Enabling Computational Dynamics in Distributed Computing Environments Using a Heterogeneous Computing Template

摘要

This paper describes a software infrastructure made up of tools and libraries designed to assist developers in implementing computational dynamics applications running on heterogeneous and distributed computing environments. Together, these tools and libraries compose a so called Heterogeneous Computing Template (HCT). The heterogeneous and distributed computing hardware infrastructure is assumed herein to be made up of a combination of CPUs and Graphics Processing Units (GPUs). The computational dynamics applications targeted to execute on such a hardware topology include many-body dynamics, smoothed-particle hydrodynamics (SPH) fluid simulation, and fluid-solid interaction analysis. The underlying theme of the solution approach embraced by HCT is that of partitioning the domain of interest into a number of subdomains that are each managed by a separate core/accelerator (CPU/GPU) pair.Five components at the core of HCT enable the envisioned distributed computing approach to large-scale dynamical system simulation: (a) the ability to partition the problem according to the one-toone mapping; i.e., spatial subdivision, discussed above (pre-processing); (b) a protocol for passing data between any two co-processors; (c) algorithms for element proximity computation; (d) numerical methods for solving the specific dynamics problem of interest; and (e) the ability to carry out post-processing in a distributed fashion.In this contribution the components (a) and (b) of the HCT are demonstrated via the example of the Discrete Element Method (DEM) for rigid body dynamics with friction and contact. The collision detection task required in frictionalcontact dynamics (task (c) above), is shown to benefit on the GPU of a two order of magnitude gain in efficiency when compared to traditional sequential implementations.