APPM Department Colloquium - Joseph F. Grcar

Joseph F. Grcar, Center for Computational Science and Engineering, Lawrence Berkeley National Laboratory

John von Neumann and the Origins of Scientific Computing

Scientific computing as we understand it today began to be practiced around the time of Carl Friedrich Gauss in the form of astronomical calculations based on Isaac Newton’s laws, and geodetic calculations for cartography. Of course, the invention of modern computers (those digital, electronic, and programmable) in the 1940’s represented a paradigm shift in what could be achieved through calculation. From his wartime military duties John von Neumann acquired what he described as an “obscene” interest in mechanized calculations. No one was better situated that he to understand the advances that could be realized, but also the whole range of technical obstacles that had to be overcome. Thus von Neumann and his principal collaborator in this work, Herman Goldstein, largely reinvented scientific computing and created computer science in the late 1940’s.

This talk will survey the development of scientific computing from Gauss to the first electronic calculations in the early 1950’s. The emphasis is not on von Neumann himself but rather on the conceptual changes that occurred. Von Neumann arises naturally in the story because many of those changes can be traced to him. If time permits it is possible to sketch von Neumann’s biography and the subject to which both he and Gauss contributed: “Gaussian” elimination.