Jump to content

Cray Computer Corporation

From Alfawiki
Seymour Cray
Born (1925-09-28) 28 September 1925 (age 101)
Chippewa Falls, Wisconsin, U.S.
Died5 October 1996(1996-10-05) (aged 71)
Occupation(s)Electrical engineer, computer architect
Known forDesigning supercomputers, founding Cray Research

Seymour Roger Cray (September 28, 1925 – October 5, 1996) was a pioneering American electrical engineer and architect of supercomputers, renowned for designing some of the fastest computing machines globally and establishing Cray Research. Often hailed as "the father of supercomputing," Cray is recognized for founding the supercomputer industry. Joel S. Birnbaum, former Hewlett-Packard CTO, highlighted his significant contributions to high-performance computing, while Larry Smarr, director of the National Center for Supercomputing Applications, referred to him as "the Thomas Edison of the supercomputing field."

Early life

Seymour Cray was born in 1925 in Chippewa Falls, Wisconsin, to civil engineer Seymour R. Cray and Lillian Cray. His fascination with science and engineering emerged early; by age ten, he had constructed a device that converted punched paper tape into Morse code signals using parts from an Erector Set. His family’s basement functioned as his personal workshop.

After completing high school at Chippewa Falls in 1943, Cray was drafted into World War II as a radio operator. He served in both European and Pacific theaters, contributing to efforts to decipher Japanese naval communications. Following his military service, he obtained a Bachelor of Science in electrical engineering (1949) and a Master of Science in applied mathematics (1951) from the University of Minnesota.

Engineering Research Associates

In 1950, Cray began working at Engineering Research Associates (ERA) located in Saint Paul, Minnesota. ERA originated from a former United States Navy laboratory dedicated to developing codebreaking machinery and continued this focus alongside various other engineering endeavors.

Cray rapidly became proficient in digital computer design, playing a key role in creating the ERA 1103, which was the first scientific computer to achieve commercial success. He remained with ERA through its acquisitions by Remington Rand and later Sperry Corporation, during which ERA became UNIVAC’s scientific computing division.

Control Data Corporation

Unhappy with ERA’s direction under Sperry Rand management, Cray partnered with William Norris to establish Control Data Corporation in 1957.

By 1960, Cray had finalized the design of the CDC 1604, an enhanced and affordable successor to the ERA 1103 that delivered remarkable performance relative to its cost. While the CDC 1604 was entering production, Cray pursued additional projects. He initially developed an advanced CDC 3000 series tailored for business data processing but found the market less attractive due to budget limitations and minimal scope for innovation. Consequently, he redirected his efforts toward high-performance computing with the CDC 6600.

Although its hardware components were not state-of-the-art individually, Cray prioritized overall system efficiency beyond mere CPU speed by maximizing input/output throughput to avoid processor bottlenecks. He famously remarked, "Anyone can build a fast CPU. The trick is to build a fast system."

The CDC 6600 emerged as the first commercially available supercomputer, significantly outperforming rivals despite its premium price tag. For customers prioritizing maximum speed, it stood unrivaled. Competitors such as IBM attempted to challenge this lead with models like the IBM 7030 Stretch, but struggled due to design issues including "imprecise interrupts." Cray overcame these challenges by substituting I/O interrupts with polled requests managed by ten peripheral processors—specialized mini-computers dedicated to handling all data transfers. Its successor, the CDC 7600, further boosted performance by a factor of five.

In a 1963 Business Week profile introducing the CDC 6600, Cray expressed a principle relating performance to cost that has often been mistakenly attributed to Herb Grosch:

CDC's Chippewa Falls Facility

During this timeframe, Cray became increasingly dissatisfied with the interference from CDC’s management. He insisted on a tranquil workspace with minimal supervision; however, as the organization grew, middle management often disrupted him—frequently using his presence primarily as a sales asset to engage potential clients.

To preserve his concentration, Cray chose to move from St. Paul to a location distant enough to prevent unplanned visits and calls, yet still accessible for board meetings. Norris endorsed this relocation by setting up a new laboratory on Cray’s own property in Chippewa Falls. Additionally, concerns about nuclear conflict influenced the choice, as Cray’s nearby residence featured a substantial fallout shelter.

The Chippewa Falls laboratory was established midway through the 6600 project without causing any delays. Following the shipment of the 6600, the subsequent CDC 7600 was developed there, achieving peak performance speeds ten times greater. However, the next CDC 8600 project encountered significant obstacles and ultimately ended Cray’s successful tenure at CDC in 1972.

Despite the commercial triumphs of the 6600 and 7600, their development nearly drove CDC to financial ruin. The problematic 8600 project led Cray to suggest initiating a new start, but Norris rejected this due to financial concerns and preferred to concentrate resources on the STAR-100 initiative. Unwilling to accept these limitations, Cray departed from CDC.

Cray Research

The separation was conducted amicably, and one year later Cray established Cray Research on the same Chippewa property, with Norris contributing $250,000 in investment. Similar to CDC, research and development remained in Chippewa Falls while corporate headquarters were located in Minneapolis; however, unlike CDC, manufacturing also occurred at Chippewa Falls.

Initially facing uncertainty regarding its strategic direction and funding availability, Cray Research leveraged Seymour Cray’s esteemed reputation to attract investors. Wall Street financiers were eager to provide capital for ongoing development efforts.

