High-performance computing tools for the integrated assessment and modelling of socialeecological systems q , qq Brett A. Bryan * CSIRO Ecosystem Sciences, Waite Campus, Urrbrae, SA 5064, Australia article info Article history: Received 12 July 2011 Received in revised form 17 January 2012 Accepted 13 February 2012 Available online 10 March 2012 Keywords: Graphics processing unit (GPU) Parallel programming Multi-core Cluster Grid GIS Environmental Concurrency Global challenges abstract Integrated spatio-temporal assessment and modelling of complex socialeecological systems is required to address global environmental challenges. However, the computational demands of this modelling are unlikely to be met by traditional Geographic Information System (GIS) tools anytime soon. I evaluated the potential of a range of high-performance computing (HPC) hardware and software tools to overcome these computational barriers. Performance advantages were quantified using a synthetic model. Four tests were compared, using: a) an Arc Macro Language (AML) GIS script on a single central processing unit (CPU); b) Python/NumPy on 1e256 CPU cores; c) Python/NumPy on 1e64 graphics processing units (GPUs) with high-level PyCUDA abstraction (GPUArray); and d) Python/NumPy on 1e64 GPUs with low- level PyCUDA abstraction (ElementwiseKernel). The GIS implementation effectively took 15.5 weeks to run. Python/NumPy on a single CPU core led to a speed-up of 59 compared to the GIS. On a single GPU, speed-ups of 1473 were achieved using GPUArray and 4881 using ElementwiseKernel. Parallel pro- cessing led to further performance enhancements. At best, the ElementwiseKernel module in parallel over 64 GPUs achieved a speed-up of 63,643. Open source tools such as Python applied across a spectrum of HPC resources offer transformational and accessible performance improvements for integrated assessment and modelling. By reducing the computational barrier, HPC can lead to a step change in modelling sophistication, including the better representation of uncertainty, and perhaps even new modelling paradigms. However, migration to new hardware and software environments also has significant costs. Costs and benefits of HPC are discussed and code tools are provided to help others migrate to HPC and transform our ability to address global challenges through integrated assessment and modelling. Ó 2012 Elsevier Ltd. All rights reserved. 1. Introduction Addressing global environmental challenges (e.g. climate change, food and energy security, natural resource management, and biodiversity conservation) operating within complex socialeecological systems demands the integrated assessment and modelling of biophysical, ecological, economic, and social infor- mation (Costanza, 1996; Parker et al., 2002; Jakeman and Letcher, 2003; Kumar et al., 2006; Bryan et al., 2010a, 2011a; Reid et al., 2010). Many of these processes are heterogeneous over the land- scape (Bryan, 2003) and may display spatial interconnections (e.g. topographic processes, species dispersal, supply chain analysis). Temporal dynamics may also be an important component in these processes (e.g. climate change, tree growth, cash flow) (Venevsky and Maksyutov, 2007). Data sets are often very large due to the increasing need to model processes over large extents (e.g. conti- nental and global scale) whilst maintaining high enough spatial and temporal resolution (e.g. landscape scale, daily time-steps) to adequately capture the relevant dynamics (Harris, 2002). Further, uncertainty and sensitivity are an inherent part of these socialeecological systems which policy-makers need to under- stand in order to make robust decisions (Rotmans and van Asselt, 2001; Kooistra et al., 2005; Bryan, 2010; Cheviron et al., 2010; Lilburne and North, 2010). A common way to quantify sensitivity and uncertainty is by undertaking multiple model simulations using varying input parameter values (Lilburne and Tarantola, 2009). In concert, these characteristics of the integrated Abbreviations: HPC, high-performance computing; CPU, central processing unit; GPU, graphics processing unit; NPV, net present value; AML, arc macro language; GIS, geographic information system. q Computer codes are available in machine-readable form on request from the author. qq Thematic Issue on the Future of Integrated Modeling Science and Technology. * Tel.: þ61 8 8303 8581; fax: þ61 8 8303 8582. E-mail address: brett.bryan@csiro.au. Contents lists available at SciVerse ScienceDirect Environmental Modelling & Software journal homepage: www.elsevier.com/locate/envsoft 1364-8152/$ e see front matter Ó 2012 Elsevier Ltd. All rights reserved. doi:10.1016/j.envsoft.2012.02.006 Environmental Modelling & Software 39 (2013) 295e303