THE CHOSEN ONE: Bhatkar and C-DAC had to battle numerous odds, including a blockade of critical components from the US, to produce the supercomputerIf ever there was a physical manifestation of the links between high science and the deep passions of spiritualism, it was played out for many years in the confines of the Centre for the Development of Advanced Computing (C-DAC), which catapulted India to a supercomputing power. A young and high-strung team of scientists and technocrats, with an average age of 23 or so, battled numerous odds, including a blockade of critical components from the US, through the late-1980s and the 1990s to craft Param, India's first supercomputer. Many versions followed.
THE CHOSEN ONE:
When first unveiled in 1991, Param wowed the world, and even countries like Russia and Japan were bested in the race to build a supercomputer. Quaintly, through all of this, it was a certain belief in a supreme entity and "divine grace" that propelled a core group, including C-DAC Director Vijay Bhatkar, to scale new heights in supercomputing technology. "Success has a time and place dimension," says Bhatkar. "It depends on the actors, their actions, the tools used and God's grace."
Bhatkar, now Chairman of the Pune-based ETH Research Lab, engaged in harnessing technology for the spread of education, was religious, but the crises he surmounted while building Param only reinforced his beliefs. "His deep spiritualism enabled him to endure the intense pressures of the task at hand. Anyone else in his place would have collapsed, taking the project down with him," insists PR Eknath, MD and CTO, DiviNet Access Technologies, who was Bhatkar's confidant and colleague during the C-DAC days.
Spiritualism ingrained equanimity in Bhatkar and enabled him to take a detached view of issues. "It brought much clarity and also helped me reach out to people, and society at large," he says. "I don't look at a machine. I look upon it as an abstraction, a mathematical entity." He was always a mathematics genius of sorts; even as a student in IIT, Delhi. People who knew him in his early years, especially his stint with Electronics Research and Development Centre (ER & DC), Thiruvananthapuram, recall it was often difficult to grasp what he was saying—when thinking, only garbled words came out as speech. His speech couldn't keep pace with his thoughts. Only those in sync with him could get it.
Bhatkar's and his team's journey in building the Param, a gigaflop range machine, was seeded in the 1987 high technology accord between then Prime Minister Rajiv Gandhi and US President Ronald Reagan. At the meeting, Gandhi was humiliated by the Americans, who, instead of offering India the current Cray supercomputer, offered an older version. The PM was also warned that if it was used for purposes other than weather forecasting, "the US will ensure not even a pin reaches India."
A few days later, Bhatkar received a call from his mentor KPP Nambiar, then Secretary, Department of Electronics (DoE), and was asked a pointed question, "Can you build a supercomputer?" Bhatkar had just quit as director of the ER& DC and had joined the Tatas in Pune. Bhatkar could only say: "I haven't seen one. But I have been reading about it." Only supercomputer applications were in the public domain; very little on its architecture was available. A meeting with Rajiv Gandhi ensued, in which Nambiar, KR Narayanan, then Minister of State for Technology, and Bhatkar were present. Bhatkar, without hesitation, agreed to undertake the mission in three years with Rs 34 crore.
"Three years, because that was about the time it took to import a supercomputer then, and Rs 34 crore because it was the prevailing price of the machine," explains Bhatkar, who banked on little else but intuitive gut feel. "If I had said five years or beyond, the PM would have lost interest," he says. For Rs 34 crore then, he not only agreed to build the supercomputer, but also an entire institution and infrastructure with manpower costs.
Starts The Super Mission
He had decided to take the parallel processing route to the machine—basically an ensemble of processors to undertake a supercomputing task. "This was the only way for us, considering the technology blockade," says Bhatkar, who eventually had to bank on friends, German and British professors, to source some critical components and route them through different countries to India. Moreover, he was certain that parallel processing was the future of supercomputing, despite scepticism within India and across the world.
A PROTOTYPE of Param was ready two years after the mission started. To prove the prowess of the true blue supercomputer, scientists took it to the 1990 Zurich supercomputing show, where it beat Japan, Germany and Russia. It landed second in the competition, only next to the US. Subsequently, its brilliance prompted even the Russians to purchase the Indian Param. The success heralded the emergence of India as a supercomputer power. Param's key applications are used in several signifi cant programmes in the country such as weather forecasting, remote sensing, molecular modelling and drug design.C-DAC, with its open and non-hierarchical approach, attracted the best of talent from across the country. It was open 24 hours and employees could choose the time of work. Children and family were welcome in the premises. Dinner was also served on campus. All of this, now taken for granted, was a novelty in the late-1980s. Engineers and design experts quit permanent jobs to be on contract for three years. Bhatkar led by example. He was the first to sign up on contract, when he could have sought governmental cushion. In fact, the first set of 20 people worked without pay for six months due to procedural delays at the ministry level.
A PROTOTYPE of Param was ready two years after the mission started. To prove the prowess of the true blue supercomputer, scientists took it to the 1990 Zurich supercomputing show, where it beat Japan, Germany and Russia. It landed second in the competition, only next to the US. Subsequently, its brilliance prompted even the Russians to purchase the Indian Param. The success heralded the emergence of India as a supercomputer power. Param's key applications are used in several signifi cant programmes in the country such as weather forecasting, remote sensing, molecular modelling and drug design.
Attracting the brightest from across India was easy, but channelling their individual brilliance towards a common goal was tough. "Dealing with thinkers is a different ballgame. Their thinking, right or wrong, has to be valued, argued out. They are an emotional lot," explains Eknath. The play of brilliance, egos, idiosyncrasies, emotional outbursts and deep sulks were the order of the day. Fostering an open culture helped. It wasn't uncommon at C-DAC. Even a junior engineer would barge into Bhatkar's room, wag a finger at him, and engage in a loud argument on any issue related to the project. "The biggest challenge was to bring coherence to the diversity of views," recalls Bhatkar. "You can only influence the course of an argument by your own knowledge, not by pulling rank."
