hmaly, Re: "With 500 million transistors; I would consider it unlikely that Madison will ever got into the desktop arena, but then I have been wrong before."
I agree. Madison makes no sense on the desktop - who said it would? My point is that you cannot ignore its prospects for the server market.
Re: "So, why worry about something Hammer really isn't going up against. Yamhill and Xeon will be the competition."
You seem to see things similar to the way that I do. It's AMD who wants Opteron to be an Itanium 2 competitor, simply because it's a no-lose proposition. If they can, they will try and force Intel to a common x86-64 platform, where they can compete on equal ground, and if not, nothing is lost by making false promises. It's more of a strategic move for the future - not because AMD is in danger of losing market share to these products today, but because of their fear is that if IA-64 becomes pervasive enough to be considered ubiquitous in the high end markets, then it gives the industry a sort of "dependence" on a new architecture that AMD would be years behind in trying to compete. They would rather spend their resources on one architecture - x86 - and proliferate their design that way.
Similarly, it's a great strategic move for Intel to try and create a sort of "dependence" on their new architecture. Not because they plan on keeping it exclusive, but because they've spent years in development, and they know that anyone who wishes to license the technology would be years behind in "catching up". That would basically give Intel a leadership position for years to come.
I think that prior to the last couple years, Intel had really underestimated how the CPU market would evolve. They originally bet against themselves, and figured that RISC based processors would control more of the future than x86, so Itanium was more of an insurance against being left behind with an aging x86 architecture. But x86 didn't end up losing to RISC, and consequently, Intel had to scale back quite a bit on their original prospects for the design. At one point, they might have thought that IA-64 would make a great replacement for all their markets, but I think it's obvious by now that x86 still has a lot of life in it - so much so that the only market left for a high powered, no limitation design is in the high end enterprise. Thus, IA-64 is firmly fixed in a specialized market, at least until Intel can drive down the costs and the power requirements significantly. Perhaps several process shrinks in the future will allow this to be possible. Montecito is two process generations past McKinley, and it is scheduled to debut in only two more years. Talk about fast progress... it's one of the reasons why the industry continues to be excited about IA-64. There is simply so much in the short term to look forward to.
Re: "Frankly, it might be to AMD's best interest if Madison is a smashing success, as then Intel will have to put a lot of resources into a non-competitor."
I don't know about that. Intel has the resources to throw around, that's for sure, and developing a new architecture side by side with their current one is far from being outside of their capability. If the worst happens, and IA-64 fails, it sure won't mean the end of Intel. They will sweep it under the rug, along with Rambus and Timna and all the other failed designs that analysts don't seem to remember any more. Something has to be said for Intel that they can continue to cast such a positive impression despite all their recent failings.
But if Madison is a smashing success, it leaves the door wide open for Montecito and others, and I think that Intel can continue a strong product line. I just hope that success doesn't go too much to their heads, and they decide to divert resources from future products, and make certain assumptions like they will be "guaranteed successes". Only when management is stupid enough to that is when they deserve to lose market share to stronger competitors. But as long as Intel can introduce a new core every year, the IA-64 line will grow very quickly, and Intel may very well find new markets to put it in.
Re: "Nah, I consider you a bud. I will forgive your silliness, if you forgive mine."
Fair enough. We all have our moments, don't we.... :-)
wbmw |