After several years dedicated to R&D, Cray Research introduced its inaugural product—the Cray-1—in 1976. The system design prioritized overall speed rather than focusing solely on processor speed. It outstripped nearly all competitors at the time, including CDC’s STAR-100 and the specialized ILLIAC IV.

The first unit (serial number 001) was loaned to Los Alamos National Laboratory in 1976; that summer saw the sale of the first complete system to the National Center for Atmospheric Research for $8.8 million. Initial sales forecasts anticipated about a dozen units based on prior CDC models; however, over 80 Cray-1 systems were eventually sold. This success earned Seymour Cray the moniker "The Wizard of Chippewa Falls."

Later products encountered more formidable challenges. While Cray worked on developing the Cray-2, competing teams launched the two-processor Cray X-MP, which achieved considerable success and was later expanded to four processors. Released after six years of development, the Cray-2 offered only modest speed gains over the X-MP due to its sophisticated memory architecture but sold fewer units because of its programming complexity.

When work began on the Cray-3, Cray found managing daily operations increasingly distracting and stepped down as CEO in 1980 to concentrate solely on design as an independent contractor. In 1988, he relocated the Cray-3 project to Colorado Springs.

By 1989, confronting familiar difficulties with competing projects and constrained budgets—the Cray-3 struggling while an X-MP upgrade advanced—management chose a more conservative investment path that resulted in the development of the Cray Y-MP.

Cray Computer Corporation

The Colorado Springs research facility was separated from Cray's main operations to establish Cray Computer Corporation, which assumed responsibility for advancing the Cray-3 initiative.

The Cray-3, operating at 500 MHz, marked the second significant setback for Cray. Seeking a tenfold boost in performance, Cray chose to utilize gallium arsenide semiconductor technology rather than relying on his customary, well-established methods. This approach necessitated the creation of all components from scratch, including the semiconductor chips themselves.

Despite numerous obstacles, the team succeeded in delivering the initial Cray-3 unit to NCAR on 24 May 1993. This system functioned primarily as a prototype for debugging purposes. Meanwhile, competing massively parallel architectures entered the marketplace, offering superior cost-to-performance ratios. In reaction, Cray initiated development on the Cray-4, aiming to achieve 1 GHz speeds and outperform rivals irrespective of expense.

By 1995, no further sales of the Cray-3 had been recorded, and the diminished post-Cold War demand for high-cost supercomputers like the Cray-4 impacted market viability. Confronted with financial strain, Cray Computer Corporation filed for Chapter 11 bankruptcy protection on 24 March 1995.

SRC Computers

Initially skeptical of massively parallel systems, Cray maintained that a single high-speed processor outmatched multiple slower units. However, by the mid-1990s, improvements in compiler technologies and parallel system designs challenged this perspective. To pursue massively parallel computing optimized for communication and memory bottlenecks, Cray founded SRC Computers. Following his untimely death in a vehicular accident, SRC continued its work with a focus on reconfigurable computing technologies.

Technical approaches

Cray’s design philosophy centered on two critical elements: efficient heat dissipation and accurate signal timing.[1] His machines incorporated integrated cooling solutions ranging from coolant channels embedded within chassis structures to circuit boards thermally bonded to metal plates and immersion cooling techniques. He innovated by synchronizing computer operation with cooling system status to prevent hardware damage during power interruptions.

To mitigate clock skew, Cray meticulously ensured that all signal pathways maintained equal electrical lengths, occasionally routing traces multiple times across circuit boards. Applying Maxwell's equations, he accounted for radio frequency phenomena that influenced signal velocity and timing precision.

Remarkably, Cray favored simple design tools—using pads of 8½″ × 11″ lightly ruled -inch quadrille sheets—instead of relying on sophisticated CAD programs.[2]

List of computers

Cray played a key role in developing these significant computer systems:

Personal life

In 1947, Cray wed Verene Voll, whom he had known since childhood. Verene was a nutritionist and daughter of a Methodist minister—mirroring Cray’s maternal background—and together they had three children before divorcing around 1978. Subsequently, Cray married Geri M. Harrand. He was also the grandfather of LGBTQ rights advocate Andrew Cray.

Known for shunning public attention, Cray pursued hobbies such as skiing, windsurfing, tennis, and excavating a tunnel beneath his residence—a pastime he credited with inspiring creative problem-solving at work. Stories shared by then-CEO John A. Rollwagen illustrate Cray’s distinctive management style characterized by brief planning documents and status updates focused on ambitious objectives with minimal elaboration.

Cray succumbed to injuries sustained in a rollover collision caused by a reckless driver while merging onto Interstate 25 near the United States Air Force Academy in Colorado. He died on October 5, 1996, approximately two weeks after the accident and shortly after turning 71 years old.

Posthumous Recognition

Since 1997, the IEEE Computer Society has presented the Seymour Cray Computer Engineering Award to acknowledge groundbreaking achievements in the realm of high-performance computing systems that embody the inventive spirit exemplified by Cray.

See also

References

External links

  1. ↑ Customer presentation by Seymour Cray, c1979
  2. ↑ Science, 1978
Retrieved from "https://www.alfawiki.org/index.php?title=Cray_Computer_Corporation&oldid=23159"