Package org.apfloat.aparapi
package org.apfloat.aparapi
Transforms for using a GPU via the aparapi library.
This package contains Number-Theoretic Transform implementations that can use the GPU (Graphics Processing Unit) for executing the transforms. There is significant overhead in invoking the GPU, e.g. in transferring the data between the main memory and the GPU memory, so for small data sets there is usually no performance improvement, and in many cases performance can be even significantly slower. However for very large calculations, e.g. one billion digits, using the GPU can improve the performance noticeably, depending on the hardware used.
-
ClassDescriptionSix-step NTT implementation that processes the data in the columns of the matrix.Two-pass NTT implementation that processes the data in the columns of the matrix.Builder factory for aparapi transform implementations for the
int
element type.Steps for the factor-3 NTT using the GPU, for theint
element type.Factor-3 NTT implementation for theint
element type.Matrix transposition in the GPU for theint
type.NTT Builder for aparapi transform implementations for theint
element type.NTT steps for theint
element type aparapi transforms.Six-step NTT implementation for theint
element type.Builder factory for aparapi transform implementations for thelong
element type.Steps for the factor-3 NTT using the GPU, for thelong
element type.Factor-3 NTT implementation for thelong
element type.Matrix transposition in the GPU for thelong
type.NTT Builder for aparapi transform implementations for thelong
element type.NTT steps for thelong
element type aparapi transforms.Six-step NTT implementation for thelong
element type.