RE: Paul Graham's PyCon Keynote & The Programmer's Apprentice

["Christopher Barber" <address@hidden>]

> From there, Paul said that if Moore's law holds, computers will be
> 78 quintillion times faster in 100 years (with similarly large
> increases in memory and storage capacities).

Of course, Moore's law will not hold.  The rate of doubling of the amount of
transisters you can fit on a chip has slowed since Moore's original observation.
It is quite possible that the growth could hit a brick wall at some point in the
future.  Of course, computers could become faster for other reasons, such as
increased parallelism.

It should be noted that increases in the speed of memory have not kept pace with
processor speed, otherwise we would not have to resort to the complex
multi-level memory caching schemes, which are now common.  Furthermore, while
communication bandwidth has increased enourmously in recent years, latency has
not improved much at all, and is ultimately bounded by the speed of light.  Thus
it will be latency rather than CPU power that will increasingly be the bounding
factor in computer performance.

- Christopher