Safety in Numbers: Creating and Contesting the Los Alamos Approach to Supercomputing, 1943 to 1980

Abstract

Since its origins during World War II as the primary R&D site of the Manhattan Project, the Laboratory at Los Alamos has pushed the limits of computational technologies and systems. In response to its heavy computing demands, and the infancy of the computing industry during the early Cold War, the Theoretical Physics (T) Division at Los Alamos, which provided computing resources for all divisions at the Lab, played an active role in researching and developing the computing technologies that its scientists required. T Division's hardware and software R&D effort formed the basis of the distinctive Los Alamos approach to computing, which placed computing on an equal footing with other scientific research, and encouraged the formation of local, world-class expertise in computing technologies essential to the Lab's mission. As particular computing technologies matured, and vendors were able to provide commercial products adequate for the Lab's needs, T Division shifted its R&D focus to other, less mature lines of computing research. This approach provided Los Alamos computer users with technologies that were otherwise unavailable or unsuitable from outside sources, and ensured that T Division could rapidly adapt the Lab's computing operation to unpredictable changes in the technical and strategic demands of the Cold War. The Los Alamos approach to computing remained largely uncontested until its transfer in 1968 with T-Division personnel to the Lab's new Computing (C) Division, where the inherited approach became a point of contention for a growing number of weapons-program computer users who opposed the Lab's basic computer-science research and C Division's operational independence. Radical changes to the Lab's administrative structure over the 1970s upended the power dynamics between the supporters of the traditional Los Alamos approach to computing and those who sought to bring the Lab's computing resources under the direct control of the weapons program. A focal point for critics of C Division's R&D policies, the local development of a cutting-edge operating system for the iconic Cray-1 supercomputer became the final battle ground over the decades-old Los Alamos approach to computing.