NSF-CNPq Collaborative Research: Implementation and Compilation of High-Performance, Scalable Parallel Constraint Programming Systems

点击翻译按钮获取中文摘要

Proposal Number: CCR-9900320TITLE: High-performance, Scalable Parallel Constraint Programming SystemsPI (and CoPI's): Gopal Gupta and Enrico PontelliSummary:Constraint programming incorporates efficient techniques forsolving problems that involve a large amount of search. Evenafter incorporating these efficient constraint-solvingtechniques, the search space can still be very large.This project proposes to perform research on design andimplementation of constraint programming systems on shared memorymultiprocessors (SMPs) so that these compute-intensiveconstraint programs can be executed much faster. Efficientimplementation techniques for realizing scalable constraintprogramming systems on distributed memory multiprocessors (DMPs)as well as on distributed clusters of SMPs will also be designed.The project's eventual aim is to build a unified parallel systemthat seamlessly runs on a distributed network of shared memorymultiprocessors. The project will be pursued in formal cooperationwith the Federal University of Rio de Janeiro (UFRJ) and the FederalUniversity of Rio Grande do Sul (UFRGS) in Brazil. The researchwill make contributions to the area of constraint programming,logic programming, parallel processing, and programming languagesthrough the new techniques developed for scalable implementation,compilation, and optimizations.