Seymour Cray

Electrical Engineer

• Born: September 28, 1925
• Birthplace: Chippewa Falla, Wisconsin
• Died: October 5, 1996
• Place of death: Colorado Springs, Colorado

Designer of supercomputers and founder of Cray Research

Born: September 28, 1925; Chippewa Falls, Wisconsin

Died: October 5, 1996; Colorado Springs, Colorado

Primary Field: Computer science

Specialty: Computer hardware

Primary Company/Organization: Cray Research

Introduction

Seymour Cray is acknowledged in the history of computer science as the father of supercomputing and the founder of the supercomputer industry. Without him, it is quite possible that we would not have the type of computers or the electronic devices we use to this day. Joel Birnbaum famously stated that “Many of the things that high performance computers now do routinely were at the farthest edge of credibility when Seymour envisioned them.” Cray led a private life away from the computer science industry until his tragic and unexpected death in 1996 as the result of a traffic accident.

Early Life

Seymour Roger Cray was born on September 28, 1925, in the northwest of Wisconsin in Chippewa Falls. His father was a civil engineer. With his father's occupation and Cray's fascination with radios, motors, and electrical circuits, Cray developed an interest and talent in engineering and science. This passion led his parents to allow Cray to use the basement of their house as a laboratory for his studies and research. Cray's love of inventions was manifested early, and throughout his life he enjoyed tinkering with machines that could assist him or others.

When Cray was a young boy, he rigged a Morse code connection between his bedroom and that of his sister so that they could talk after lights were out. His father discovered the late-night clicking and told Cray to shut down the system because it was bothering the rest of the family. Cray's response to his father's request was to convert the clickers to lights and to continue talking with his sister. Cray composed this device out of Erector Set components that converted punched paper tape passed through it into Morse code signals.

Cray attended grade school and high school in Chippewa Falls. He completed high school in 1943 and was then drafted into the U.S. Army to serve in World War II.

Life's Work

Cray's technical abilities gave him the opportunity to become a radio operator for the U.S. Army. Once he had received his training, Cray was assigned to Europe and, although he served on the front lines, did not sustain any injuries. Cray was later transferred to the Pacific theater, where he worked on breaking coded messages that the Japanese were sending. At the conclusion of the war, he returned to the United States.

Once at home, Cray decided to attend the University of Minnesota, where he majored in electrical engineering. In 1949, he received his bachelor of science degree and then studied applied mathematics, receiving his master of science degree in 1951. After receiving his master's, Cray he joined Engineering Research Associates (ERA) in Saint Paul, Minnesota, at the suggestion of one of his university instructors. ERA was located in a renovated and restructured wooden glider factory. At ERA, Cray worked on digital computers, which were cutting-edge at the time. ERA was formed out of a former U.S. Navy laboratory that had devised and manufactured code-breaking machines. ERA specialized in computer technology but also developed a wide variety of projects that utilized basic engineering. During his tenure at ERA, Cray became known as an expert on digital technology. He was one of the leading designers of the ERA 1103, which became one of the first successful computers for scientific purposes.

In 1955, ERA was acquired by Remington Rand Company, a maker of business machines, which was then in turn acquired by Sperry Corporation to form Sperry Rand Corporation. All remaining staff from ERA were placed in Sperry Rand's scientific computing division, including Cray. Two years later, in 1957, the scientific division was closed and Cray and four other employees sought work at the Control Data Corporation (CDC). The other four were hired, but Cray was denied employment by CDC's chief executive officer (CEO), William Norris, until Cray finished a project that he had begun for the Navy. Norris had a good relationship with the Navy and wanted that relationship to remain intact. The project, the Naval Tactical Data System, was finished early 1958; thereafter, CDC hired Cray.

While at CDC, Cray became the lead designer of the next product after the ERA 1103, the CDC 1604. This product allowed CDC to become a major computer manufacturer. The CDC 1604 became one of the first, if not the first, computer to replace vacuum tubes with smaller transistors. Before the product shipped for sale, Cray set to work on the next version, the CDC 6600. The CDC 6600 was the computer first dubbed a supercomputer. The 6600 debuted as a hugely powerful machine, largely because of Cray's work on optimizing the system, and it outperformed everything else on the market at that time. In response to feeble attempts by competitors to outdo the 6600, Cray designed the CDC 7600, which proved to be five times faster than the 6600. The last system that Cray completed at CDC was the 8600. He began work on it in 1968 and realized that the enhanced clock speed would not allow him to reach his goals. Recognizing the need to enhance the design, Cray designed the 8600 with four processors, all sharing the same memory.

By 1970, Cray had become responsible for systems that would shape the high-performance computer industry for years to come. Around this period, Cray became exasperated with interference by CDC's upper management and lobbied for a laboratory to be built in his hometown of Chippewa Falls. He got his wish, and CDC built a new laboratory on land that Cray owned.

Upon his departure from CDC, Cray founded his own company, Cray Research, Inc., with the single goal of building the fastest computers on the globe. Norris invested $300,000 in start-up money for Cray's company. Research and development, as well as manufacturing, were based in Chippewa Falls, while the firm's business headquarters were in Minneapolis. When he started Cray Research, Cray shelved the 8600 design, primarily because at that time he believed that the software problems were too great for the company and industry to handle. Cray's company did not have the money to design a new computer, but after a couple of business meetings with Wall Street investment bankers, it was able to gain enough funding to continue work on new products. Upon approaching investors on Wall Street, Cray was surprised to find that his reputation had preceded him and that venture capitalists were delighted and willing to lend him the money he needed.

