Any reader of Physics in Canada is well aware of the incredible advances in computing and communications that have occurred during our lifetimes, and how they have revolutionized, not only science, but our daily lives as well. There are so many examples that illustrate. In less than 40 years, our common tools have changed from slide rules to massively parallel computer systems, connected across the country by networks with unimaginable capacity and speeds. In barely 20 years, memory costs have gone from roughly $1,000 for a 20 MB disk to $10 for a 2 GB stick, a reduction in per unit cost of a factor of 10,000, or even more if you choose to incorporate general inflation in the calculation. We are all familiar with these kinds of numbers, but it is still worth pausing from time to time to reflect on just how incredible they are -- and to speculate on how far and fast these trends will continue.
Our interest in this issue of Physics in Canada is in the realm of the big systems, or high performance computing (HPC) as it has come to be known in recent years. Our primary goals are to summarize the computing systems and technologies that are available to us now, and to illustrate some of the breadth of the physics that is being done with them. Our additional goals are to describe the Canadian journey in HPC, with emphasis on the process and achievements of the past decade, and point potential users to the access points for the resources that are now available.
Our issue opens with an overview of HPC technologies, their components, and how to choose different ones for different applications. It includes a description of grid services used to coordinate access to computers distributed around the world. Our issue closes with the story of HPC in Canada to date, how our various resources are organized and available nationally, and the emergence of a new era and a new organization, Compute Canada. This story has been one of, probably unparalleled, cooperation amongst interested scientists across the country, and tremendous support from organizations including NSERC, the National Research Council of Canada, the Canada Foundation for Innovation, CANARIE, provincial funding organizations, and numerous computer manufacturers who have provided generously of both financial support and time. Above all else, the key to success has been the cooperation across all these sectors.
Between these two "bookends" of this issue, we have a collection of five papers illustrating the techniques of HPC, and applications to different sciences. They vary in both style and content, but they have important common and complementary features. They are all intended to appeal to a broad audience of physicists with different areas of expertise. They range from the enormous scales of astrophysics, to the minute scales of nuclei. They cover condensed matter physics, equilibrium and non-equilibrium physics, the finite-element analysis of engineering, astrophysics, nuclear physics, and the ever-improving climate modeling, in particular the Canadian Regional Climate Model. Some emphasize technique, some emphasize results, and some illustrate the place of simulations within the broader spectrum from "theory" to "experiment".
A major goal of all academic work is not just to predict our future as a planet and society, but to control our future. Computers play a major role in this enterprise because they enable an ever more detailed modeling of the physical world in which we find ourselves, and a cataloguing of its rich diversity (such as the human genome). Recent advances are analogous suddenly to having at our disposal a microscope or telescope that is orders of magnitude more powerful than what was previously available, and so we can see further and do more to control our own destiny than was previously possible. The Canadian Regional Climate Model is an obvious example of this, but all the articles illustrate in one way or another how computers are playing an increasingly important role as an extension of our own senses to monitor and control the world in which we live. The pace of development will only quicken in the future, and the Canadian academic community has an important role to play at the leading edge of these developments.
We hope that you find this collection of articles enjoyable and informative. For those of you who have not been part of the exciting Canadian HPC story, we invite you to reflect on what a diverse community can accomplish when it works tirelessly and cooperatively towards a common cause.
Mark Whitmore, University of Manitoba
Gordon Drake, University of Windsor
Guest Editors, Physics in Canada
We invite our readers to submit comments on the contents of this editorial.
The Editorial Board welcomes articles from readers suitable for, and understandable to, any practising or student physicist. Review papers and contributions of general interest are particularly welcome.