#include <sdfg/data_flow/pointer_metadata.h>

Inheritance diagram for sdfg::data_flow::MemoryAccessPattern:

Public Member Functions
virtual	~MemoryAccessPattern ()=default
virtual bool	empty () const =0
	There are no elements in this pattern = no accesses of this type.
virtual bool	every_element_accessed () const =0
	Every element that is part of this pattern is actually accessed.
virtual const math::tensor::TensorLayout *	layout () const
	The structured affine layout (shape/strides/offset, in elements) backing this pattern, or nullptr when the pattern carries no explicit layout (a flat convex span or no access).
MemoryAccessPatternType	ref () const
virtual void	replace (const symbolic::Expression old_expression, const symbolic::Expression new_expression)=0
virtual void	replace (const symbolic::ExpressionMapping &replacements)=0
virtual MemoryAccessPatternType	clone () const =0
virtual void	serialize_to_json (nlohmann::json &entry)=0

Constructor & Destructor Documentation

◆ ~MemoryAccessPattern()

virtual sdfg::data_flow::MemoryAccessPattern::~MemoryAccessPattern ( )

virtualdefault

Member Function Documentation

◆ clone()

virtual MemoryAccessPatternType sdfg::data_flow::MemoryAccessPattern::clone ( ) const

pure virtual

Implemented in sdfg::data_flow::ConvexAccessPattern, sdfg::data_flow::NoAccessPattern, and sdfg::data_flow::TensorLayoutPattern.

◆ empty()

virtual bool sdfg::data_flow::MemoryAccessPattern::empty ( ) const

pure virtual

There are no elements in this pattern = no accesses of this type.

Implemented in sdfg::data_flow::ConvexAccessPattern, sdfg::data_flow::NoAccessPattern, and sdfg::data_flow::TensorLayoutPattern.

◆ every_element_accessed()

virtual bool sdfg::data_flow::MemoryAccessPattern::every_element_accessed ( ) const

pure virtual

Every element that is part of this pattern is actually accessed.

If false, this is just an approximation that may contain elements in between the bounds that are not guaranteed to be accessed Ex. indicating write of a[0] and a[9] as ConvexPattern(10) is approximate. Writes a[0], a[1], .. [9] would be every element being accessed. Bounds should never be violated regardless. We may in the future

Implemented in sdfg::data_flow::ConvexAccessPattern, sdfg::data_flow::NoAccessPattern, and sdfg::data_flow::TensorLayoutPattern.

◆ layout()

virtual const math::tensor::TensorLayout * sdfg::data_flow::MemoryAccessPattern::layout ( ) const

inlinevirtual

The structured affine layout (shape/strides/offset, in elements) backing this pattern, or nullptr when the pattern carries no explicit layout (a flat convex span or no access).

Only TensorLayoutPattern provides one.

Reimplemented in sdfg::data_flow::TensorLayoutPattern.