Seymour Cray

Seymour Roger Cray was a pioneering American electrical engineer and computer scientist.

He is widely recognized as the "father of supercomputing" because he helped create the entire supercomputer industry. Cray designed a series of machines that remained the fastest in the world for several decades. His influence on the field was so profound that experts compared him to Thomas Edison. He possessed a unique ability to envision high-performance computing capabilities that others thought were impossible. His work focused on pushing the absolute limits of speed and efficiency in digital technology.

Cray’s approach to design was based on a fundamental understanding of system bottlenecks. He believed that a fast central processing unit, or CPU, was not enough on its own. If the data could not move quickly enough to reach the processor, the CPU would be "starved" of information. To prevent this, he focused on maximizing I/O bandwidth, which is the speed of data moving in and out of the system. In his CDC 6600 design, he solved the problem of "imprecise interrupts" that had hindered competitors like IBM. He did this by using ten peripheral processors, which were small, built-in mini-computers. These processors handled all data transfers to and from the central memory, allowing the main system to run much faster.

Cray’s interest in engineering began in his childhood in Chippewa Falls, Wisconsin. Born in 1925, he used an Erector Set at age ten to build a device that converted paper tape into Morse code. During World War II, he served as a radio operator and worked on breaking Japanese naval codes. After the war, he attended the University of Minnesota. He earned a B.Sc. in electrical engineering in 1949 and an M.Sc. in applied mathematics in 1951. These technical foundations allowed him to become an expert in digital computer technology during his early career.

His professional career moved through several major companies. He first joined Engineering Research Associates (ERA), where he designed the ERA 1103, the first commercially successful scientific computer. After ERA was acquired, he helped found the Control Data Corporation (CDC) in 1957 with William Norris. At CDC, Cray designed the CDC 6600, which became the first commercial supercomputer in 1964. The 6600 was nearly three times faster than the next fastest machine, the IBM 7030 Stretch. This success established Cray as a leader in high-performance computing.

Cray eventually left CDC due to disagreements with management and a desire for a quieter work environment. He wanted to avoid the interruptions of middle managers to focus on pure design. He moved his laboratory to his hometown of Chippewa Falls to create distance from the corporate office. In 1976, he founded Cray Research, a new company that also operated out of Chippewa Falls. His first major product from this company was the Cray-1. This machine was a massive financial and technical success, with over 80 units sold. One system was sold to the National Center for Atmospheric Research for $8.8 million.

Throughout his later career, Cray continued to experiment with new technologies. He worked on the Cray-2, which featured very fast and large main memory and a deep pipeline. Later, he formed the Cray Computer Corporation to pursue the Cray-3 project. For this machine, he attempted to use gallium arsenide semiconductors instead of traditional silicon. He wanted to achieve a tenfold increase in performance through this new material. While the Cray-3 was a technical achievement, it faced difficult competition from massively parallel machines. This era marked a shift in how the industry approached high-speed computing.

Seymour Cray's legacy is found in the very architecture of modern high-performance computing. He proved that a computer is a complete system where every part must work in harmony to achieve speed. His transition from a young boy building Morse code devices to the "Wizard of Chippewa Falls" shows a lifelong dedication to engineering. The industry he helped build continues to use the principles of maximizing bandwidth and efficient data movement that he pioneered. His work provided the computational power necessary for major scientific advancements around the world.