9320427 Cremer This award is for creating a mixed shared and distributed memory computing cluster to be used in support of the following research projects in experimental parallel and distributed computing: (1) Physical System Simulation and Applied Numerical Computing: While faster and more powerful hardware for scientific computation continues to be developed, that task of creating software to take full advantage of such hardware remains quite difficult. This project aims to develop fundamental techniques and software tools that facilitate the development of high-performance physical system simulation environments. Particular emphasis is on development of high-performance mechanical systems simulators, parallel real-time integration methods, and a framework for scenario control in real-time driving simulation. (2) Automated Reasoning: Reasoning algorithms are used in many application areas of computer science and artificial intelligence, such as hardware and software verification, deductive data bases, constraint solving, deductive plan generation, program specification and generation, and logic programming. The objective of this project is to develop new high performance reasoning algorithms, focusing on user-friendly induction provers, satisfiability decision procedures, and distributed/parallel theorem provers. Each of these requires a combination of implementation, experimentation, and theoretical analysis. (3) Self-Stabilizing Systems: Self-stabilizing programs are programs that require no initialization and can recover from arbitrary transient failures. In large systems, software failures ma y occur due to timing problems that can only be prevented by costly synchronization procedures. Using the techniques of self-stabilization, the need for such synchronization procedures can be avoided by tolerating and recovering from faults. Moreover, it has been such programs have applications beyond fault-tolerance -- stabilization can be useful to adapt to dynamic environments where changes to the system resemble transient faults. This project examines the following problems: stabilizing simulation; stabilizing adaptive programs; and discrepancies between theoretical and empirical models of self-stabilization. *** @ j SOUNDRECEXE @ j i SSMARQUESCR @ j B SSMYST SCR @ j L SSSTARS SCR @ j D TADA WAV @ j & l TARTAN BMP @ j 4 v TERMINALEXE @ j E B TERMINALHLP @ j 9320427 Cremer This award is for creating a mixed shared and distributed memory computing cluster to be used in support of d n U W . 0 _ d $ $ $ G d n d Times Symbol " Helvetica Chicago Times New Roman & Arial 5 Courier New R ZapfDingbats Palatino Greek GenMath MathMeteor MT Extra " + e+ e 4 Cremer/Iowa Mark Purvis Mark Purvis
