• Parallelizing Irregular Applications
• Lonestar Benchmark Suite

• Compiling for Distributed Memory Computers
• Control Dependence Computation
• Empirical Optimization
• Fault Tolerance
• Fractal Symbolic Analysis
• Functional Languages with Logic Variables
• Memory Hierarchy Optimizations
• Model Checking for Hierarchical State Machines
• Program Representation
• Static Single Assignment Form
• Synthesis of Irregular Codes
• Web Services for Distributed Scientific Computing

Compiling for Distributed-Memory Computers

Our project implemented one of the first compilers that generated message-passing code from sequential shared-memory programs with data distribution directives. We introduced several concepts such as runtime resolution and the owner-computes rule that have become part of the standard terminology in this area. Other projects with similar goals included FORTRAN-D (Ken Kennedy), Vienna FORTRAN (Hans Zima), and Kali (Piyush Mehrotra).

The lessons learnt from all these projects were incorporated into the High-Performance FORTRAN (HPF) language which is still supported by several software vendors.

Publications:

• David Bau, Induprakas Kodukula, Vladimir Kotlyar, and Keshav Pingali, Paul Stodghill Solving Alignment Using Elementary Linear Algebra Languages and Compilers for Parallel Computing, 08/08/1994
• Anne Rogers and Keshav Pingali Compiling for Distributed Memory Architectures IEEE Transactions on Parallel and Distributed Systems, 06/10/1994
• Keshav Pingali and Anne Rogers Compiler Parallelization of SIMPLE for a Distributed Memory Machine Technical Report TR90-1084, Cornell Computer Science Department, 01/01/1990
• Anne Rogers and Keshav Pingali Process Decomposition Through Locality of Reference Programming Language Design and Implementation, 06/21/1989