High - performance computing and geography : developments , issues , and case studies

In this paper we outline some of the results that were obtained by the application of a Cray T3D parallel supercomputer to human geography problems. We emphasise the fundamental importance of high-performance computing (HPC) as a future relevant paradigm for doing geography. We offer an introduction to recent developments and illustrate how new computational intelligence technologies can start to be used to make use of opportunities created by data riches from geographic information systems, artificial intelligence tools, and HPC in geography. Research Council's High Performance Computing Initiative (HPCI) in 1994. This consortium had three main aims: (1) to port a selection of existing vector and serial geography codes to the 512-processor Cray T3D; (2) to preserve the programming and porting effort by conforming to programming standards [Message Passing Interface (MPI) and Fortran 77/90]; (3) to use these codes to carry out new science that had previously been considered computationally infeasible; and (4) to use the results and the experience to raise awareness of high-performance computing (HPC) in the geography community, particularly to illustrate what is now possible in the rapidly developing field of parallel supercomputing. This fourth aim is the purpose of this paper. The world of computing underwent a quiet revolution of profound long-term significance in the early 1990s. Hillis (1992) argues that it was then in the throes of a major technological change characterised by the development of highly parallel super-computing hardware which was about to change significantly how science is done. This process is now complete. Faster supercomputers have stimulated new ways of doing science in areas that are just too complex to be handled by any other means or where there is need to analyse large volumes of data or where there is need for real-time analysis and modelling. Computer-based experimentation and simulation is increasingly being regarded as a cost-effective and very useful means of creating new knowledge. Computation has become a scientific tool of equal importance to theory and experimentation. HPC is also widely acknowledged as one of the key information technologies of the future. But so far HPC has had a minimal impact on geography and most of the social sciences. In this paper we demonstrate by reference to four case studies that HPC could also be extremely relevant to many areas of geography with a potential that goes far beyond the narrow domain of quantitative geography and geographic information systems (GIS). It can …