Benjamin Ong

Yingda Cheng, Andrew J. Christlieb, Wei Guo et al.

In plasma simulations, where the speed of light divided by a characteristic length is at a much higher frequency than other relevant parameters in the underlying system, such as the plasma frequency, implicit methods begin to play an important role in generating efficient solutions in these multi-scale problems. Under conditions of scale separation, one can rescale Maxwell's equations in such a way as to give a magneto static limit known as the Darwin approximation of electromagnetics. In this work, we present a new approach to solve Maxwell's equations based on a Method of Lines Transpose (MOL$^T$) formulation, combined with a fast summation method with computational complexity $O(N\log{N})$, where $N$ is the number of grid points (particles). Under appropriate scaling, we show that the proposed schemes result in asymptotic preserving methods that can recover the Darwin limit of electrodynamics.

Benjamin Ong, Andrew Melfi, Andrew Christlieb

In this paper, we further develop a family of parallel time integrators known as Revisionist Integral Deferred Correction methods (RIDC) to allow for the semi-implicit solution of time dependent PDEs. Additionally, we show that our semi-implicit RIDC algorithm can harness the computational potential of multiple general purpose graphical processing units (GPUs) in a single node by utilizing existing CUBLAS libraries for matrix linear algebra routines in our implementation. In the numerical experiments, we show that our implementation computes a fourth order solution using four GPUs and four CPUs in approximately the same wall clock time as a first order solution computed using a single GPU and a single CPU.

Benjamin Ong, Andrew Christlieb, Bryan Quaife

In this paper, a new two-parameter family of regularized kernels is introduced, suitable for applying high-order time stepping to N-body systems. These high-order kernels are derived by truncating a Taylor expansion of the non-regularized kernel about $(r^2+ε^2)$, generating a sequence of increasingly more accurate kernels. This paper proves the validity of this two-parameter family of regularized kernels, constructs error estimates, and illustrates the benefits of using high-order kernels through numerical experiments.