The Evolution of Computer Architecture: Past, Present, and Future by Carol Ho
“Every fact that is learned becomes a key to other facts.” – Edward L. Youmans
To create a better future, we must first understand our past, what it has to offer, and where it could take us. In the field of computer architecture, this better future means advancements in (but not limited to) performance and cost. Through the years, the performance and cost has been redefined to create the next benchmark based on current technology. Let's keep raising the bar! It's human nature to want more. This hunger is fed by a continuous demand for faster and cheaper devices. The demand creates a need for businesses who in turn fund our research and allow us to keep evolving technologically. Now, let's hop in my DeLorean and travel back to 1981! Image processing demands fueled ideas in parallelism (MIMD/SIMD) and reconfigurable architectures (PASM) . Now, we'll catch up to Marty McFly in 1985! Architects working with signal processing applications realize that real-time calculations are indeed a need. The increasing algorithm complexity drove the need for more performance. More research is put into novel architectures . As more time progressed, so did electrical improvements. The architectural improvements did not advance as quickly to make use of the new technology [12,16]. Breaking the bottleneck issue continues to be a problem. Working around physical pin bandwidth limitations brought a cause to creating more integration on a chip . The multiprocessor on a chip idea contributed to higher performance . Mobile devices emerged and so did the demand for embedded platforms drove new architectures to better utilize them . Chip sizes began to shrink, leading more to embedded applications . Presently, the need for smaller devices continues to grow. The desire for embedded use leads to more research for smaller microprocessors. Of course, cheap is always good, so the number of...
Please join StudyMode to read the full document