A flexible framework for multidimensional DFTs
From MaRDI portal
Abstract: Multi-dimensional discrete Fourier transforms (DFT) are typically decomposed into multiple 1D transforms. Hence, parallel implementations of any multi-dimensional DFT focus on parallelizing within or across the 1D DFT. Existing DFT packages exploit the inherent parallelism across the 1D DFTs and offer rigid frameworks, that cannot be extended to incorporate both forms of parallelism and various data layouts to enable some of the parallelism. However, in the era of exascale, where systems have thousand of nodes and intricate network topologies, flexibility and parallel efficiency are key aspects all multi-dimensional DFT frameworks need to have in order to map and scale the computation appropriately. In this work, we present a flexible framework, built on the Redistribution Operations and Tensor Expressions (ROTE) framework, that facilitates the development of a family of parallel multi-dimensional DFT algorithms by 1) unifying the two parallelization schemes within a single framework, 2) exploiting the two different parallelization schemes to different degrees and 3) using different data layouts to distribute the data across the compute nodes. We demonstrate the need of a versatile framework and thus a need for a family of parallel multi-dimensional DFT algorithms on the K-Computer, where we show almost linear strong scaling results for problem sizes of 1024^3 on 32k compute nodes.
Recommendations
- P3DFFT: A framework for parallel computations of Fourier transforms in three dimensions
- PFFT: An extension of FFTW to massively parallel architectures
- A decomposition method with minimum communication amount for parallelization of multi-dimensional FFTs
- A PARALLEL FAST FOURIER TRANSFORM
- scientific article; zbMATH DE number 991433
Cites work
- \texttt{NWChem}: a comprehensive and scalable open-source solution for large scale molecular simulations
- A domain decomposition method for pseudo-spectral electromagnetic simulations of plasmas
- A Rewriting System for the Vectorization of Signal Transforms
- A simple and efficient parallel FFT algorithm using the BSP model
- An Algorithm for the Machine Calculation of Complex Fourier Series
- Elemental, a new framework for distributed memory dense matrix computations
- FFTs and three-dimensional Poisson solvers for hypercubes
- High performance computational chemistry: An overview of NWChem a distributed parallel application
- scientific article; zbMATH DE number 53687 (Why is no real title available?)
- Modeling the viscoplastic micromechanical response of two-phase materials using fast Fourier transforms
- Multiprocessor FFTs
- P3DFFT: A framework for parallel computations of Fourier transforms in three dimensions
- Parallel Algorithms for the Spectral Transform Method
- Parallel empirical pseudopotential electronic structure calculations for million atom systems
- Parallel implementation of 3D FFT with volumetric decomposition schemes for efficient molecular dynamics simulations
Cited in
(4)
This page was built for publication: A flexible framework for multidimensional DFTs
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5131983)
