— Comment by ptmmac on Horace Dediu’s ‘The Price Is Right’
The Manufacturers may finally be getting a break in price drops because Moore’s Law is no longer ruling the roost.
The factor not included in the data is the lack of increase in processor speeds since those are limited by frequency. If you were graphing the quality of the transistor per dollar, rather than the price per “machine” you would have seen a very different story. Customers are no longer buying machines that offer a compelling improvement in processing speed. A 4 year old computer today is now more analogous to a 4 year old TV rather than a 4 year old computer from 1999. The power envelope is the limiting factor not the geometry that drove the earlier heyday of Moore’s Law. One result of this is low powered devices were not limited by the same problems and actually were able to ramp frequency while computer’s could not. A cell phone now offers much of the functionality of a $400 computer. Sales of traditional computer’s will continue to drop because the product turnover is dropping, and there is more competition from other kinds of devices. Dropping demand for computer’s definitely means there is less investment money to pay for the next processor node.
What I am curious about is why has the obvious slow down in increases of computer speeds been so ignored by the press? There has been more technology innovation than in the 90’s, but we have seen less growth in productivity. Perhaps if we could name the problem, the solution would get more attention. Two types of processors have continued to improve while CPU’s of average PC’s have stagnated. One is the graphics processor because of it’s ability to harness more improvement from running multiple cores, and the other is the low power computing devices. Area’s economic productivity have followed in the wake of each of these two exceptions. Petroleum exploration has benefited from the ability to find oil that was not practical to extract because this is processed with multicore machines with large graphics processors. Cell phones are another growth area with smart phones leading the way towards new business innovation. Apple’s big bet on ARM makes much more sense if we look at what happened to the 5th generation PowerPC chip and how it failed to match the growth in processor speeds that Intel was producing.