPipelines()

◆ registerTransformDialectExtension()

◆ removeBufferizationAttributesInModule()

◆ replaceOpWithBufferizedValues()

◆ replaceOpWithNewBufferizedOp()

◆ runOneShotBufferize()

◆ runOneShotModuleBufferize()