Young engineers were also impatient. When they wanted a component, they wanted it 'now,' little realising it was an institution under the government of India after all, still under the license–permit raj. Nambiar and Bhatkar, nevertheless, did away with the usual government tendering system for purchases; such was the faith they had in the team. Purchasing teams at C-DAC lived up to the faith as they negotiated hard with vendors. Eknath remembers how they bargained with Synopsis and Mentorgraphics, design software majors, to purchase a $1 million pack for a mere $150,000. Not a single charge of corruption came up through the years, though government auditors were straining to seek indiscretions.
During the years, Bhatkar also took on the role of a mother hen, insulating his team from the wiles of bureaucracy. "I took all the hurt myself and never allowed it to spread to my team," he says. Even those who harboured professional jealousies and wanted to pull him down were treated with kindness. "This is indeed a rare trait in a leader," says Eknath. Earlier, as ER&DC Director, he was the only one treated with deference by the firebrand trade union leaders of Kerala.
Within two years, a prototype was ready, but none in India was willing to believe that it was indeed a true blue supercomputer. Therefore, Bhatkar took a strategic decision to benchmark his machine at the 1990 Zurich supercomputing show, in which the US, Japan, Germany, Russia and an Anglo- French consortium were to participate.
Again, C-DAC was embroiled in delays, typical of the government, with the result that its machine, in disassembled form, landed in Zurich at 2 am on the day the benchmarking test was scheduled. Twelve engineers worked through the early hours to put it together for the 11 am test. It refused to function at the first go. Surprisingly, it started humming when the benchmarking exercise started. The US led. India was accorded second place, followed by the Anglo-French consortium. Even the Germans and Russians were behind India. The Japanese failed to launch. "All indication was of a failure. It was nothing else but divine intervention," says Bhatkar.
A year later, in 1991, the full scale Param was launched, and the world media screamed: "An angry India does it."
Subsequently, even the Russians purchased the Indian Param. The ceremony in Russia was the towering moment for Bhatkar personally, as giants from the Russian Academy of Sciences—the very professors whose texts and theorems he had pored over, when he started his venture—gave him a standing ovation. "That's a moment I cherish to this day," says Bhatkar.
The C-DAC mission was accomplished, and it was the reason why the government wanted to shut shop, although Bhatkar wanted to scale up to a faster teraflop machine. His support system had slowly crumbled by then. Rajiv Gandhi was assassinated, Sam Pitroda, the tech-czar, was sidelined by the successor government, and Nambiar had moved on. On an occasion, a secretary to the DoE also flew down to Pune to preside over the closure of C-DAC. The institution, however, survived.
Bhatkar delivered the teraflop architecture machine in 1998, and stepped out of the institution, after conducting a phenomenal technological opera, to the applause of the scientific world, for a decade.
We at Outlookindia.com welcome feedback and your comments, including scathing criticism
1. Scathing, passionate, even angry critiques are welcome, but please do not indulge in abuse and invective. Our Primary concern is to keep the debate civil. We urge our users to try and express their disagreements without being disagreeable. Personal attacks are not welcome. No ad hominem please.
2. Please do not post the same message again and again in the same or different threads
3. Please keep your responses confined to the subject matter of the article you are responding to. Please note that our comments section is not a general free-for-all but for feedback to articles/blogs posted on the site
4. Our endeavour is to keep these forums unmoderated and unexpurgated. But if any of the above three conditions are violated, we reserve the right to delete any comment that we deem objectionable and also to withdraw posting privileges from the abuser. Please also note that hate-speech is punishable by law and in extreme circumstances, we may be forced to take legal action by tracing the IP addresses of the poster.
5. If someone is being abusive or personal, or generally being a troll or a flame-baiter, please do not descend to their level. The best response to such posters is to ignore them and send us a message at Mail AT outlookindia DOT com with the subject header COMPLAINT
6. Please do not copy and paste copyrighted material. If you do think that an article elsewhere has relevance to the point you wish to make, please only quote what is considered fair-use and provide a link to the article under question.
7. There is no particular outlookindia.com line on any subject. The views expressed in our opinion section are those of the author concerned and not that of all of outlookindia.com or all its authors.
8. Please also note that you are solely responsible for the comments posted by you on the site. The comments could be deleted or edited entirely at our discretion if we find them objectionable. However, the mere fact of their existence on our site does not mean that we necessarily approve of their contents. In short, the onus of responsibility for the comments remains solely with the authors thereof. Outlookindia.com or any of its group publications, may, however, retains the right to publish any of these comments, with or without editing, in any medium whatsoever. It is therefore in your own interest to be careful before posting.
9.Outlookindia.com is not responsible in any manner whatsoever for how any search engine -- such as Google, Bing etc -- caches or displays these comments. Please note that you are solely responsible for posting these comments and it is a privilege being granted to our registered users which can be withdrawn in case of abuse. To reiterate:
a. Comments once posted can only be deleted at the discretion of outlookindia.com
b. The comments reflect the views of the authors and not of outlookindia.com
c. outlookindia.com is not responsible in any manner whatsoever for the way search engines cache or display these comments
d. Please therefore take due caution before you post any comments as your words could potentially be used against you
10. We have an online thread for our comments policy:
You are welcome to post your suggestions here or in case you have a specific issue, to directly email us at Mail AT outlookindia DOT com with the subject header COMPLAINT