Was Moore's law, now something else, parallelism

[Richard Pieri]

On Jul 13, 2010, at 1:41 PM, Bill Bogstad wrote:
> 
> I don't know exactly how much faster i3/i5/i7 CPUs are on single
> threaded code vs. the original core 2 CPUs.  However, everything I've
> read in the last ten years, supports the thesis that single-core
> performance increases are not occurring at the same rate as in the
> past.

It's simply an evolutionary dead-end.  Intel took that concept to one practical limit with the Pentium 4 and ran into a brick wall.  Not because of cooling.  Because they couldn't get data to flow to go any faster.  They reached the practical limit of the few long pipelines architecture philosophy... and abandoned it for the many parallel short pipelines architecture derived from Pentium III and Pentium M, the same architecture philosophy used by IBM in the POWER 4 and POWER 5 architectures.

The past five years have seen the chip industry undergo a drastic change.  The focus has shifted from raw processing power to performance per Watt and more dedicated purpose packages bundling CPU, GPU and other systems.  The result has been an explosion of inexpensive portable and mobile electronics.  It is these, the iPads and iPhones, the Droids and Blackberries, the netbooks and Kindles that are the driving force in the computer industry today.

--Rich P.