-
Interweaving PFASST and Parallel Multigrid
Authors:
Michael Minion,
Robert Speck,
Matthias Bolten,
Matthew Emmett,
Daniel Ruprecht
Abstract:
The parallel full approximation scheme in space and time (PFASST) introduced by Emmett and Minion in 2012 is an iterative strategy for the temporal parallelization of ODEs and discretized PDEs. As the name suggests, PFASST is similar in spirit to a space-time FAS multigrid method performed over multiple time-steps in parallel. However, since the original focus of PFASST has been on the performance…
▽ More
The parallel full approximation scheme in space and time (PFASST) introduced by Emmett and Minion in 2012 is an iterative strategy for the temporal parallelization of ODEs and discretized PDEs. As the name suggests, PFASST is similar in spirit to a space-time FAS multigrid method performed over multiple time-steps in parallel. However, since the original focus of PFASST has been on the performance of the method in terms of time parallelism, the solution of any spatial system arising from the use of implicit or semi-implicit temporal methods within PFASST have simply been assumed to be solved to some desired accuracy completely at each sub-step and each iteration by some unspecified procedure. It hence is natural to investigate how iterative solvers in the spatial dimensions can be interwoven with the PFASST iterations and whether this strategy leads to a more efficient overall approach. This paper presents an initial investigation on the relative performance of different strategies for coupling PFASST iterations with multigrid methods for the implicit treatment of diffusion terms in PDEs. In particular, we compare full accuracy multigrid solves at each sub-step with a small fixed number of multigrid V-cycles. This reduces the cost of each PFASST iteration at the possible expense of a corresponding increase in the number of PFASST iterations needed for convergence. Parallel efficiency of the resulting methods is explored through numerical examples.
△ Less
Submitted 30 March, 2015; v1 submitted 24 July, 2014;
originally announced July 2014.
-
High order schemes based on operator splitting and deferred corrections for stiff time dependent PDEs
Authors:
Max Duarte,
Matthew Emmett
Abstract:
We consider quadrature formulas of high order in time based on Radau-type, L-stable implicit Runge-Kutta schemes to solve time dependent stiff PDEs. Instead of solving a large nonlinear system of equations, we develop a method that performs iterative deferred corrections to compute the solution at the collocation nodes of the quadrature formulas. The numerical stability is guaranteed by a dedicate…
▽ More
We consider quadrature formulas of high order in time based on Radau-type, L-stable implicit Runge-Kutta schemes to solve time dependent stiff PDEs. Instead of solving a large nonlinear system of equations, we develop a method that performs iterative deferred corrections to compute the solution at the collocation nodes of the quadrature formulas. The numerical stability is guaranteed by a dedicated operator splitting technique that efficiently handles the stiffness of the PDEs and provides initial and intermediate solutions to the iterative scheme. In this way the low order approximations computed by a tailored splitting solver of low algorithmic complexity are iteratively corrected to obtain a high order solution based on a quadrature formula. The mathematical analysis of the numerical errors and local order of the method is carried out in a finite dimensional framework for a general semi-discrete problem, and a time-stepping strategy is conceived to control numerical errors related to the time integration. Numerical evidence confirms the theoretical findings and assesses the performance of the method in the case of a stiff reaction-diffusion equation.
△ Less
Submitted 1 April, 2016; v1 submitted 1 July, 2014;
originally announced July 2014.
-
Inexact spectral deferred corrections
Authors:
Robert Speck,
Daniel Ruprecht,
Michael Minion,
Matthew Emmett,
Rolf Krause
Abstract:
Spectral deferred correction (SDC) methods are an attractive approach to iteratively computing collocation solutions to an ODE by performing so-called sweeps with a low-order time stepping method. SDC allows to easily construct high order split methods where e.g. stiff terms of the ODE are treated implicitly. This requires the solution to full accuracy of multiple linear systems of equations durin…
▽ More
Spectral deferred correction (SDC) methods are an attractive approach to iteratively computing collocation solutions to an ODE by performing so-called sweeps with a low-order time stepping method. SDC allows to easily construct high order split methods where e.g. stiff terms of the ODE are treated implicitly. This requires the solution to full accuracy of multiple linear systems of equations during each sweep, e.g. with a multigrid method. In this paper, we present an inexact variant of SDC, where each solve of a linear system is replaced by a single multigrid V-cycle and thus significantly reduces the cost for each sweep. For the investigated examples, this strategy results only in a small increase of the number of required sweeps and we demonstrate that "inexact spectral deferred corrections" can provide a dramatic reduction of the overall number of multigrid V-cycles required to complete an SDC time step.
△ Less
Submitted 12 June, 2014; v1 submitted 30 January, 2014;
originally announced January 2014.
-
High-Order Algorithms for Compressible Reacting Flow with Complex Chemistry
Authors:
M. Emmett,
W. Zhang,
J. B. Bell
Abstract:
In this paper we describe a numerical algorithm for integrating the multicomponent, reacting, compressible Navier-Stokes equations, targeted for direct numerical simulation of combustion phenomena. The algorithm addresses two shortcomings of previous methods. First, it incorporates an eighth-order narrow stencil approximation of diffusive terms that reduces the communication compared to existing m…
▽ More
In this paper we describe a numerical algorithm for integrating the multicomponent, reacting, compressible Navier-Stokes equations, targeted for direct numerical simulation of combustion phenomena. The algorithm addresses two shortcomings of previous methods. First, it incorporates an eighth-order narrow stencil approximation of diffusive terms that reduces the communication compared to existing methods and removes the need to use a filtering algorithm to remove Nyquist frequency oscillations that are not damped with traditional approaches. The methodology also incorporates a multirate temporal integration strategy that provides an efficient mechanism for treating chemical mechanisms that are stiff relative to fluid dynamical time scales. The overall methodology is eighth order in space with option for fourth order to eighth order in time. The implementation uses a hybrid programming model designed for effective utilization of many-core architectures. We present numerical results demonstrating the convergence properties of the algorithm with realistic chemical kinetics and illustrating its performance characteristics. We also present a validation example showing that the algorithm matches detailed results obtained with an established low Mach number solver.
