case study1

CASE STUDY #1:
PAST AND PRESENT TRENDS OF COMPUTER ARCHETICTURE
Past:
The computer started with its large, heavy machines composed of thousands of vacuum tubes.and has a development of the transistor created the next evolution in computer architecture, the microchip which is currently used in the generation of computers. Like its vacuum tube predecessor, this architecture of utilizing transistors, can only go so far. At this rate each switch will eventually become the size of an atom. When this happens the laws of quantum mechanics must be used. A new evolution in computer architecture will need to be developed to handle the unique laws of quantum mechanics. This architecture is already being developed and is called a quantum computer.
Quantum computers work in a rather distinctive way. Instead of using traditional bits, they use quantum bits, or qubits. Qubits are particles that can take on the unique states required for quantum computing. The best way to understand how a quantum computer works is by example. A basic example is to take a register composed of 2 bits. Using a classic register, these two bits can have a value of 0,1,2, or 3. Now using a quantum register with two qubits, the register can have a value of 0, 1, 2 and 3.
Remarkable developments in semiconductor technology enabling the implementation of ideas that were previously beyond the computer architect's grasp. Ideas like multilevel memory hierarchies, pipelining, multiple instruction , and speculative execution are just a few examples of architectural innovations that have become commonplace in high-end computers.
The computer "revolution" has been driven by the remarkable growth in semiconductor technologies. The common denominator has been the constant reduction in the size of electronic and magnetic devices that can be manufactured inexpensively. If we take our time horizon to be the next decade as an appropriate strategic point in the future, we expect to see feature sizes of 0.05 microns and chips with 100 million devices operating at several gigahertz. Technical obstacles need to be overcome for this progress to occur, but it is our expectation that the overwhelming resources in industry and academia that are deployed in the computer industry.
The two major obstacles that we see are the issues of power and memory latency. Low-power design is important because portable computing will become ubiquitous in the next five years. Even in the realm of high-performance computing, power issues are important because power consumption is linearly related to clock frequency.