Three years after those meetings, Cray Research debuted the Cray-1, the first in Cray Research's line of supercomputers. The Cray-1 could perform more than 200 million calculations per second. It was released in 1976 and was used to solve large-scale scientific problems. It was sold to government and university laboratories. The Cray-1 outperformed every other computer on the market at that time. While designing the Cray-1, Cray made sure that the entire Cray-1 computer ran at a fast speed. In the past, only the processor had run; now other systems in the computer could keep up.

While Cray worked on the Cray-2, other teams in his corporation delivered the two-processor Cray X-MP, which was a huge success and was then followed by the four-processor X-MP. When the Cray-2 was finally released after six years of development, it was only marginally faster than the X-MP, largely because of the latter's very fast and large main memory. As a result, the Cray-2 sold in much smaller numbers. The Cray-2 processed at 250 megahertz (MHz) with a much deeper pipeline, making it more difficult to code than the short-pipe X-MP. The X-MP was also the first computer ever to provide a huge increase in central memory size. For the next few years, Cray's company came out with more products that had higher computing speed.

In 1980, Cray resigned as chairman and CEO of his growing firm to work in a new laboratory in Colorado Springs, Colorado. Cray stated that he had been distracted by day-to-day tasks and wanted to focus more fully on design. He became an independent contractor, designing even faster machines than those at his laboratory in Chippewa Falls. He took the Cray-3 design with him to his new laboratory. Cray later named his new facility the Cray Computer Corporation. Because Cray was unwilling to use new state-of-the-art technology, the Cray-3 was not commercially successful, although it displayed reliable operation at a 500 MHz clock speed. When Cray's last system, the Cray-4, was almost completed, it operated at a clock rate of 1 gigahertz (GHz), a clock rate that no one else to date had achieved. However, with Cray products declining in sales and market share as the Cold War wound down in the late 1980s and early 1990s, and with the lowered revenues that made it difficult to provide new development funds, Cray Computer Corporation ran out of money. It filed for Chapter 11 bankruptcy on March 24, 1995. Cray then established a new company named SRC Computers, designing massively parallel machines. The new design concentrated on communications and memory performance, two problems that had shelved previous attempts at parallel designs.

Cray liked to use simple tools when working on his inventions, usually only a piece of paper and a pencil. However, he admitted that some of his work required more sophisticated devices. One day, someone told Cray that the Apple Computer company had purchased a Cray computer and was planning to simulate the next Apple computer design using the Cray computer. Cray expressed amusement, as he had been using an Apple to simulate the Cray-3. Cray's selection of people for his projects also reflected fundamentals. Once asked why he often hired new college graduates, he replied that they did not know that what he was asking them to do was impossible, so they tried.

No one in Cray's field has accomplished the constant success that he achieved during his lifetime. He dedicated his entire professional career to the designs and development of large-scale, high-performance systems. He often said that he was put on earth to do that job. Cray died at the age of seventy-one; he had sustained head and neck injuries in a traffic accident on September 22, 1996. Although he underwent emergency surgery to alleviate his injuries, he succumbed to those injuries two weeks after the crash.

SRC Computers continued development after Cray's death and now specializes in reconfigurable computing. The Computer Society of the Institute of Electrical and Electronics Engineers honored Cray with the Seymour Cray Computer Engineering Award. Established in late 1997, the award includes a crystal memento, an illuminated certificate, and a $10,000 honorarium; it is awarded to recognize innovative contributions to high-performance computing systems that best display the creative spirit shown by Seymour Cray.

Personal Life

Cray and his first wife, Verene, had three children, including a son and two daughters. After his marriage to Verene ended in divorce, Cray married Gerri M. Harman in 1976.

Cray valued his time away from his work, although he had little time to enjoy many of his favorite activities because of his dedication to his company. An avid sportsman, Cray enjoyed skiing, tennis, windsurfing, and hiking. He also enjoyed fishing and traveling. He was known both for his down-to-earth demeanor and his unusual hobbies, such as digging a tunnel beneath his home. Although this activity might have been regarded as somewhat eccentric, Cray claimed that his ideas often came to him while he was digging.

Bibliography

Ceruzzi, Paul E. A History of Modern Computing. 2nd ed. Cambridge: MIT, 2012. Print. A narrative history of the computer industry from 1945 to 1995, including Cray in chapters headed “From Mainframe to Minicomputer” and “The Go-Go Years.”

Murray, C. J. The Supermen: The Story of Seymour Cray and the Wizards Behind the Supercomputer. New York: Wiley, 1997. Print. The primary biography of the supercomputer pioneer, wherein the supercomputer itself is as much the focus as Cray—from its origins during World War II to its modern applications in climate modeling. The result provides a solid context for the history of modern computing generally. Includes Cray's final interview.

Scientific American, ed. Understanding Supercomputing. New York: Byron Preiss Visual, 2002. Print. A collection of fifteen articles published in Scientific American between 1995 and 2001 that together provide basic grounding in the technology of supercomputers.

Slater, R. “The Hermit of Chippewa Falls and His ‘Simple, Dumb Things.’” Portraits in Silicon. Cambridge: MIT, 1989. 195–206. Print. A brief biography of Cray in this collection of sketches on pioneers in computer science.

Worthy, J. C. William C. Norris: Portrait of a Maverick. New York: Ballinger, 1987. Print. Biography of the founder of Control Data Corporation, where Cray designed the CDC 1604 and which he left to form his own company.