OT possibly Celeron benchmarks

How about doing an apples-and-apples comparision? Windows 98 = Oranges. Windows XP = Apples. No benchmark comparision has validity if the operating systems are different. Even if one XP version was running with SP1 and the other with SP2, the operating systems would be considered different by those of use with decades of hardware and software benchmarking experience. Finally, one would also control the precise numbers of memory resident programs and services loaded by the operating system at startup.

Also, Windows 98 does not have preemptive dispatching quite as effective as the NT family's. As a consequence, SuperPi runs virtually uninterrupted on Windows 98, but gets interrupted regularly by the XP operating system which looks around to see if there are other tasks which need processor time slices.

I am not defending the Celery here. It is not well-suited for heavy duty math intensive computations. But the benchmark methodology as described is highly flawed... Ben Myers