Zum Inhalt springen

Von Neumann's Memory: The Human Calculator

Zusammenfassung

John von Neumann could multiply 8-digit numbers in his head, recite texts he had read once from memory years later, and had memorized the Budapest phone directory as a parlor trick. During the Manhattan Project, he was widely considered the fastest human calculator alive. His ability to hold complex structures in working memory — entire proofs, large tables of figures, lengthy texts — was the cognitive foundation for the intellectual output that defined theoretical computer science, quantum mechanics, and game theory.

The Prodigy

John von Neumann was born in 1903 in Budapest to a wealthy Jewish banking family. His memory was noticed before he could read: he reportedly memorized a page of the phone book at age six as a game. His father’s dinner guests would test him by asking him to recite columns of figures he had seen earlier in the evening; he did so reliably.

He published his first mathematical paper at age 18 (a 1922 joint paper with Michael Fekete on the zeros of polynomials), held doctorates in both mathematics (Budapest, 1926) and chemistry (Zürich, 1926), and was appointed to the Institute for Advanced Study at Princeton in 1933 at age 29 — one of the first six professors appointed, alongside Einstein.

His memory allowed him to produce proofs at speed that colleagues found unsettling. He would listen to a problem, sometimes briefly, and begin dictating a complete proof without notes. Physicist George Gamow described von Neumann as “the only man I know who can count fast enough to keep up with computers.”

The Manhattan Project Calculator

During Los Alamos, von Neumann performed calculations that teams of human “computers” (people whose job was computation) took hours to complete. He mentally evaluated the implosion geometry of the plutonium bomb — a problem requiring iterative solutions to partial differential equations — by holding intermediate results in memory and working through successive approximations. His numbers were used as checks on the mechanical calculator output.

This capacity directly shaped his work on stored-program computers. Having lived the experience of human memory as the bottleneck in computation, he understood intuitively what a computer needed architecturally: a memory that could hold both instructions and data, fast enough to keep pace with an arithmetic unit. The Von Neumann Architecture — the design that all modern computers follow — reflects the perspective of someone who had been, for years, the fastest available computational resource.

The Architecture Named After Him

The First Draft of a Report on the EDVAC (1945), written by von Neumann and widely circulated, described the stored-program concept: a computer that kept its program in the same memory as its data, allowing programs to modify themselves and making reprogramming a matter of loading new data rather than rewiring circuits. The report was written quickly, in days, and reflected von Neumann’s ability to hold the entire design structure in mind as he wrote.

The document was circulated under his name alone, which caused lasting resentment among Eckert and Mauchly at the University of Pennsylvania who had contributed substantially to the underlying ideas. The naming controversy has never been fully resolved; the architecture is called “von Neumann” because the document that described it bore his name.

📚 Sources