GPU-based parallel integration of large numbers of independent ODE systems

DOI10.1007/978-3-319-06548-9_8zbMATH Open1317.65157OpenAlexW3102910020MaRDI QIDQ5261402FDOQ5261402

Authors Kyle E. Niemeyer, Chih-Jen Sung 

Publication date 3 July 2015

Published in Numerical Computations with GPUs (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/1611.02274

zbMATH Keywords

parallel computationstiff ordinary differential equationnonstiff systembackwards differentiation formulaefifth-order adaptive Runge-Kutta-Cash-Karp algorithmRunge-Kutta-Chebyshev algorithm

Mathematics Subject Classification ID

Parallel numerical computation (65Y05) Nonlinear ordinary differential equations and systems (34A34) Numerical methods for stiff equations (65L04) Multistep, Runge-Kutta and extrapolation methods for ordinary differential equations (65L06) Numerical methods for initial value problems involving ordinary differential equations (65L05)

Abstract: The task of integrating a large number of independent ODE systems arises in various scientific and engineering areas. For nonstiff systems, common explicit integration algorithms can be used on GPUs, where individual GPU threads concurrently integrate independent ODEs with different initial conditions or parameters. One example is the fifth-order adaptive Runge-Kutta-Cash-Karp (RKCK) algorithm. In the case of stiff ODEs, standard explicit algorithms require impractically small time-step sizes for stability reasons, and implicit algorithms are therefore commonly used instead to allow larger time steps and reduce the computational expense. However, typical high-order implicit algorithms based on backwards differentiation formulae (e.g., VODE, LSODE) involve complex logical flow that causes severe thread divergence when implemented on GPUs, limiting the performance. Therefore, alternate algorithms are needed. A GPU-based Runge-Kutta-Chebyshev (RKC) algorithm can handle moderate levels of stiffness and performs significantly faster than not only an equivalent CPU version but also a CPU-based implicit algorithm (VODE) based on results shown in the literature. In this chapter, we present the mathematical background, implementation details, and source code for the RKCK and RKC algorithms for use integrating large numbers of independent systems of ODEs on GPUs. In addition, brief performance comparisons are shown for each algorithm, demonstrating the potential benefit of moving to GPU-based ODE integrators.

Recommendations

• The art of solving a large number of non-stiff, low-dimensional ordinary differential equation systems on GPUs and CPUs
• Massive parallel implementation of ODE solvers.
• Program package MPGOS: challenges and solutions during the integration of a large number of independent ODE systems using GPUs
• Exploiting limited access distance of ODE systems for parallelism and locality in explicit methods
• Solving ordinary differential equations on GPUs

Cited in

(9)

• Program package MPGOS: challenges and solutions during the integration of a large number of independent ODE systems using GPUs
• Exploiting limited access distance of ODE systems for parallelism and locality in explicit methods
• Automated translation and accelerated solving of differential equations on multiple GPU platforms
• GPU-acceleration of waveform relaxation methods for large differential systems
• Massive parallel implementation of ODE solvers.
• OpenCL parallel integration of ordinary differential equations: applications in computational dynamics
• Solving ordinary differential equations on GPUs
• Solving random ordinary differential equations on GPU clusters using multiple levels of parallelism
• The art of solving a large number of non-stiff, low-dimensional ordinary differential equation systems on GPUs and CPUs

Describes a project that uses

Uses Software

• RKC
• RKCK

This page was built for publication: GPU-based parallel integration of large numbers of independent ODE systems

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5261402)