△ Less
Submitted 21 April, 2014; v1 submitted 27 September, 2013;
originally announced September 2013.
-
A space-time parallel solver for the three-dimensional heat equation
Authors:
Robert Speck,
Daniel Ruprecht,
Matthew Emmett,
Matthias Bolten,
Rolf Krause
Abstract:
The paper presents a combination of the time-parallel "parallel full approximation scheme in space and time" (PFASST) with a parallel multigrid method (PMG) in space, resulting in a mesh-based solver for the three-dimensional heat equation with a uniquely high degree of efficient concurrency. Parallel scaling tests are reported on the Cray XE6 machine "Monte Rosa" on up to 16,384 cores and on the…
▽ More
The paper presents a combination of the time-parallel "parallel full approximation scheme in space and time" (PFASST) with a parallel multigrid method (PMG) in space, resulting in a mesh-based solver for the three-dimensional heat equation with a uniquely high degree of efficient concurrency. Parallel scaling tests are reported on the Cray XE6 machine "Monte Rosa" on up to 16,384 cores and on the IBM Blue Gene/Q system "JUQUEEN" on up to 65,536 cores. The efficacy of the combined spatial- and temporal parallelization is shown by demonstrating that using PFASST in addition to PMG significantly extends the strong-scaling limit. Implications of using spatial coarsening strategies in PFASST's multi-level hierarchy in large-scale parallel simulations are discussed.
△ Less
Submitted 14 July, 2014; v1 submitted 30 July, 2013;
originally announced July 2013.
-
A multi-level spectral deferred correction method
Authors:
Robert Speck,
Daniel Ruprecht,
Matthew Emmett,
Michael Minion,
Matthias Bolten,
Rolf Krause
Abstract:
The spectral deferred correction (SDC) method is an iterative scheme for computing a higher-order collocation solution to an ODE by performing a series of correction sweeps using a low-order timestepping method. This paper examines a variation of SDC for the temporal integration of PDEs called multi-level spectral deferred corrections (MLSDC), where sweeps are performed on a hierarchy of levels an…
▽ More
The spectral deferred correction (SDC) method is an iterative scheme for computing a higher-order collocation solution to an ODE by performing a series of correction sweeps using a low-order timestepping method. This paper examines a variation of SDC for the temporal integration of PDEs called multi-level spectral deferred corrections (MLSDC), where sweeps are performed on a hierarchy of levels and an FAS correction term, as in nonlinear multigrid methods, couples solutions on different levels. Three different strategies to reduce the computational cost of correction sweeps on the coarser levels are examined: reducing the degrees of freedom, reducing the order of the spatial discretization, and reducing the accuracy when solving linear systems arising in implicit temporal integration. Several numerical examples demonstrate the effect of multi-level coarsening on the convergence and cost of SDC integration. In particular, MLSDC can provide significant savings in compute time compared to SDC for a three-dimensional problem.
△ Less
Submitted 25 August, 2014; v1 submitted 4 July, 2013;
originally announced July 2013.
-
PyClaw: Accessible, Extensible, Scalable Tools for Wave Propagation Problems
Authors:
David I. Ketcheson,
Kyle T. Mandli,
Aron Ahmadia,
Amal Alghamdi,
Manuel Quezada,
Matteo Parsani,
Matthew G. Knepley,
Matthew Emmett
Abstract:
Development of scientific software involves tradeoffs between ease of use, generality, and performance. We describe the design of a general hyperbolic PDE solver that can be operated with the convenience of MATLAB yet achieves efficiency near that of hand-coded Fortran and scales to the largest supercomputers. This is achieved by using Python for most of the code while employing automatically-wrap…
▽ More
Development of scientific software involves tradeoffs between ease of use, generality, and performance. We describe the design of a general hyperbolic PDE solver that can be operated with the convenience of MATLAB yet achieves efficiency near that of hand-coded Fortran and scales to the largest supercomputers. This is achieved by using Python for most of the code while employing automatically-wrapped Fortran kernels for computationally intensive routines, and using Python bindings to interface with a parallel computing library and other numerical packages. The software described here is PyClaw, a Python-based structured grid solver for general systems of hyperbolic PDEs \cite{pyclaw}. PyClaw provides a powerful and intuitive interface to the algorithms of the existing Fortran codes Clawpack and SharpClaw, simplifying code development and use while providing massive parallelism and scalable solvers via the PETSc library. The package is further augmented by use of PyWENO for generation of efficient high-order weighted essentially non-oscillatory reconstruction code. The simplicity, capability, and performance of this approach are demonstrated through application to example problems in shallow water flow, compressible flow and elasticity.
△ Less
Submitted 12 May, 2012; v1 submitted 27 November, 2011;
originally announced November 2011.
