High-Performance Computing (HPC) and Problem Solving Environments (PSEs)

Research Associate: A. Catlin

Sponsors: ARPA, NSF

Problem solving environments are systems that provide all the computational facilities for solving a target class of problems. See the general PSE web site and the Purdue PSE research group web site for more information. Scientific PSEs are intrinsically linked with high performance computing, indeed, one of their primary goals is to hide the complexity of parallel machines. This project has several focus areas

• Applications modeled by partial differential equations (PDEs), as in the Parallel ELLPACK projects.
• Management of parallel computing resources, see the slides at http://www.cs.purdue.edu/homes/jrr/slides/Salishan1.html and http://www.cs.purdue.edu/research/cse/hpc-pses.
• Computational intelligence to support scientific PSEs, the PYTHIA project.
• Network based scientific computing as in the Net //ELLPACK and SciAgents projects.