I develop and publish one of the programs profiled in this article (Clear Speed). I'm currently corresponding with the author for a follow up piece he's going to run (hopefully). The CPU used in the Pocket PC (PXA250) is completely clock configurable as far as memory speed and core multiplier. Just imagine an unlocked Pentium with memory speed settings in the CPU itself. It sucks that Intel gets a free ride on the whole "overclocking" issue. We all know what goes on between chip manufacturing and marketing.

Some of my main issues:

- Intel says a heat sink isn't necessary on the PXA250 so I can't see how they can claim heat is an issue.

- Can the PXA250 actually be damaged by software, even inadvertently? How does Intel feel about designing a CPU this way and then documenting how to do it?

- Assuming that somehow each CPU off a wafer has performance characteristics that vary by a factor of 2 (200 mhz to 400 mhz) how is Intel's testing any different than just running the CPU at a speed and seeing if it works? Does Intel suggest that their processors degrade over time and can fail at some future point? How do they account for such poor manufacturing techniques that would result in such a variation in the parts produced?

Thanks for listening!

Hayes