Seymour Roger Cray (September 28, 1925 – October 5, 1996) was an American electrical engineer, computer scientist, mathematician, and supercomputer architect. He designed a series of computers that were the fastest in the world for many years and founded Cray Research, which built many of these machines. Known as "the father of supercomputing," Cray is credited with creating the supercomputer industry. Joel S. Birnbaum, then chief technology officer of Hewlett-Packard, said of him: "It seems impossible to exaggerate the effect he had on the industry; many of the things that high performance computers now do routinely were at the farthest edge of credibility when Seymour envisioned them." Larry Smarr, then director of the National Center for Supercomputing Applications at the University of Illinois, said that Cray is "the Thomas Edison of the supercomputing industry."
Early life
Cray was born in 1925 in Chippewa Falls, Wisconsin, to Seymour R. and Lillian Cray. His father was a civil engineer who encouraged Cray's interest in science and engineering. At the age of ten, Cray built a device using parts from an Erector Set that changed punched paper tape into Morse code signals. The basement of the family home was used as a "laboratory" for Cray's experiments.
Cray graduated from Chippewa Falls High School in 1943. He was drafted into World War II as a radio operator and fought in Europe. Later, he worked in the Pacific theater, where he helped break Japanese naval codes. After returning to the United States, he got a B.Sc. in electrical engineering from the University of Minnesota, graduating in 1949. He then earned a M.Sc. in applied mathematics in 1951.
Career
In 1950, Cray joined Engineering Research Associates (ERA) in Saint Paul, Minnesota. ERA was formed from a former United States Navy laboratory that built machines to break codes. This tradition continued at ERA whenever similar work was available. ERA learned about computer technology during one such project, but at other times, the company worked on many different engineering tasks.
Cray quickly became known as an expert in digital computer technology, especially after his work on the ERA 1103, the first commercially successful scientific computer. He stayed at ERA when the company was bought by Remington Rand and later by Sperry Corporation in the early 1950s. At the newly created Sperry Rand, ERA became the part of the company responsible for scientific computing in the UNIVAC division.
Cray, along with William Norris, later became unhappy with ERA and left to form Sperry Rand. In 1957, they started a new company called Control Data Corporation.
By 1960, Cray had completed the design of the CDC 1604, an improved and less expensive version of the ERA 1103 that performed well for its cost. Even as the CDC 1604 was being sent to customers in 1960, Cray began working on new computer designs. He first worked on an upgraded version of the CDC 1604 (the CDC 3000 series), but company leaders wanted these machines to focus on "business and commercial" tasks for regular customers. Cray disliked designing these simpler machines, which required low costs to sell many units. His goal was to build the fastest computer in the world. After some initial work on the CDC 3000 series, he handed the project to others and moved on to the CDC 6600. Some features of the 6600 first appeared in the 3000 series.
Although the CDC 6600 was not using the newest technology, Cray worked hard to make the machine run as quickly as possible. He understood that performance was not just about the speed of the computer’s main processor, but also about maximizing the speed of data transfers. He later said, "Anyone can build a fast CPU. The trick is to build a fast system."
The CDC 6600 was the first commercial supercomputer, outperforming all other machines at the time by a large amount. While it was expensive, it was the only option for those who needed the fastest computer available. When other companies, like IBM, tried to create similar machines, they faced problems (such as the IBM 7030 Stretch). Cray solved a major design challenge called "imprecise interrupts," which had caused IBM’s failure. He did this by replacing I/O interrupts with a system using ten mini-computers called peripheral processors. These mini-computers handled all data transfers to and from the 6600’s main memory. The next CDC 7600 improved the speed of the 6600 by five times.
In 1963, a Business Week article announced the CDC 6600 and included a quote from Seymour Cray:
— Seymour Cray, "Computers get faster than ever," Business Week (31 August 1963): p. 28.
During this time, Cray became frustrated with what he saw as interference from CDC management. He needed a quiet work environment with little management involvement, but as the company grew, he was often interrupted by middle managers who, according to Cray, did little but show off and used him as a sales tool by introducing him to customers.
