itk::OpenCLSampler Class Reference

#include <itkOpenCLSampler.h>

Detailed Description

The OpenCLSampler class represents an OpenCL sampler object.

A sampler object describes how to sample an image when the image is read in the kernel. The built-in functions to read from an image in a kernel take a sampler as an argument. The sampler arguments to the image read function can be sampler objects created using OpenCL functions and passed as argument values to the kernel or can be samplers declared inside a kernel. In this section we discuss how sampler objects are created using OpenCL functions.

The GetFilterMode() specifies the type of filter that must be applied when reading an image. This can be {CL_FILTER_NEAREST}, or {CL_FILTER_LINEAR}.

The GetAddressingMode() specifies how out-of-range image coordinates are handled when reading from an image. This can be set to {CL_ADDRESS_MIRRORED_REPEAT}, {CL_ADDRESS_REPEAT}, {CL_ADDRESS_CLAMP_TO_EDGE}, {CL_ADDRESS_CLAMP} and {CL_ADDRESS_NONE}.

Samplers are created using OpenCLContext::CreateSampler(), as follows:

OpenCLSampler sampler = context.CreateSampler
    (false, OpenCLSampler::ClampToEdge, OpenCLSampler::Linear);

Samplers can also be defined as literals in the OpenCL kernel source code, which avoids the need to create an explicit OpenCLSampler value:

__constant sampler_t imageSampler =
CLK_NORMALIZED_COORDS_FALSE | CLK_ADDRESS_CLAMP | CLK_FILTER_NEAREST;

The main advantage of OpenCLSampler over literal sampler values is that OpenCLSampler allows the pixel derivation strategy to be modified at runtime.

See also

OpenCLContext

Definition at line 71 of file itkOpenCLSampler.h.

Public Types
enum	AddressingMode {   None = 0x1130 , ClampToEdge = 0x1131 , Clamp = 0x1132 , Repeat = 0x1133 ,   MirroredRepeat = 0x1134 }
enum	FilterMode { Nearest = 0x1140 , Linear = 0x1141 }
using	Self = OpenCLSampler

Public Member Functions
OpenCLSampler::AddressingMode	GetAddressingMode () const
OpenCLContext *	GetContext () const
OpenCLSampler::FilterMode	GetFilterMode () const
bool	GetNormalizedCoordinates () const
cl_sampler	GetSamplerId () const
bool	IsNull () const
	OpenCLSampler ()
	OpenCLSampler (const OpenCLSampler &other)
	OpenCLSampler (OpenCLContext *context, cl_sampler id)
OpenCLSampler &	operator= (const OpenCLSampler &other)
	~OpenCLSampler ()

Private Attributes
OpenCLContext *	m_Context
cl_sampler	m_Id

Member Typedef Documentation

Self

using itk::OpenCLSampler::Self = OpenCLSampler

Standard class typedefs.

Definition at line 75 of file itkOpenCLSampler.h.

Member Enumeration Documentation

AddressingMode

enum itk::OpenCLSampler::AddressingMode

Enumerator
None
ClampToEdge
Clamp
Repeat
MirroredRepeat

Definition at line 112 of file itkOpenCLSampler.h.

FilterMode

enum itk::OpenCLSampler::FilterMode

Enumerator
Nearest
Linear

Definition at line 127 of file itkOpenCLSampler.h.

Constructor & Destructor Documentation

OpenCLSampler() [1/3]

itk::OpenCLSampler::OpenCLSampler ( )

inline

Constructs a null OpenCL sampler object.

Definition at line 78 of file itkOpenCLSampler.h.

OpenCLSampler() [2/3]

itk::OpenCLSampler::OpenCLSampler ( OpenCLContext * context, cl_sampler id )

inline

Constructs an OpenCL sampler object from the native identifier id. This class takes over ownership of id and will release it in the destructor. The sampler id will be associated with context.

Definition at line 86 of file itkOpenCLSampler.h.

OpenCLSampler() [3/3]

itk::OpenCLSampler::OpenCLSampler

(

const OpenCLSampler &

other

)

Constructs a copy of other. The {clRetainSampler()} function will be called to update the reference count on GetSamplerId().

~OpenCLSampler()

itk::OpenCLSampler::~OpenCLSampler

(

)

Releases this OpenCL sampler object by calling {clReleaseSampler()}.

Member Function Documentation

GetAddressingMode()

OpenCLSampler::AddressingMode itk::OpenCLSampler::GetAddressingMode

(

)

const

Returns the addressing mode for out-of-range co-ordinates when reading from an image in OpenCL.

See also

GetNormalizedCoordinates(), GetFilterMode()

GetContext()

OpenCLContext * itk::OpenCLSampler::GetContext ( ) const

inline

Returns the OpenCL context that this sampler was created for; null if not yet created within a context.

Definition at line 167 of file itkOpenCLSampler.h.

GetFilterMode()

OpenCLSampler::FilterMode itk::OpenCLSampler::GetFilterMode

(

)

const

Returns the type of filter to apply when reading from an image in OpenCL.

See also

GetNormalizedCoordinates(), GetAddressingMode()

GetNormalizedCoordinates()

bool itk::OpenCLSampler::GetNormalizedCoordinates

(

)

const

Returns true if this sampler is using normalized co-ordinates, false otherwise.

See also

GetAddressingMode(), GetFilterMode()

GetSamplerId()

cl_sampler itk::OpenCLSampler::GetSamplerId ( ) const

inline

Returns the native OpenCL identifier for this sampler; or 0 if the sampler is null.

Definition at line 159 of file itkOpenCLSampler.h.

IsNull()

bool itk::OpenCLSampler::IsNull ( ) const

inline

Returns true if this OpenCL sampler object is null, false otherwise.

Definition at line 135 of file itkOpenCLSampler.h.

operator=()

OpenCLSampler & itk::OpenCLSampler::operator=

(

const OpenCLSampler &

other

)

Assigns other to this OpenCL sampler object. The current samplerId() will be released with {clReleaseSampler()}, and the new samplerId() will be retained with {clRetainSampler()}.

Field Documentation

m_Context

OpenCLContext* itk::OpenCLSampler::m_Context

private

Definition at line 173 of file itkOpenCLSampler.h.

m_Id

cl_sampler itk::OpenCLSampler::m_Id

private

Definition at line 174 of file itkOpenCLSampler.h.