Hi Nick,
It may be a 64bit CPU, but that doesn't necessarily mean that it is using 64bit Windows; yet. There are driver issues (like some being beta) with some systems, so using 64bit Windows right now is not for the faint hearted. The DataCAD hardlock drive
may now be 64bit compatible, but the one person I know of who tried to use 64bit Windows had to give up a little while before this (possible) update became available.
As for the Xeon -- the problem is that there are so
many variations on the specs on all the Intel chips that it is hard to know much by the name only! What is the model number? They have taken to having a 300 series, 500 600 series, and a 700 series (gee, where did they get that from, I wonder?), and there are a myriad of subtle differences in the way these are set up. Bus speeds, cache sizes, and the like are all over the place, and they also mix in different architectures: the P4 and all it's derivatives are in the 300, 500, 600, and 700 ranges, and the CPU's that are based on the Pentium M are in the 700 series, too. And now they also have the dual core models -- and only a few of all these are 64 bit capable.
AMD, OTOH has a much simpler and more understandable line of products: there are Sempron (silly name, I know) which is at the bottom, the Athlon 64 which comes with different cache sizes, but has it's relative performance taken into account in the model number, the Athlon 64 FX which is essentially an Opteron at the fastest clockspeed, the Athlon 64 X2 (dual core, natch), and the Opteron, which is the "server" chip. Ironically, the model number system that AMD is using currently is based on the performance relative to the P4 -- but the fastest ones are way above where any P4 ever was: 4400+, 4600+, and 4800+...
With the onset of dual core processors, it is inevitable that CPU designs will get "wider" and not higher clockspeed. The Pentium 4 has three pipelines that are very deep (31 stages) in order to get to high clockspeeds, while the approach that AMD has taken has
nine pipelines that are not as deep (I think they now have 20 or so?), so it does more work on each "tick of the clock". This is referred to as IPC (instructions per clock), and the the other major issue with really deep pipelines is a mispredictions: if you come to a logic dead end, you have to wait longer to clear out the long pipeline; as each stage takes one clock tick, so 11 more stages means 1/3 more clock ticks to clear things out -- and then 1/3 more clock ticks until you get something out at the end of the pipeline!
The biggest practical advantages that the Athlon 64 have is they run much cooler and use less power doing so. This can be dramatic: some Athlon 64's use as much power at 100% utilization as the hottest P4's do at idle -- which is
half as much! They will also run the 64bit programs better, as they now have 16 registers, while the Intel 64bit CPU's still have the 8 registers that all the 32bit CPU's have, AFAIK.
