Murray, Charles J.
John Wiley & Sons 1997
Date finished: 2015-12-30
A corporate history of the Minnesota/Wisconsin-based supercomputing companies, which all turned on the axis of designer and architect Seymour Cray. After the cryptographic programs of WW2 ended, it became clear that computing resources would continue to be useful for the Cold War and the US Navy didn't want the engineering groups to dissolve. They funded Engineering Research Associates, a new company based in a Minneapolis warehouse. ERA was sold to Remington Rand, and after a few years some executives and engineers were dissatisfied enough to leave and form a new company, Control Data. This pattern repeated, as Cray Research was spun out and then again to Cray Computer Corporation. It was a remarkable time, when a central focus of computing was centered in the Midwest instead of Silicon Valley. The history extends from early vacuum-tube models, which were assembled into garage-sized processors, through the introduction of increasingly dense integrated circuits on silicon wafers, to late experiments with gallium arsenide for faster switching. (Gallium arsenide is still thought of as a successor to silicon, but the transition hasn't happened yet.)
It really is a corporate history, though, and its largest flaw is the lack of technical detail. Cray once had to work with unreliable 37-cent transistors and managed to combine them so that they worked more reliably, but this isn't explained. At one point an engineer is tasked with writing a new compiler; we are not told what language it was for or even what a compiler is. There are allusions to the difficulty of multiprocessing, but no explanations. I wished for more, which would probably better explain why Cray was considered so brilliant and so critical to the company.
%T The Supermen %S The Story of Seymour Cray and the Technical Wizards Behind the Supercomputer %A Murray, Charles J. %K computing %G ISBN 0-471-04885-2 %I John Wiley & Sons %P 224pp %D 1997 %@ 2015-12-30