Re: Unlike some of this thread, I don't think there will be a 90nm quad core CPU from AMD. While I think it would be feasible, the payoff for such a project in terms of resources expended is not great.
If that's the case, I certainly believe in a "first things first" mentality from AMD, in that I don't expect a 65nm quad core as one of the first 65nm offerings. If AMD executes, it will be with a 65nm dual core server and premium mobile offerings in Q4 (and probably an FX, since it's essentially a 1xx Opteron), followed by 65nm volume desktop offering (dual/single) in Q1, and then 65nm quad cores later in Q1 or Q2. I don't think they can get all these designs out within weeks of one another and expect a smooth 65nm ramp at the same time. IMO, it will be a painful crunch time to get competitive products to market, and AMD will launch the highest profile, highest margin, lowest risk parts first, followed by the higher volume parts, followed by the specialty skus later.
Re: It seems that some are jumping the Itanic even before the inevitable collision with the iceberg.
Whatever, Joe, it's just the same story all over again. Intel and AMD regularly trade fellows, and I'm sure each company is actively headhunting talent from the other at all times. The fact that it's in Fort Collins doesn't necessarily signal anything bad about IPF. IPF is a strange bird, one whose performance prospects are hard to ignore, but at the same time seems to have a curse of poor execution on half the projects they've tried to release (Merced and Montecito being the ones with poor execution, while McKinley and Madison seemed on target, AFAICT). Even so, in spite of Intel's best efforts to undermine IPF (sarcasm), it still beats Opteron rather handily in TPC-C with an ancient 4 year old FSB shared 4-socket architecture. Paul is right when he claims it should be even more impressive when HP launches their new chipset for Montecito.
But while that's interesting for a while, it's barely relevant for anything short term. It's fun to look at IPF's growth, poke fun at it and at the same time marvel, depending on your perspective. I don't believe the FUD for a minute, nor do I believe we'll see anything relevant this decade. It has a certain class of buyers who are eager to make use of it, and still others that would avoid it like the Plague. And I have never seen such an avid group of critics and advocates simultaneously try to argue over the fate of the architecture.
I used to be an IPF advocate, but now I don't care. I think I would like to see Montecito win a few benchmarks this Summer, but I know it will be at a tug of war with Power, and eventually lose hard-core with Power 6. Of course, IBM is really brute forcing Power 6, even more than they have with Power 4/5, so I think it's going to be expensive, and IPF will continue to grow in certain segments.
Intel has no influence with IPF, if that's what you're still worried about. In fact, they seem to be trying their darnedest just to garner widely embraced 3rd party industry support, though even this is difficult with IBM being so two-faced, and the sky falling on SGI. In the end, IPF has HP and a group of small Japanese OEMs... but maybe that's enough, who knows...?
My opinion is that the whole "Itanic hits an iceburg" routine is so old it has gone from satirical, to ironic, to FUD, to entertaining, to just plain stupid. If you want to carry on as if it's not long for this world, go ahead, but I guarantee you you'll be wrong. All IPF needs to do is double its growth one more time (it has already done this for several years in a row), and it will have sold as many units as Sun. And all it has to do is grow 3x, and it will have sold as many units as Power (high end Power, excluding PPC or embedded). This will be the more difficult ramp, but what's to stop it? The darned thing keeps growing, even with meager little Madison holding the fort. Montecito has a lot of interest out there, and I think you'll see a hell of a lot more PA-RISC customers transition this year.
What's more to say...? It's hard to call IPF successful, but it's hanging in there, and there's still no shortage of critics or advocates to keep the candle burning. Will people lose interest? Maybe eventually, but by then, it will have been entrenched as a high end niche product that nobody cares about. Sure, it will make good money, but once the major players are all vying for the last slice of the pie, the battle can go on forever....
News 
Market Data 
Discover 
AI
