Cryptography, meteorology, national security, simulation of nuclear explosions, scientific research in the subatomic field: these were (and still are) the main fields of application of supercomputers , computers that are very different from our Personal Computers.
Supercomputers have features that allow them to perform a considerable number of calculations in a very short time : multiple processors, miniaturized circuit components, particular cooling systems, speed in recovering stored data, use of vector arithmetic.
Even more fascinating, however, is the history of these machines.
Seymour Cray and supercomputers
We are in the midst of World War II: the US military, through its team of experts, is constantly working to decipher the secret messages of the Nazi army. These experts are essentially mathematicians, physicists, engineers, and anyone with a strong logical and mathematical ability. In a short time, we have gone from manual cryptography to the creation of machines capable of probing all the combinations necessary to decrypt messages in code , such as those coming from the famous Enigma. A sort of pre-supercomputer was therefore used.
But the war, just as it had a beginning, also had an end. What happened to all these cryptography specialists? Many of them led very different lives before the war: mathematicians and physicists were academics, academics, researchers. The engineers, on the other hand, came from humbler family situations than their more "theoretical" colleagues, and were therefore salesmen, workers, or ran the family business. Like Seymour Cray, who initially ran his father's farm.
The birth of the ERA
After the war, many returned to their former lives, but others were not, such as William Norris and Howard Engstrom. The two had particular skills, exceptional machines, and a dream: to find a market for these machines outside the military sphere. And so, after a desperate search for funding, some compromise with the US Air Force and the acquisition of the company by a glider factory, they founded ERA ( Engineering Research Associates ), a company that had the purpose of realizing ad hoc machines and sell these supercomputers to airlines, the NSA, and others . The watchword was freedom . All engineers had to have as their sole concern the construction and improvement of their machines; thus the first tube supercomputers were born.
In this context, Seymour Cray appears for the first time : fresh from a degree in electronic engineering at the University of Minnesota, he was hired in ERA and, at the age of 25, he immediately distinguished himself both from a technical point of view and as a leader. Brilliant, dedicated to work but not workaholic, he loved to immerse himself in his circuits without any kind of distraction; for this reason, since he hated the constant interference of managers, at one point he began to work alone in a laboratory in Chippewa.
He had little patience with colleagues who he thought were asking "stupid" questions or weren't quick enough to complete a task, but he had everyone's respect. After two weeks, if Cray said something could be done better and needed to be changed, then it had to be. Within a few weeks he won the trust of all his colleagues , even the older and more experienced ones, and with him they built ever more powerful and faster machines.
Seymour Cray's machines
In particular we recall the computer 1103, the first to mount transistors in place of valves ; this machine represented a real watershed between the “old” tube technology and the “new” transistor. After the sale of the ERA in 1957, the marriage between Cray and the company ended, but that did not stop him: he worked at the CDC ( Control Data Corporation ) and then founded his company in 1972, Cray Research Inc. , With the purpose to develop ever more powerful and faster machines and in total creative freedom.
Among his most famous machines are the Cray-1 (capable of performing 240 million calculations per second), the Cray X-Mp (which first exploited the multiprocessor principle), the Cray-2 (which used the fluorinert as coolant) and the Cray Y-Mp.
But like all good stories, Seymour Cray's also has an end. In the 1980s the market had changed a lot; personal computers had begun to take hold, and over time they became a much cheaper machine: smaller, equally fast, and much cheaper.
Cray had continued to dream the same, and already had his new machine in mind: the Cray-3 , a new computer whose circuitry was supposed to be gallium arsenide; this inorganic compound is now widely used in photovoltaic cells, DVD players, microwaves and laser technology, but had limited success at the time, and had never been used for a computer. For the first time, Seymour Cray did not find the support of his colleagues: his project was considered too risky.
Cray with great stoicism decided not to give up, and left with a smile the company he had created to found another, the Cray Computer Corporation , and thus continue his work. And he succeeded, developing both Cray-3 and Cray-4; but these machines were considered too expensive, they did not sell, and so in 1995 the Cray Computer Corporation had to file for bankruptcy. Once again Cray went ahead and founded SRC Computers in 1996 , but when the project had just started the engineer died of a car accident, putting the final point on his story. However, his legacy is still alive today.
Supercomputers today and the new IT revolution
Despite the end of the first boom of these machines, the research and development of supercomputers has never stopped . Recently Zuckerberg announced the RSC ( Research SuperCluster ), a Meta supercomputer for the development of AI and the Metaverse.
We could think of a new computer revolution thanks to next generation supercomputers, although we are pushing more and more for quantum computers. In 2019, Google announced that it had built its own quantum computer called Sycamore. So who knows who will be the leader of this new revolution, which technology will be the master, and if we will have another Seymour Cray as a beacon .
Article by Antonella Babbone