ITensorHandle Class Reference

#include <ITensorHandle.hpp>

Inheritance diagram for ITensorHandle:

Public Member Functions
virtual	~ITensorHandle ()
virtual void	Manage ()=0
virtual void	Allocate ()=0
virtual ITensorHandle *	GetParent () const =0
virtual const void *	Map (bool blocking=true) const =0
virtual void	Unmap () const =0
	Unmap the tensor data. More...
void *	Map (bool blocking=true)
void	Unmap ()
	Unmap the tensor data that was previously mapped with call to Map(). More...
virtual TensorShape	GetStrides () const =0
virtual TensorShape	GetShape () const =0
virtual void	CopyOutTo (void *memory) const =0
virtual void	CopyInFrom (const void *memory)=0
virtual unsigned int	GetImportFlags () const
	Get flags describing supported import sources. More...
virtual bool	Import (void *memory, MemorySource source)

Detailed Description

Definition at line 16 of file ITensorHandle.hpp.

Constructor & Destructor Documentation

~ITensorHandle()

virtual ~ITensorHandle ( )

inlinevirtual

Definition at line 19 of file ITensorHandle.hpp.

References ITensorHandle::Allocate(), ITensorHandle::GetParent(), ITensorHandle::Manage(), ITensorHandle::Map(), and ITensorHandle::Unmap().

{}

Member Function Documentation

Allocate()

virtual void Allocate ( )

pure virtual

Indicate to the memory manager that this resource is no longer active. This is used to compute overlapping lifetimes of resources.

Implemented in NeonSubTensorHandle, ClSubTensorHandle, ConstPassthroughCpuTensorHandle, PassthroughCpuTensorHandle, ScopedCpuTensorHandle, ClTensorHandle, NeonTensorHandle, SampleTensorHandle, and RefTensorHandle.

Referenced by AllocateAndCopyDataToITensorHandle(), Graph::AllocateDynamicBuffers(), ConstPassthroughCpuTensorHandle::ConstPassthroughCpuTensorHandle(), PassthroughCpuTensorHandle::PassthroughCpuTensorHandle(), and ITensorHandle::~ITensorHandle().

CopyInFrom()

virtual void CopyInFrom ( const void *  memory )

pure virtual

Referenced by CopyDataToITensorHandle(), and ITensorHandle::Unmap().

CopyOutTo()

virtual void CopyOutTo ( void *  memory ) const

pure virtual

Referenced by CopyDataFromITensorHandle(), and ITensorHandle::Unmap().

GetImportFlags()

virtual unsigned int GetImportFlags ( ) const

inlinevirtual

Get flags describing supported import sources.

Reimplemented in NeonTensorHandle, SampleTensorHandle, and RefTensorHandle.

Definition at line 70 of file ITensorHandle.hpp.

Referenced by LoadedNetwork::EnqueueWorkload().

{ return 0; }

GetParent()

virtual ITensorHandle* GetParent ( ) const

pure virtual

Get the parent tensor if this is a subtensor.

Returns

a pointer to the parent tensor. Otherwise nullptr if not a subtensor.

Implemented in NeonSubTensorHandle, ClSubTensorHandle, NeonTensorHandle, ClTensorHandle, ConstCpuTensorHandle, SampleTensorHandle, and RefTensorHandle.

Referenced by Graph::AllocateDynamicBuffers(), and ITensorHandle::~ITensorHandle().

GetShape()

virtual TensorShape GetShape ( ) const

pure virtual

Get the number of elements for each dimension ordered from slowest iterating dimension to fastest iterating dimension.

Returns

a TensorShape filled with the number of elements for each dimension.

Implemented in NeonSubTensorHandle, ClSubTensorHandle, NeonTensorHandle, ClTensorHandle, ConstCpuTensorHandle, SampleTensorHandle, and RefTensorHandle.

Referenced by armnn::CopyTensorContentsGeneric(), armnnUtils::FindMinMax(), and ITensorHandle::Unmap().

