COMP 370 – Microprocessors
Moin Fatehali Khaki
The brains of a computer is the semiconductor chip. Much of the progress over the last 35 years in making computers faster, smaller and cheaper has been a numbers game, squeezing ever more transistors and other electronic devices onto this postage-stamp-- sized piece of silicon. Today's PCs pack tens of millions of transistors onto a chip, each transistor as small as a few hundred nanometers (billionths of a meter). But continuing this miniaturization will not be easy-or cheap. Some experts predict that by 2010 it will cost from $30 to $50 billion to build a manufacturing facility to fabricate even smaller chips. Then again, a number of physicists contend the price tag won't matter, because silicon-based devices are fast reaching their fundamental physical size limits. [Molecular memory; Rotman, David]
A large portion of the silicon area of many contemporary digital designs is dedicated to the storage of data values and program instructions. More than half of the transistors in today’s high-performance microprocessors are devoted to cache memories and this ratio is expected to further increase. The situation is even more dramatic at the system level. High performance workstations and computers containing several Giga-bytes of information and to implement this information on a device needs a large amount of memory. Memory is required in every digital device, be it an MP3 player or an MPEG-4 player. The demand for ‘non-volatile’ memory has expanded drastically.
There are means of storing Boolean values based on either positive feedback or capacitive storage. While semiconductor memories are built on the same concepts, the simple use of a register cell as a means for mass storage leads to excessive area requirements. Memory cells are therefore combined into large array which minimizes the overheads by peripheral circuitry and increases the storage density.