org.apfloat.spi
Interface ConvolutionStrategy

All Known Implementing Classes:

Double3NTTConvolutionStrategy, DoubleKaratsubaConvolutionStrategy, DoubleMediumConvolutionStrategy, DoubleShortConvolutionStrategy, Float3NTTConvolutionStrategy, FloatKaratsubaConvolutionStrategy, FloatMediumConvolutionStrategy, FloatShortConvolutionStrategy, Int3NTTConvolutionStrategy, IntKaratsubaConvolutionStrategy, IntMediumConvolutionStrategy, IntShortConvolutionStrategy, Long3NTTConvolutionStrategy, LongKaratsubaConvolutionStrategy, LongMediumConvolutionStrategy, LongShortConvolutionStrategy

public interface ConvolutionStrategy

Generic convolution strategy. To perform the convolution, an implementing class could use e.g.

• A simple long multiplication convolution with O(n²) complexity
• An O(n^log2(3)) Karatsuba type algorithm, e.g. as desribed in Knuth's Seminumerical Algorithms
• Floating-point Fast Fourier Transform (FFT) based convolution
• Number-Theoretic Transform (NTT) based convolution, with the Chinese Remainder Theorem used

Version:

1.0

Author:

Mikko Tommila

Method Summary

DataStorage

convolute(DataStorage x, DataStorage y, long resultSize) Convolutes the two sets of data.

Method Detail

convolute

DataStorage convolute(DataStorage x,
                      DataStorage y,
                      long resultSize)
                      throws ApfloatRuntimeException

Convolutes the two sets of data.

Parameters:

x - First data set.

y - Second data set.

resultSize - Number of elements needed in the result data.

Returns:

The convolved data.

Throws:

ApfloatRuntimeException