Geoscience Reference
In-Depth Information
3 Parallel Computing
in Geography
Muhammed Adnan, Paul A. Longley,
Alex D. Singleton and Ian Turton
CONTENTS
Abstract ............................................................................................................................................ 49
3.1 Introduction ............................................................................................................................ 49
3.2 Types of Parallel Computing .................................................................................................. 51
3.3 Short History of Parallel Computing ...................................................................................... 52
3.4 Parallel Computing and Geography ....................................................................................... 53
3.5 When Not to Use Parallel Computing .................................................................................... 55
3.6 When to Use Parallel Computing and How ............................................................................ 55
3.7 GPGPU Geodemographic Information System ...................................................................... 57
3.8 Towards Real-Time GeoComputation of Geodemographics Using GPU .............................. 59
3.9 Conclusions ............................................................................................................................. 63
References ........................................................................................................................................64
ABSTRACT
Parallel computing has been around for more than 20 years. Yet geography and GIScience have not
employed this technology widely or to its full potential. One reason for this may be due to a per-
ceived lack of access to parallel computing resources in social science departments despite the fact
that dual-core and multicore processors, and hyper-threading technology generating virtual cores,
have been standard equipment for a number of years. However, this may change in the near future
with improved accessibility to technologies such as graphics processing units (GPUs). This chapter
provides an overview of parallel computing including a presentation of different types of parallel
processing followed by a brief history of the field. The chapter then attempts to set out when paral-
lel computing should be used and how, ending with an example of the use of general-purpose GPU
for the development of geodemographic classifications. In an era of government initiatives on open
data, the rise of big spatial data and pressing global geographical challenges to solve, greater adop-
tion of parallel computing technologies could deliver numerous gains.
3.1 INTRODUCTION
In the IT world, sequential computing is the use of a single processing unit to perform a single
or multiple tasks, and this has historically been the standard mode of computation. In contrast,
parallel computing (or parallel processing*) makes use of, at the same time, more than one
central processing unit (CPU) in order to allow users to complete lengthy computational tasks
more quickly. This is achieved by taking the problem and dividing it into smaller tasks, which
are then solved simultaneously (or in parallel) via these different CPUs. Although parallelism
* An older expression that is frequently used as a synonym for parallel computing.
49
 
Search WWH ::




Custom Search