The HPC Technology Roadmap

Supercomputing Today and Tomorrow:

Seminar series on current and future high performance computing systems

This seminar series is arranged within the HPC Technology Roadmap technology watch study, undertaken by NSC and CSCISM. The study is a part of the European ENACTS project. Please refer to the HPC Technology Roadmap home page for further information about the study and the ENACTS project.

All seminars take place in the Physics building (Fysikhuset) on campus Valla, Linköping University.

Programme

"The Future of Scalable Scientific Computing"

Dr. Martin Walker

Time:    June 8, 2001 at 10.15
Place:    Room Schrödinger, Physics building (Fysikhuset), Campus Valla, Linköping University

Extrapolating from existing technology and products, and those in an advanced stage of development by Compaq Computer Corporation, Dr. Walker will extrapolate and speculate on what the next decade may bring for high-performance computing. Some current inhibitors to the interaction of humans with machines, like bandwidth of access, will effectively disappear. Current emphasis on data and information will be replaced by a focus on knowledge and its management. Current sharp distinction between real and virtual will become fuzzy, and the transition from virtual to real greatly eased for many products.

"Supercomputing Today and Tomorrow"

Dr. Wolfgang Mertz

Time:    June 11, 2001 at 10.15
Place:    Kepler 1, Physics building (Fysikhuset), Campus Valla, Linköping University

The presentation will first discuss SGI's current technology. Second, an outlook on future technologies from the point of view of SGI will be presented. Topics related to future technologies that will be discussed include chip technology, networks, programming models, graphics, storage, architecture, and more.

"Making the Transition from Vector to Scalar at Fujitsu"

This seminar is cancelled!

Dr. David Snelling

Time:    June 14, 2001 at 10.15
Place:    Room Schrödinger, Physics building (Fysikhuset), Campus Valla, Linköping University

"HPC Strategy Evolution"

Dr. Benoit Marchand

Time:    June 18, 2001 at 10.15
Place:    Room Schrödinger, Physics building (Fysikhuset), Campus Valla, Linköping University

Since the beginning of 'compute' time, compute intensive system solutions were either based on specialized costly HW or on commercial systems.  The result is a misfit in all cases. Sun believes that a third solution exists: specialized commercial derivatives. We also believe that customization of technical solutions will become key in the future. That is, no longer will a single system type be able to address most markets. Finally, we foresee the emergence of VLSS (Very Large Scale Systems) and we're actively pursueing development of novel solutions to address future requirements.

"Supercomputing: Possibilities and Pitfalls"

Dr. Jamshed Mirza

Time:    June 19, 2001 at 10.15
Place:    Room Schrödinger, Physics building (Fysikhuset), Campus Valla, Linköping University

Scalable Supercomputers today hold the possibility of "unlimited" performance in a cost-effective way. That opens up the possibility of not only solving traditional Grand Challenge problems, but also to use these systems in new, innovative, and more pervasive ways to solve problems that will impact our lives in a more direct way and change the way we live, work and conduct business.

But this promise of ever increasing performance will only come with many 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 system issues that can derail the performance trend we are on?

"How Shall We Program High Performance Computers?"

Dr. Burton J. Smith

Time:    June 20, 2001 at 10.15
Place:    Room Planck, Physics building (Fysikhuset), Campus Valla, Linköping University

Uniprocessor computer architecture has traditionally been motivated by programming languages and operating systems, with benchmarks written in the usual languages also having some influence. In high performance computing the situation is curiously reversed, with architecture determining the principal characteristics of programming languages, operating systems, and benchmarks. The result has been chaos; a "software crisis" has been declared, and better tools for the development of parallel software have been demanded. The outlook for good tools is bleak without a new approach to the problem, which should include the engineering of computer systems with both system and application software in mind and the development of programming abstractions that are both effective and efficient on hardware we can build.

Burton Smith is Chief Scientist of Cray Inc. He received the BSEE from the University of New Mexico in 1967 and the Sc.D. from MIT in 1972. From 1985 to 1988 he was Fellow at the Supercomputing Research Center of the Institute for Defense Analyses in Maryland. Before that, he was Vice President, Research and Development at Denelcor, Inc. and was chief architect of the HEP computer system. Dr. Smith is a Fellow of both the ACM and the IEEE, and winner of the IEEE-ACM Eckert-Mauchly award in 1991. His main interest is general purpose parallel computer architecture.