Marty's Singularity Blog
The Singularity is happening now!
December 20, 2010 at 5:17 am
filed under Computing, Singularity
Tagged Computing, Singularity
We talked some about Moore’s Law in the previous post but it is definitely worth a more detailed look. Gordon Moore was one of the founders of Intel and thus one of the early pioneers of integrated circuit technology. Integrated circuits contain multiple semiconductor devices that combine to form an electronic circuit. The invention of the integrated circuit in 1958 is credited to Jack Kilby who won the Nobel Prize in Physics in 2000 for this invention. Kilby’s circuit used the semiconductor germanium. Robert Noyce, another one of the founders of Intel, working independently created his own version of an integrated circuit. Noyce’s device was based on another semiconductor, silicon. Silicon has become far more commonly used in integrated circuits because it is both much less expensive and much easier to work with than germanium.
Moore made the observation a few years later which came to be known as Moore’s Law. His observation was that the number of transistors in an integrated circuit was doubling every two years or so. This was occurring mainly because engineers were leaning how to make semiconductor devices like transistors smaller and smaller. At the same time they were also learning how to make them faster, plus the reduced dimensions meant that signals arrived faster because they had less far to travel. The end result was that the amount of computation available increased even faster than the feature size decreased.
In the early 1960’s the state of the art for integrated circuits was approximately 100 devices per circuit. Today there are many integrated circuits with over a billion devices such as multi-core microprocessors. This shows the power of exponential increases and even more basically the power of human ingenuity.
Ray Kurzweil made a further observation which he documents in The Law of Accelerating Returns. He found that Moore’s Law is actually part of a much larger phenomenon which stretches back over a century. Starting with punched card mechanical devices in 1890, then relay devices in WWII, followed by vacuum tube computers in the late 1940’s, and discrete transistor devices in the late 1950’s, and finally integrated circuits starting in the 1960’s, there has been a steady increase in price/performance of computing devices over five different paradigms. And as we shall see later, there are several candidates for the next paradigm once we reach the limits of semiconductors.
you must be logged in to post a comment.
no comments
RSS / trackback