Cray decided he needed to move from St. Paul to a location far enough that it would be too long to visit quickly and expensive to call by phone, but close enough for important meetings. After discussion, Norris supported the move and helped set up a new laboratory on land Cray owned in his hometown of Chippewa Falls. Part of the reason for the move may have been Cray’s concerns about a possible nuclear war, which he felt made the Twin Cities unsafe. His home, built near the new CDC laboratory, included a large bomb shelter.
The new Chippewa Lab was set up during the middle of the 6600 project, but it did not slow the project down. After the 6600 was completed, the next project, the CDC 7600, was developed in Chippewa Falls. The 7600 had speeds ten times faster than the 6600. The failed follow-up to the 7600, the CDC 8600, ended Cray’s streak of successes at CDC in 1972.
Although the 6600 and 7600 were major successes, both projects almost caused the company to go bankrupt during their development. The 8600 faced similar problems, and Cray decided the only solution was to start over. This time, Norris was unwilling to take the risk, and another project, the CDC STAR-100, seemed to be progressing better. Norris agreed to keep the STAR-100 project alive at a low level until it was completed, then fund the 8600 fully. Cray refused to work under these conditions and left the company.
The split was friendly, and a year later, Cray started a new company, Cray Research, on the same Chippewa Falls property. Norris invested $250,000 in the new company. Like CDC, Cray Research had its research and development in Chippewa Falls, and its business headquarters in Minneapolis. Unlike CDC, Cray’s manufacturing was also in Chippewa Falls.
At first, it was unclear what the new company would do. It seemed unlikely they could afford to develop a new computer, as the large CDC had struggled to support even one. When the company’s president went to Wall Street to find funding, he was surprised to learn that Cray’s reputation was very strong. Far from struggling to find a role in the market, Cray found that financial support was readily available for a new machine.
After several years of development, their first product was released in 1976 as the Cray-1. Like earlier Cray designs, the Cray-1 focused on making the entire computer fast, not just the processor. When it was released, it outperformed nearly every other machine in terms of speed, including the STAR-100 that had helped fund the 8600 project. The only machine that could match the Cray-1’s performance was the ILLIAC IV, a specialized machine that rarely reached its maximum speed except for specific tasks. In general, the Cray-1 far outperformed all other machines on the market.
The first Cray-1, serial number 001, was sent to Los Alamos National Laboratory in 1976. That summer, the first full system was sold to the National Center for Atmospheric Research (NCAR) for $8.8 million. The company had expected to sell about a dozen such machines, based on sales from the CDC era, so the price was set accordingly. Eventually, more than 80 Cray-1s were sold, the company became financially successful, and Cray’s work with supercomputers earned him the nickname
Personal life
Cray married Verene Voll in 1947. They had known each other since childhood. She was the daughter of a Methodist minister, as was Cray's mother, and Verene worked as a nutritionist. They had three children. Cray and Voll divorced around 1978. He later married Geri M. Harrand. Cray was the grandfather of Andrew Cray, an LGBTQ rights activist.
Cray avoided publicity. There are some unusual stories about his life away from work, called "Rollwagenisms," from John A. Rollwagen, who was then the CEO of Cray Research. Cray enjoyed skiing, windsurfing, tennis, and other sports. Another favorite pastime was digging a tunnel under his home. He said the elves gave him ideas while he worked in the tunnel: "While I'm digging in the tunnel, the elves will often come to me with solutions to my problem."
One story says that when Cray was asked by management to provide detailed one-year and five-year plans for his next machine, he simply wrote, "Five-year goal: Build the biggest computer in the world. One year goal: One-fifth of the above." Another time, when expected to write a multi-page detailed status report for the company executives, Cray's two-sentence report read: "Activity is progressing satisfactorily as outlined under the June plan. There have been no significant changes or deviations from the June plan."
Cray was mortally wounded in a car accident where his vehicle flipped over while he was merging his Jeep Cherokee onto Interstate 25 near the Air Force Academy in Colorado. The accident was caused by another vehicle trying to overtake Cray's car. Police gave the at-fault driver a citation for careless driving. Cray died of his injuries on October 5, 1996, two weeks after the accident and one week after his 71st birthday.
Posthumous
The IEEE Computer Society's Seymour Cray Computer Engineering Award was created in late 1997. It honors new and creative ideas in advanced computer systems that process information quickly. The award is inspired by Seymour Cray's innovative approach to computing.