IBM ACS System: A Pioneering Supercomputer Project of the 1960's

The showcase IBM effort at high-performance computing in the 1960's has traditionally been considered the IBM S/360 Model 91. That machine well-deserves the attention it has received. In fact, in the field of computer architecture, the decade of the sixties is known for the CDC 6600 and the IBM Model 91, and many modern processors trace features back to those machines.

However, there was another relatively unknown IBM effort that operated in parallel with the deployment of the Model 91. It was launched by IBM Chairman T. J. Watson, Jr., “to come up with something so much better than the [just announced] 6800 as to once more, in the eyes of the public, put IBM far away in the prestige league.”

The project was called Advanced Computing Systems (ACS). It was set up in California specifically to be located far from normal mainframe development on the East Coast as well as to be close to the Livermore National Laboratories and the advanced work on disk drives at IBM’s San Jose facility.

ACS built upon earlier IBM work on Stretch and Stretch-Harvest influenced by the legendary John Cocke, and on IBM’s follow-on “Project Y” at the T. J. Watson Research Center. Most of the Project Y personnel moved to California in 1965 to launch ACS. Many other designers and engineers were recruited and one visitor from Livermore Labs commented that he “had not seen such a high concentration of talent since the Manhattan Project.”

The ACS architecture incorporated innovations that remain important today, including multiple out-of-order dynamic instruction scheduling, multiple condition codes, a decoupled branch architecture, and instruction pre-fetching. Advanced ECL circuits and optimizing compilers were also crucial parts of the plans for ACS. On reflection it appears to have been the first “superscalar” design, and yet its story remains virtually untold to this day.

Please join us to hear former project members describe the exciting atmosphere of the ACS team and the computer design innovations that ACS created.

This event will present two consecutive panel discussions:

Panel 1- Early ACS / also ACS-1 architecture and SW

- Fran Allen, IBM Fellow, Moderator
- Lynn Conway, Emerita Professor of EE/CS, University of Michigan
- Brian Randell, Emeritus Professor of CS and Senior Research Investigator, Newcastle University

Panel 2- Later ACS and dispersion of folks elsewhere after
1969 cancellation / also ACS HW

-Russ Robelen, Moderator
-Bill Mooney
-John Zasio