Package org.flag4j.operations

Class TransposeDispatcher

java.lang.Object

org.flag4j.operations.TransposeDispatcher

public final class TransposeDispatcher
extends Object

Provides a dispatch method for dynamically choosing the best matrix transpose algorithm.

Nested Class Summary

Nested Classes

Modifier and Type	Class	Description
private static enum	TransposeDispatcher.Algorithm	Simple enum class containing available algorithms for computing a matrix transpose.

Field Summary

Fields

Modifier and Type	Field	Description
private static final int	COMPLEX_BLOCKED_THRESHOLD	Threshold for using complex blocked algorithm.
private static final int	CONCURRENT_THRESHOLD	Threshold for number of elements in matrix to use concurrent implementation.
private static final int	HERMATION_BLOCKED_THRESHOLD	Threshold for using blocked hermitian algorithm
private static final int	STANDARD_THRESHOLD	Threshold for using standard transpose implementation.

Constructor Summary

Constructors

Modifier	Constructor	Description
private	TransposeDispatcher()

Method Summary

Modifier and Type	Method	Description
private static TransposeDispatcher.Algorithm	chooseAlgorithm(Shape shape)	Chooses the appropriate matrix transpose algorithm based on the shape of a matrix.
private static TransposeDispatcher.Algorithm	chooseAlgorithmComplex(Shape shape)	Chooses the appropriate matrix transpose algorithm based on the shape of a matrix.
private static TransposeDispatcher.Algorithm	chooseAlgorithmHermitian(Shape shape)	Chooses the appropriate matrix hermitian transpose algorithm based on the shape of a matrix.
private static TransposeDispatcher.Algorithm	chooseAlgorithmTensor(int numEntries)	Chooses the appropriate algorithm for computing a tensor transpose.
private static TransposeDispatcher.Algorithm	chooseAlgorithmTensor(int length1, int length2)	Chooses the appropriate algorithm for computing a tensor transpose.
static CMatrix	dispatch(CMatrix src)	Dispatches a matrix transpose problem to the appropriate algorithm based in its shape and size.
static Matrix	dispatch(Matrix src)	Dispatches a matrix transpose problem to the appropriate algorithm based on its shape and size.
static CMatrix	dispatchHermitian(CMatrix src)	Dispatches a matrix hermitian transpose problem to the appropriate algorithm based on its shape and size.
static double[]	dispatchTensor(double[] src, Shape shape, int[] axes)	Dispatches a tensor transpose problem to the appropriate algorithm based on its shape and size.
static CTensor	dispatchTensor(CTensor src, int axis1, int axis2)	Dispatches a tensor transpose problem to the appropriate algorithm based on its shape and size.
static Tensor	dispatchTensor(Tensor src, int axis1, int axis2)	Dispatches a tensor transpose problem to the appropriate algorithm based on its shape and size.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Details

COMPLEX_BLOCKED_THRESHOLD

private static final int COMPLEX_BLOCKED_THRESHOLD

Threshold for using complex blocked algorithm.

See Also:

• Constant Field Values

HERMATION_BLOCKED_THRESHOLD

private static final int HERMATION_BLOCKED_THRESHOLD

Threshold for using blocked hermitian algorithm

See Also:

• Constant Field Values

STANDARD_THRESHOLD

private static final int STANDARD_THRESHOLD

Threshold for using standard transpose implementation.

See Also:

• Constant Field Values

CONCURRENT_THRESHOLD

private static final int CONCURRENT_THRESHOLD

Threshold for number of elements in matrix to use concurrent implementation.

See Also:

• Constant Field Values

Constructor Details

TransposeDispatcher

private TransposeDispatcher()

Method Details

dispatch

public static Matrix dispatch(Matrix src)

Dispatches a matrix transpose problem to the appropriate algorithm based on its shape and size.

Parameters:

src - Matrix to transpose.

Returns:

The transpose of the source matrix.

dispatch

public static CMatrix dispatch(CMatrix src)

Dispatches a matrix transpose problem to the appropriate algorithm based in its shape and size.

Parameters:

src - Matrix to transpose.

Returns:

The transpose of the source matrix.

dispatchHermitian

public static CMatrix dispatchHermitian(CMatrix src)

Dispatches a matrix hermitian transpose problem to the appropriate algorithm based on its shape and size.

Parameters:

src - Matrix to transpose.

Returns:

The transpose of the source matrix.

dispatchTensor

public static Tensor dispatchTensor(Tensor src,
 int axis1,
 int axis2)

Dispatches a tensor transpose problem to the appropriate algorithm based on its shape and size.

Parameters:

src - Tensor to transpose.

axis1 - First axis in tensor transpose.

axis2 - Second axis in tensor transpose.

Returns:

The result of the tensor transpose.

Throws:

ArrayIndexOutOfBoundsException - If either axis is not within the src tensor.

dispatchTensor

public static double[] dispatchTensor(double[] src,
 Shape shape,
 int[] axes)

Dispatches a tensor transpose problem to the appropriate algorithm based on its shape and size.

Parameters:

src - Entries of tensor to transpose.

shape - Shape of the tensor to transpose.

axes - Permutation of axes in the tensor transpose.

Returns:

The result of the tensor transpose.

Throws:

ArrayIndexOutOfBoundsException - If either axis is not within the src tensor.

dispatchTensor

public static CTensor dispatchTensor(CTensor src,
 int axis1,
 int axis2)

Dispatches a tensor transpose problem to the appropriate algorithm based on its shape and size.

Parameters:

src - Tensor to transpose.

axis1 - First axis in tensor transpose.

axis2 - Second axis in tensor transpose.

Returns:

chooseAlgorithmTensor

private static TransposeDispatcher.Algorithm chooseAlgorithmTensor(int length1,
 int length2)

Chooses the appropriate algorithm for computing a tensor transpose.

Parameters:

length1 - Length of first axis in tensor transpose.

length2 - Length of second axis in tensor transpose.

Returns:

chooseAlgorithmTensor

private static TransposeDispatcher.Algorithm chooseAlgorithmTensor(int numEntries)

Chooses the appropriate algorithm for computing a tensor transpose.

Parameters:

numEntries - Total number of entries in tensor to transpose.

Returns:

The algorithm to use for the tensor transpose.

chooseAlgorithm

private static TransposeDispatcher.Algorithm chooseAlgorithm(Shape shape)

Chooses the appropriate matrix transpose algorithm based on the shape of a matrix.

Parameters:

shape - Shape of matrix to transpose.

Returns:

The appropriate matrix transpose algorithm.

chooseAlgorithmHermitian

private static TransposeDispatcher.Algorithm chooseAlgorithmHermitian(Shape shape)

Chooses the appropriate matrix hermitian transpose algorithm based on the shape of a matrix.

Parameters:

shape - Shape of matrix to transpose.

Returns:

The appropriate matrix transpose algorithm.

chooseAlgorithmComplex

private static TransposeDispatcher.Algorithm chooseAlgorithmComplex(Shape shape)

Chooses the appropriate matrix transpose algorithm based on the shape of a matrix.

Parameters:

shape - Shape of matrix to transpose.

Returns:

The appropriate matrix transpose algorithm.