The Big Leagues, the Supercomputers

The first time we saw the 'Control Data 6600" we traveled to the University of Texas at Austin just to look at it. That was in the days when all computers where shown off behind glass walls. Nobody would talk to us, we just looked at it.

The opportunity came in 1970 to join a startup scientific data processing company in San Antonio, who purchased the CDC 6400, a slower version of the 6600, (only one CPU functional unit instead of 10 parallel units on the 6600). We were hired as a computer operator and fired a few months later for tinkering with the computer causing the operator to perform a deadstart (now called boot) and inconveniencing the clients (maybe 20-50) who where using it at during much of the day.

We where then rehired as a trainee systems programmer, maintaining the operating system using Assembly language. We had the source code for every line of every program ( a few million lines of code).

The company didn't last very long, but we went on to Control Data, then another system programmers gig specializing in Data Communications on the CDC machines, and in 1974 at the age of 23 we decided to start our own company to replace the slow multiplexors Control Data used for Data Communications with an Intelligent front-end based mostly on existing Mini Computer hardware.

We did design a special high direct channel interface from the CDC 6000 series to the Modcomp mini-computers we used. Nobody would do it for us so we got into the hardware manufacturing business.

The Control Data 6600

CDC 6600 CPU 131K 60 bit word Core Memory, 1 or 2 CPU's, 10 PPU's,
Two 60MB 12 head parallel Hard Disks (in one cabinet, the 6638) & miscellaneous. About $15,000,000 at the time. Pictured below is a similar setup.

A Logical Diagram of the CDC 6600

CDC 6600 logical.tiff

The core memory modules of a CDC 6600

4k x 12 core memory 4096 12 bit words of memory, 5 would be stacked to make a 60 bit word, cost about $20,000 each the computer had 170 of these.