Tony, Re: "Actually, I was thinking that Hammer may be the next Betamax or Apple...better than the competition but flunks anyway because of things like infrastructure support, maybe even production problems with SOI. I know, same-o, same-o."
I disagree. Hammer will be a great chip in and of itself, and AMD will find a way to make the performance look very attractive. x86-64, on the other hand, is something I am very skeptical about. Most of the AMDroids are hoping that it will somehow cause a paradigm shift in consumer demand, and everybody will be asking for Hammer, because it has 64-bit extensions. Big whoop, if you ask me. MMX was said to be 128-bit, but that didn't sell it any better.
Intel has had only limited success pushing instruction extensions, mostly because the effort required is immense. Developers like to try new technology a) when they have extra time, b) when other projects aren't standing in the way, and c) in between project life-cycles. a+b+c rarely happens, and trying to brute force developers by offering incentives and support is not something I see AMD capable of doing. Finally, expecting the whole industry to move at once is sort of like hoping the Atlantic ends up in the Pacific - it's just not going to happen, even with a very large bucket.
Of course, the main argument here is if AMD can't get developers to program for x86-64, how will Intel get them to do it for IA-64. I think it has to do with a) experience, b) incentives, c) money, and d) momentum. With IA-64, Intel has all these; with x86-64, AMD has none.
wbmw |
