Pleasant Memories of C-DAC

Dr. Paulraj was associated with C-DAC as its Director and was actively involved in the first mission. Presently he is Professor at the Information Systems Laboratory at Stanford University, USA.

It is with pleasure and pride that I recall my brief association with C-DAC between 1987-90. In 1987 Dr. S.G. Pitroda invited me to join the newly organized Governing Council for C-DAC. The vision for C-DAC set by the council was to develop a high speed computing capability to support "production" computing needs such as weather forecasting and oil exploration. In short, free India of the need to import (with frequent export license denials by the US) of high end computers and associated application software for critical national requirements. Consensus quickly converged on a microprocessor based parallel architecture, as this was the only pragmatic approach given the devices for which India could obtain export licenses. The more flexible and software friendly (CRAY-like) vector architecture was dropped since the required semiconductor devices were not importable. As we explored parallel architecture tradeoffs, we decided on an even simpler hardware approach using a message passing architecture instead of the more software friendly-shared memory architecture. The huge technical challenge for effectively solving target production applications now shifted to good software tools and more importantly carefully written applications software, which would offer acceptable performance to the end production users. So in 1988, this appeared to be an opportunity in which India could be a world leader given the low costs and the innate talent we could harness in software development, and thus beyond serving our own needs, we could also create a large international market for our machines.

In 1988, C-DAC began operations in Pune and I often commuted from Bangalore to help build the Pune team and start on the system design. I have very pleasant memories of those days and many teammates we brought into C-DAC served the institution with great dedication and some remain even today. Later in 1988, we set up a center in Bangalore for the system software group and thereafter I primarily focused my efforts at this center (C-DAC was a part time effort for me as I was primarily assigned to BEL, Bangalore). The Bangalore center built the systems software suite of parallel tools, compilers and operating system. A small hardware group also prototyped a fast processing node based i860/i960 and a homegrown cut-through mesh switch. The Pune center built an Inmos-transputer based machine and started on applications software projects.

Over the years I have been delighted to see how well the C-DAC family has grown and I am proud to have been part in shaping its initial years. Message passing parallel computers are now not favored except in a few special applications. Even large vector machines have become extinct in favor of uniprocessor or small-shared memory multiprocessor machines. But computing technology itself whether in desktops, servers or dedicated applications in communications and networking have seen explosive growth. C-DAC has rightly refocused and widened its goals to encompass larger domains of computing and information technology. C-DAC is also building a strong revenue stream in recent years and this is an important measure of its value in a market economy. Mr. Arora states in his message on the C-DAC web site  - the opportunities are mind boggling. I heartily agree. I am indeed happy to renew my contacts and stand ready to support C-DAC as it grows and reaches for new horizons.