Massively parallel landscape-evolution modelling using general purpose graphical processing units

Abstract

Abstract-As our expectations of what computer systems can do and our ability to capture data improves, the desire to perform ever more computationally intensive tasks increases. Often these tasks, comprising vast numbers of repeated computations, are highly interdependent on each other -a closely coupled problem. The process of Landscape-Evolution Modelling is an example of such a problem. In order to produce realistic models it is necessary to process landscapes containing millions of data points over time periods extending up to millions of years. This leads to non-tractable execution times, often in the order of years. Researchers therefore seek multiple orders of magnitude reduction in the execution time of these models. The massively parallel programming environment offered through General Purpose Graphical Processing Units offers the potential for multiple orders of magnitude speedup in code execution times.In this paper we demonstrate how the time dominant parts of a Landscape-Evolution Model can be recoded for a massively parallel architecture providing two orders of magnitude reduction in execution time.

Two decades of numerical modelling to understand long term fluvial archives: Advances and future perspectives

“…It does however simulate local fan morphology realistically but again without simulating detailed sedimentological patterns. There are also attempts underway to use the more detailed, but also more parameter/input demanding Wainwright (2006) model for larger spatio-temporal scales using parallel processing (PARALLEM;McGough et al, 2012). Unfortunately these attempts have not yielded realistic landscapes yet.…”

“…This is representative of the trade-off between complexity and numerical simulation feasibility, which is one underlying reason that no model is able to simulate detailed realistic landscapes over long time spans. A technical approach to reduce run time is currently sought in parallel processing (McGough et al, 2012) and has been developed for versions of the CAESAR-Lisflood model. However, this technique also requires a complete recoding of existing models, discouraging its widespread adoption.…”