mlir::dataflow Namespace Reference

Classes
class	ConstantValue
	This lattice value represents a known constant value of a lattice. More...
class	SparseConstantPropagation
	This analysis implements sparse constant propagation, which attempts to determine constant-valued results for operations using constant-valued operands, by speculatively folding operations. More...
class	Executable
	This is a simple analysis state that represents whether the associated lattice anchor (either a block or a control-flow edge) is live. More...
class	PredecessorState
	This analysis state represents a set of live control-flow "predecessors" of a program point (either an operation or a block), which are the last operations along all execution paths that pass through this point. More...
class	CFGEdge
	This lattice anchor represents a control-flow edge between a block and one of its successors. More...
class	DeadCodeAnalysis
	Dead code analysis analyzes control-flow, as understood by RegionBranchOpInterface and BranchOpInterface, and the callgraph, as understood by CallableOpInterface and CallOpInterface. More...
class	AbstractDenseLattice
	This class represents a dense lattice. More...
class	AbstractDenseForwardDataFlowAnalysis
	Base class for dense forward data-flow analyses. More...
class	DenseForwardDataFlowAnalysis
	A dense forward data-flow analysis for propagating lattices before and after the execution of every operation across the IR by implementing transfer functions for operations. More...
class	AbstractDenseBackwardDataFlowAnalysis
	Base class for dense backward dataflow analyses. More...
class	DenseBackwardDataFlowAnalysis
	A dense backward dataflow analysis propagating lattices after and before the execution of every operation across the IR by implementing transfer functions for opreations. More...
class	IntegerValueRangeLattice
	This lattice element represents the integer value range of an SSA value. More...
class	IntegerRangeAnalysis
	Integer range analysis determines the integer value range of SSA values using operations that define InferIntRangeInterface and also sets the range of iteration indices of loops with known bounds. More...
struct	Liveness
	This lattice represents, for a given value, whether or not it is "live". More...
class	LivenessAnalysis
	An analysis that, by going backwards along the dataflow graph, annotates each value with a boolean storing true iff it is "live". More...
struct	RunLivenessAnalysis
	Runs liveness analysis on the IR defined by op. More...
class	AbstractSparseLattice
	This class represents an abstract lattice. More...
class	Lattice
	This class represents a lattice holding a specific value of type ValueT. More...
class	AbstractSparseForwardDataFlowAnalysis
	Base class for sparse forward data-flow analyses. More...
class	SparseForwardDataFlowAnalysis
	A sparse forward data-flow analysis for propagating SSA value lattices across the IR by implementing transfer functions for operations. More...
class	AbstractSparseBackwardDataFlowAnalysis
	Base class for sparse backward data-flow analyses. More...
class	SparseBackwardDataFlowAnalysis
	A sparse (backward) data-flow analysis for propagating SSA value lattices backwards across the IR by implementing transfer functions for operations. More...

Enumerations
enum class	CallControlFlowAction { EnterCallee , ExitCallee , ExternalCallee }
	Indicates whether the control enters, exits, or skips over the callee (in the case of external functions). More...

Functions
LogicalResult	staticallyNonNegative (DataFlowSolver &solver, Operation *op)
	Succeeds if an op can be converted to its unsigned equivalent without changing its semantics. More...
LogicalResult	staticallyNonNegative (DataFlowSolver &solver, Value v)
	Succeeds when a value is statically non-negative in that it has a lower bound on its value (if it is treated as signed) and that bound is non-negative. More...
LogicalResult	maybeReplaceWithConstant (DataFlowSolver &solver, RewriterBase &rewriter, Value value)
	Patterned after SCCP. More...

Enumeration Type Documentation

CallControlFlowAction

enum mlir::dataflow::CallControlFlowAction

strong

Indicates whether the control enters, exits, or skips over the callee (in the case of external functions).

Enumerator
EnterCallee
ExitCallee
ExternalCallee

Definition at line 32 of file DenseAnalysis.h.

Function Documentation

maybeReplaceWithConstant()

LogicalResult mlir::dataflow::maybeReplaceWithConstant	(	DataFlowSolver &	solver,
		RewriterBase &	rewriter,
		Value	value
	)

Patterned after SCCP.

Definition at line 62 of file IntRangeOptimizations.cpp.

References copyIntegerRange(), mlir::DataFlowSolver::eraseState(), mlir::DenseIntElementsAttr::get(), mlir::Builder::getContext(), mlir::Value::getDefiningOp(), mlir::Operation::getDialect(), mlir::Builder::getIntegerAttr(), mlir::MLIRContext::getLoadedDialect(), mlir::Value::getLoc(), getMaybeConstantValue(), mlir::Region::getParentOp(), mlir::Value::getParentRegion(), mlir::Operation::getResult(), mlir::Value::getType(), mlir::DataFlowSolver::lookupState(), materializeConstant(), mlir::Dialect::materializeConstant(), mlir::RewriterBase::replaceAllUsesWith(), and mlir::Value::use_empty().

staticallyNonNegative() [1/2]

LogicalResult mlir::dataflow::staticallyNonNegative	(	DataFlowSolver &	solver,
		Operation *	op
	)

Succeeds if an op can be converted to its unsigned equivalent without changing its semantics.

This is the case when none of its openands or results can be below 0 when analyzed from a signed perspective.

Definition at line 49 of file IntegerRangeAnalysis.cpp.

References mlir::Operation::getOperands(), and mlir::Operation::getResults().

staticallyNonNegative() [2/2]

LogicalResult mlir::dataflow::staticallyNonNegative	(	DataFlowSolver &	solver,
		Value	v
	)

Succeeds when a value is statically non-negative in that it has a lower bound on its value (if it is treated as signed) and that bound is non-negative.

Note, the results of this query may not be accurate for index if you plan to use a non-64-bit index.

Definition at line 41 of file IntegerRangeAnalysis.cpp.

References mlir::DataFlowSolver::lookupState(), and mlir::ConstantIntRanges::smin().