Accelerating solutions of one-dimensional unsteady PDEs with GPU-based swept time-space decomposition

DOI10.1016/J.JCP.2017.12.028zbMATH Open1422.65474arXiv1705.03162OpenAlexW2768347146MaRDI QIDQ1699485FDOQ1699485

Daniel J. Magee, Kyle E. Niemeyer

Publication date: 23 February 2018

Published in: Journal of Computational Physics (Search for Journal in Brave)

Abstract: The expedient design of precision components in aerospace and other high-tech industries requires simulations of physical phenomena often described by partial differential equations (PDEs) without exact solutions. Modern design problems require simulations with a level of resolution difficult to achieve in reasonable amounts of time---even in effectively parallelized solvers. Though the scale of the problem relative to available computing power is the greatest impediment to accelerating these applications, significant performance gains can be achieved through careful attention to the details of memory communication and access. The swept time-space decomposition rule reduces communication between sub-domains by exhausting the domain of influence before communicating boundary values. Here we present a GPU implementation of the swept rule, which modifies the algorithm for improved performance on this processing architecture by prioritizing use of private (shared) memory, avoiding interblock communication, and overwriting unnecessary values. It shows significant improvement in the execution time of finite-difference solvers for one-dimensional unsteady PDEs, producing speedups of 2--9$imes$ for a range of problem sizes, respectively, compared with simple GPU versions and 7--300$imes$ compared with parallel CPU versions. However, for a more sophisticated one-dimensional system of equations discretized with a second-order finite-volume scheme, the swept rule performs 1.2--1.9$imes$ worse than a standard implementation for all problem sizes.

Full work available at URL: https://arxiv.org/abs/1705.03162

zbMATH Keywords

domain decompositionpartial differential equationscomputational fluid dynamicshigh-performance computingGPU computingcommunication-avoiding algorithms

Mathematics Subject Classification ID

Parallel numerical computation (65Y05) PDEs in connection with fluid mechanics (35Q35) Multigrid methods; domain decomposition for initial value and initial-boundary value problems involving PDEs (65M55) Numerical algorithms for specific classes of architectures (65Y10)

Cites Work

• PyFR: an open source framework for solving advection-diffusion type problems on streaming architectures using the flux reconstruction approach
• 50 Years of Time Parallel Time Integration
• Parallel Time Integration with Multigrid
• The swept rule for breaking the latency barrier in time advancing PDEs
• Multicore-Optimized Wavefront Diamond Blocking for Optimizing Stencil Updates

Cited In (2)

• An efficient GPU-based fractional-step domain decomposition scheme for the reaction-diffusion equation
• A new GPU-based corrected explicit-implicit domain decomposition scheme for convection-dominated diffusion problems

Uses Software

• CUDA
• PyFR
• KSIDSwept

This page was built for publication: Accelerating solutions of one-dimensional unsteady PDEs with GPU-based swept time-space decomposition

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1699485)