Sgolds, Re: if you think Intel will role the dice and ship without seeing if their implementation works on a variety of outside sites, so be it.
Like I said on the Intel thread, people are forgetting that implementing a new set of instructions is more about the microcode than anything else. I have a feeling that rumors of incompatible 64-bit instructions in Prescott was true; however, a few tweaks in the microcode is all it takes to fix that. Intel has probably had IA32E ready for some time, and they've probably tested it extensively with Dell and others. I'm sure they will have done all the validation necessary by the time it launches.
Re: It wouldn't be the first time they rushed something to market before it was ready, under pressure from AMD. (Think MTH, 1GHz PIII, P4EE.)
This has nothing to do with the argument at all. Not only that, but you are bringing up cases that have been relative successes. Only MTH was a recalled part, and an easily proven failure.
Re: IA32E? I don't think the public will buy the notion that 64 bits isn't really 64 bits because of the name. Particularly when one considers the potential Itanium buyer, who will know that IA32E = AMD64.
What is 64-bits, except for a marketing gimmick? Jeez, 64 is bigger than 32, so it must be better, huh? That is truly capitalizing on the stupidity of people, so you think de-emphasising 64-bits and calling it 32-bit Extended is going to be transparent by comparison? I think it's a wonderful way for Intel to suggest that IPF is a "true" 64-bit solution, while IA32E offers the memory addressibility and some performance features of 64-bits, but without the secret sauce that *really* makes it special.
Re: do you think they did a partial implementation (subset)?
I'd believe a superset before I believe a subset.
Market Data 
Markets 
AI
