Abstract :
The articles in this special issue on embedded memories present several answers to the eternal problem of trading off memory capacity, speed, and power. Unsurprisingly, each article describes attempts to find the best new memory technology. Test people, however, see memories differently: design people worry about how things work; test people worry about how things break. Although new memory technology is exciting, designers of embedded memories and researchers might be well advised to think more about making memories play better with the rest of the design rather than simply making them faster, denser, and less power hungry.
