The Control Unit

The control unit directs and controls the activities of the internal and external devices. It interprets the instructions fetched into the computer, determines what data, if any, are needed, where it is stored, where to store the results of the operation, and sends the control signals to the devices involved in the execution of the instructions.

The speed at which a computer can carry out its operations is referred to as its megahertz rate. A CPU that runs at one megahertz would complete one million operations per second. Many of the computers that you use on this campus are in the 200 to 400 megahertz rate. By the end of the year 2002, 2000 megahertz (e.g., 2 gigahertz) rates became available. In line with the trends noted above, ever faster systems can be designed. Because other technologies must work with the CPU, computers which arc rated with a higher megahertz rate than others do not necessarily run faster. The best test of a computer's speed is to get out your stop watch and time basic operations such as saving a file or running an animation, then run the same program on a different computer and compare.

It is also possible for many CPUs to share computing responsibilities in one computer. Soon new computer operating systems will force personal computer buyers to also make decisions about how many CPUs they want in their computer, perhaps one for every application. The world's fastest computer is a supercomputer in Yokohama that runs at over 35 trillion

operations per second for the Earth Simulator Project which does climate modeling, merging the work of over 5, 000 integrated CPUs. The term supercomputer is a relative term referring to the fastest computers available on the planet in any given year. Nevertheless, the room size computer or computer complex has never really gone away. Today's supercomputers still need much of the space and cooling equipment that the original room sized Eniac computer did.

In spite of their rapid growth in capacity, today's computers based on electrons may one day become the dinosaurs of the early history of computer technology. Think of today's computers as being in the steam engine age of the history of the automobile that came before today's internal combustion engines. The next generation of computers may be based on photons (light beams) and run hundreds of times faster than today's fastest computers. The patents for photon transistors were first filed in Scotland in 1978. One could surmise from this alone that the next fifty years of central processing unit technology will be as dramatic as the last.

Computer chip technology continues to become even more sophisticated, adding more features to the chips than just transistors storing ones or zeros. Companies are now developing three-dimensional structures on the chips at nanotechnology scale. Nanotechnology involves work with molecular or macromolecular levels in the length scale of approximately 1—100 nanometer range (1 billionth of a meter is a nanometer). One application of this capacity would be with drug companies seeking to speed up and drop the cost of research. They would like to run entire sets of research lab procedures on chips using microfluidics. Microfluidics refers to moving microscopic liquids on chips. In addition to the standard transistors holding ones and zeros, these chips are built in layers which contain miniature valves, pumps and channels that act within a chip as fluidic circuitry.

Another approach used by computer centers is blade computing. From 20 to 300 cell computers are packed in a rack like blades in a knife rack, hence the name blade computing. This concentrates the number of computers that can fit into one space and greatly reduces electrical power and management costs. Link up a cluster of really fast processors and the device becomes a supercomputer. That is, the network is the computer.

Chips