CoreImplementationsT Class

public static class CoreImplementations<T>

Public NotInheritable Class CoreImplementations(Of T)

generic<typename T>
public ref class CoreImplementations abstract sealed

[<AbstractClassAttribute>]
[<SealedAttribute>]
type CoreImplementations<'T> =  class end

Type Parameters

T

The CoreImplementationsT type exposes the following members.

	Name	Description
	AutoLoadNativeProviders	Gets or sets a value that indicates whether native (processor specific) implementation providers should be loaded automatically.
	Decompositions	Gets or sets the current implementation of matrix decompositions.
	Fft	Gets or sets the current FFT implementation.
	LinearAlgebra	Gets or sets the current implementation of the Basic Linear Algebra Subroutines (BLAS) for real and complex values.
	NativeProviderBaseName	Gets or sets the base name of the native provider assemblies.
	NativeProviderBaseNameOfSingle	Gets or sets the base name of the single-precision native provider assemblies.
	NativeProviderPath	Gets or sets the path where the assemblies that
	RawLinearAlgebra	Gets or sets the current implementation of the Basic Linear Algebra Subroutines (BLAS) for the current element type.
	SparseLinearAlgebraOperations	Gets or sets the current implementation of the sparse linear algebra routines.
	VectorFunctions	Gets or sets the current implementation of array functions.

	Name	Description
	UseFromAssembly	Instructs the library to use a custom BLAS implementation contained in a .NET assembly.
	UseHybrid	Instructs the library to use the managed implementations.
	UseManaged	Instructs the library to use the managed implementations.
	UseNative	Instructs the library to use the native BLAS implementation.
	UseProvider	Instructs the library to use the specified core provider.

Many calculations rely on a core set of algorithms. Optimized implementations exist for many platforms. The Extreme Optimization Mathematics Library for .NET provides at least two implementations. One is in 100% managed code. The other calls a native, highly optimized, processor-specific kernel based on Intel®'s Math Kernel Library®.

This class exposes properties that allow you to specify the implementations that are to be used by the other classes.

You can select the managed implementation by calling the UseManaged method. To select the native implementation, call the UseNative method. The native kernel is used by default. This is the fastest option, especially for larger matrices, but also has a larger memory footprint.

Alternatively, you can plug in platform optimized versions by inheriting from the implementation base classes, and calling UseProvider(CoreProviderT).

Reference