GetStrides()

virtual TensorShape GetStrides ( ) const

pure virtual

Get the strides for each dimension ordered from largest to smallest where the smallest value is the same as the size of a single element in the tensor.

Returns

a TensorShape filled with the strides for each dimension

Implemented in NeonSubTensorHandle, ClSubTensorHandle, NeonTensorHandle, ClTensorHandle, ConstCpuTensorHandle, SampleTensorHandle, and RefTensorHandle.

Referenced by armnn::CopyTensorContentsGeneric(), and ITensorHandle::Unmap().

Import()

virtual bool Import ( void *  memory, MemorySource  source  )

inlinevirtual

Import externally allocated memory

Parameters

memory	base address of the memory being imported.
source	source of the allocation for the memory being imported.

Returns

true on success or false on failure

Reimplemented in NeonTensorHandle, SampleTensorHandle, and RefTensorHandle.

Definition at line 76 of file ITensorHandle.hpp.

Referenced by LoadedNetwork::EnqueueWorkload().

    {
        boost::ignore_unused(memory, source);
        return false;
    };

Manage()

virtual void Manage ( )

pure virtual

Indicate to the memory manager that this resource is active. This is used to compute overlapping lifetimes of resources.

Implemented in NeonSubTensorHandle, ClSubTensorHandle, NeonTensorHandle, ClTensorHandle, ConstCpuTensorHandle, SampleTensorHandle, and RefTensorHandle.

Referenced by Graph::AllocateDynamicBuffers(), and ITensorHandle::~ITensorHandle().

Map() [1/2]

virtual const void* Map ( bool  blocking = true ) const

pure virtual

Map the tensor data for access.

Parameters

blocking

hint to block the calling thread until all other accesses are complete. (backend dependent)

Returns

pointer to the first element of the mapped data.

Implemented in NeonSubTensorHandle, ClSubTensorHandle, NeonTensorHandle, ClTensorHandle, ConstCpuTensorHandle, SampleTensorHandle, and RefTensorHandle.

Referenced by armnn::CopyTensorContentsGeneric(), LoadedNetwork::EnqueueWorkload(), SyncMemGenericWorkload::Execute(), armnnUtils::FindMinMax(), armnn::GetInputTensorData(), armnn::GetOutputTensorData(), RefTensorHandle::GetParent(), SampleTensorHandle::GetParent(), ITensorHandle::Map(), and ITensorHandle::~ITensorHandle().

Map() [2/2]

void* Map ( bool  blocking = true )

inline

Map the tensor data for access. Must be paired with call to Unmap().

Parameters

blocking

hint to block the calling thread until all other accesses are complete. (backend dependent)

Returns

pointer to the first element of the mapped data.

Definition at line 44 of file ITensorHandle.hpp.

References ITensorHandle::Map().

    {
        return const_cast<void*>(static_cast<const ITensorHandle*>(this)->Map(blocking));
    }

Unmap() [1/2]

virtual void Unmap ( ) const

pure virtual

Unmap the tensor data.

Implemented in NeonSubTensorHandle, ClSubTensorHandle, NeonTensorHandle, ClTensorHandle, ConstCpuTensorHandle, SampleTensorHandle, and RefTensorHandle.

Referenced by armnn::CopyTensorContentsGeneric(), LoadedNetwork::EnqueueWorkload(), SyncMemGenericWorkload::Execute(), armnnUtils::FindMinMax(), ITensorHandle::Unmap(), and ITensorHandle::~ITensorHandle().

Unmap() [2/2]

void Unmap ( )

inline

Unmap the tensor data that was previously mapped with call to Map().

Definition at line 50 of file ITensorHandle.hpp.

References ITensorHandle::CopyInFrom(), ITensorHandle::CopyOutTo(), ITensorHandle::GetShape(), ITensorHandle::GetStrides(), and ITensorHandle::Unmap().

    {
        return static_cast<const ITensorHandle*>(this)->Unmap();
    }

---

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

• include/armnn/backends/ITensorHandle.hpp