|Volume 7, Issue 1 -
Parallel Computing Pioneers - Arvind
Professor of Computer Science and Engineering, Massachusetts Institute of Technology
Throughout his long and distinguished career as a university professor and researcher, Arvind has been involved in all aspects of parallel computing and declarative programming languages. He has contributed to the development of data flow architectures, the implicitly parallel programming language Id, and the compilation of these types of languages on parallel machines.
Arvind became interested in computer science while working on a B. Tech. in electrical engineering, which he received from the Indian Institute of Technology in 1969. He went on to earn an M.S. and Ph.D. in computer science from the University of Minneapolis in 1972 and 1973. He was an assistant professor at the University of California, Irvine before joining the Computer Science and Engineering faculty at the Massachusetts Institute of Technology (MIT) in 1978. He has been a professor there since 1988, and has held the chair established by Charles and Jennifer Johnson since 1992.
Arvind currently heads the Computation Structures Group in the Laboratory for Computer Science. He is working on high-level specification and descriptions of architectures and protocols using a formalism known as Term Rewriting Systems (TRS's). This research encompasses hardware synthesis from TRS's and verification of their implementation against the specification TRS.
Arvind is the leader of the StarT project at MIT, an effort to develop general-purpose parallel systems based on scalable architectures, implicitly parallel programming languages, and explicitly parallel dialects of conventional programming languages. The StarT project has pursued this goal in the context of existing processor architectures and produced scalable hardware systems based on commercial workstations and servers.
"The basic idea behind the StarT project is that if a modern workstation or server is augmented to handle the specific needs of parallel processing, such as communication and remote memory accesses, then it becomes a much better building-block for a general purpose parallel machine," Arvind expains. "StarT-Voyager, our final approach, is a cluster of high performance PowerPC 604 nodes, specifically the IBM RS/6000 model 240 workstation, and includes an additional card called the Network Endpoint System (NES). The NES connects to the workstation's system bus and serves as a sophisticated network interface." An inital prototype of StarT-Voyager is scheduled for completion this fall.
Arvind is currently writing a book on implicit parallel programming with Dr. R.S. Nikhil of Compaq Cambridge Research Laboratories. He is the author or co-author of more than 80 papers, holds six patents, and serves on the editorial board of the Journal of Parallel and Distributed Computing, Journal of Functional Programming, and the International Journal of High Speed Computing. He has chaired and been a member of program committees for numerous technical meetings sponsored by ACM and IEEE. He is an ACM member, an IEEE Fellow, and was awarded the Charles Babbage Outstanding Scientist Award in 1994.