Tuesday, 4 April 2000

09:15-10:15    Keynote session I

Logic design methods for processors have changed over the years, along with changing requirements, device technologies, and CAD tools. This talk follows the evolution of logic design methods at Intel, looking ahead into impacts of Deep-Sub-Micron technology. Emphasis is put on the role of CAD and on the issue of synchronous versus asynchronous methodology.

About the speaker: Avi Kolodny (PhD, Technion) has been engaged in device research and CAD development at Intel Corp. in Santa Clara, CA and Haifa, Israel. He is currently conducting research and teaching CAD of VLSI at the Technion in Israel.

2:30-3:45    Keynote Session II

The systems considered consist of a host, comb-units, clocked registers and wires. A system is semisystolic if every directed circuit in the implied graph has at least one clocked register. The clock must be slow enough to make sure that every path which consists of comb-units alone completes its operation between two consecutive clock ticks. Retiming, as well as two techniques, tiling and bypassing, are suggested to modify the structure of a semisystolic system. The purpose of these modifications is to shortens the longest register-free path, and thus allow an increase of the frequency of the clock, without modifying the operation of the system, as the host sees it. The techniques are demonstrated on a number of examples. This talk is based on two joint papers of the author and Ami Litman.

Thursday, 6 April 2000

11:30-12:30    Keynote Session III

Static biological knowledge (DNA sequence and protein structure) have natural computer representations, enabing computers to greatly accelerate advances in these areas.  Dyanmic biological knowledge regarding biomolecular processes inside the cell and among cells have so far defied natural computer representation, thus the vast knowledge regarding biomolecular processes is accumulated and described in a way that precludes using computers as research accelerators in a similar fashion.

In this talk we describe an approach to model biomolecular processes using process algebra in an attempt to help redemy this.  In our approach molecules are prepresented by processes and molecular interaction is modelled by communication and process state change. Specific examples of using Pi-calculs to obtain qualitative models, and stochastic Pi-calculus to obtain quantitiative models, are provided.

Joint work with Aviv Regev and Bill Silverman.

About the speaker: Ehud Shapiro is an Associate Professor at the Department of Computer Scienc and Applied Math at the Weizmann Institute of Science, which he jointed in 1992 after completing his Ph.D. at Yale University. Shapiro's early research focused on logic languages for parallel and distributed programming.  Between 1994 and 1998 Shapiro took leave of absence, in which he founded Ubique, one of the first Internet software companies, that was later sold to AOL and is now a subsidiary of Lotus/IBM. Since returning to Weizmann in 1998 Shapiro's interests lie on the borderline of biology and computer science.