Page 1 of 3
We tend to think that the phenomenon of engineers and scientists being at the top of a company is something that started with Bill Gates, Steve Wozniak or Gary Kildal. But this just isn’t the case. Even back in the days when IBM was the single most important computer company, it was possible for one of its engineers to escape and make an impact that disturbed even Big Blue.
born 16 November 1922
Gene Amdahl is a physicist who got into computers becasue he wanted to work out something complicated. In 1950 he was asked by one of his professors to calculate whether the nuclear strong force was really enough to hold together a nucleus. For thirty days Amdahl slaved over a sliderule and a mechanical desk calculator to provide only two more significant digits to the solution. This is the sort of experience that drove many a scientist to become a computer pioneer!
Amdahl had taught physics, and then electronics, to the navy during the war and was well able to understand what was going on in the new field of electronic digital computers. In 1951 he started to build the Wisconsin Integrally Synchronized Computer (WISC). It had a floating point unit and could execute four instructions in parallel. This was very advanced for the time and indicated Amdahl’s flair for digital electronics and systems. The machine itself wasn’t finished for many years because it was used as a training project in electronics but it served its purpose and Amdahl got a PhD out of it - but in theoretical physics!
Wisconsin Integrally Synchronized Computer
In 1952 he joined IBM and ended up in the New York plant where the IBM 701 defence calculator was just being finished. It was already clear that something more powerful was needed and so he was set to work, as chief planner and project engineer, on its replacement, the 704.
The IBM 704 of 1954
The IBM 704 was a remarkable machine. It had three addressable index registers and a floating point hardware unit which made programming it very much easier. The 704 was an ideal scientific computer. It worked in binary and it was, for the time, fast. IBM sold a lot of 704s, partly because of its overall design and partly because it was the machine that Fortran was developed for. Without the three index registers,creating a compiler would have been much more difficult.
After the 704 Amdahl designed the 709. But after that it looked as if IBM wasn’t going to let him get another machine to himself. Machines were getting bigger all the time and designing them was becoming increasingly a team effort. When he was denied a leading roll in IBM’s biggest project of the time - the STRETCH computer later to become the 7030 - he left. In 1955 he found himself working for a small company called Ramo-Wooldridge preparing proposals and planning the RW440 process control computer.