Computational science is a rapidly growing interdisciplinary field. There are many problems in science and technology that cannot be sufficiently studied experimentally or theoretically. It may be too expensive or too dangerous, or simply impossible due to the space and timescales involved.
In fact, computational science is considered by many to be a third methodology in scientific research, along with theory and experiment, and working in tandem with them. Computational science can be used to corroborate theories that cannot be confirmed or denied experimentally, for example theories relating to the creation of the universe. On the other hand, advances in experimental techniques and the resulting data explosion, allow for data-driven modelling and simulation.
You should not confuse computational science, which uses computational methods to deal with scientific problems, and computer science, which focuses on the computer itself. Having said that, computational science draws upon computer science, as well as upon mathematics and applied sciences. Computational science typically consists of three main components:
1) algorithms and models
2) software developed to solve problems and
3) the computer and information infrastructure e.g. hardware, networking and data management components.
Clearly, computational science is an intersection between mathematics, applied disciplines and computer science.
Some of the disciplines traditionally associated with computational science include: atmospheric sciences (e.g. weather forecasting, climate and ocean modelling, seismology etc.), astrophysics, nuclear engineering, chemistry, fluid dynamics, structural analysis and material research. It’s easy to see why these disciplines were quick to take up computational science.
However, more recently, other disciplines, such as medicine (e.g. medical imaging, blood flow simulations, bone structure simulations), biology (e.g. ecosystem and environmental modelling) and economics became part of computational science as well. It is quite common nowadays to come across terms like computational economics or computational biology.
If you want to know what sorts of applications are run on the UK’s national supercomputer ARCHER, visit the relevant page on the ARCHER website. Is there anything that surprises you there?
© EPCC at The University of Edinburgh