By Cleve Moler, MathWorks
In 1994, a floating-point division error was discovered on the new Intel® Pentium® processor.
In this article, published shortly after the public disclosure of the bug, Cleve Moler recounts the people and events that led to mainstream awareness of the defect, his own role in that process, the mathematical basis for the problem, and the difficulties of explaining terms such as relative error and significant digits to the news media. Cleve also explains how MATLAB® was updated to provide a workaround for the Pentium FDIV bug.
Published 1995 - 92024v00