HPC — What Might the Future Hold?

Abstract

Scalable cluster systems today hold the promise of “unlimited” performance in a cost-effective way. This opens up the possibility of not only solving traditional grand challenge problems, but also of using these systems in new, innovative, and more pervasive ways to solve problems that impact our lives in a more direct way and change the way we live, work, and conduct business.

We will discuss some of the new ways in which HPC is being used to conduct business and provide a competitive edge. Analytic methods, statistical modeling, and pattern searching algorithms that are common in scientific computing can now be applied to the vast amounts of operational and historical data generated by business transactions to extract knowledge that can be used for competitive advantage. Even complete applications that were traditionally thought of as technical computing can be incorporated into key business processes to drive critical business decisions. For example, high resolution localized weather forecasting models enabled by these systems can now become an integral part of decision making in businesses in which weather plays an important part. Advanced pattern searching algorithms used in drug design can now be used in fraud detection in the financial industry or for fuzzy text search in the media industry. Similarly, the growing computing capability made possible by these systems is allowing the solution of larger, more complex and more accurate, business optimization problems.

We will discuss technology trends that over the coming years promise ever-more powerful supercomputers. But this promise of ever increasing performance will only come with many potential pitfalls. With current technology trends, exponential increases in parallelism is the only way to exponential increases in performance. Will we be able to put together, manage, and productively use such systems? What are the systems issues that can derail the performance trend we are on? What are the promising technologies that will mitigate the problems? We will discuss some of these issues and touch on work underway to address some of them.

We will also discuss the interplay and synergy between two key initiatives — Autonomic Computing and Grid Computing. These initiatives are expected to generate technology and insights that will define the high-end computing systems of the future. Both are aimed at making computing resources more accessible and easier to use and manage. At IBM, eLiza is a key, company-wide push in the area of autonomic computing technologies; its goal is to research, develop and deploy technologies in servers, storage and software to make computers and computing systems more “self-aware, self-managing, self-adapting, and self-healing”. Grid Computing is a new computing paradigm that will allow better sharing and utilization of computing resources. Its grand vision is to turn computing into a utility, whereby users get reliable, secure, and ubiquitous access to applications, data, servers, and storage in a completely transparent manner. The two technologies are synergistic and together will define a more complete solution for computing in the future.