Brian J. d'Auriol, Ph.D.

Seminar Details
Return to d'Auriol's Seminar Page


Brian J. d'Auriol, Systolic Computing: What it was, What it is, and What it is Becoming, Departmental Seminar, Computer Science Seminar Series, Sept. 17, 1998, Department of mathematics and Computer Science, The University of Akron, Akron, Ohio, USA.

Since its introduction in 1978 by Kung and Leiserson systolic processing has been and continues to be an important research area. A significant reason for this research interest is the potential to exploit the intrinsic parallelism in computations lending themselves to systolic processing solutions. A second important reason is that there are many application areas for systolic processing. Examples of such application areas include Digital Signal Processing (DSP), matrix computations, symbolic processing, sorting and relational database. Since 1978, significant advances in systolic processing, practical and theoretical, have contributed to a rich and relatively new field of study.

Systolic processing poses an interesting study, not only because of its potential usefulness and importance, but also as an example of research itself: from the original conception of an idea, to its elaboration in practice, the theoretical frameworks that evolve to describe various aspects and finally to unification of some or all of those frameworks. A singular point concerning such research endeavors is reached when the research is used as a basis or in support of new research directions. There is now some evidence to show that systolic processing has reached this singular point. Lengauer states ``The polytope model for loop parallelization has its origin in systolic design...''

This talk presents a special lecture on systolic computing, its origins, important developments and current interpretation

Last Updated: August 1, 2007