AI
Friday, June 24, 2005 7:15:01 PM
Re: Ok, you found one in the universe of of chips where an inferior Intel chip does not sell for premium, but the whole package, including premium chipset and or memory still push the price above, so you would still be off.
I don't know where you are coming up with this "one in the universe of chips" stuff. I am talking about 100% of Intel's dual core products, the core called Smithfield, ever heard of it? It also runs on the 945 chipset with commodity DDR2 DRAM, so there isn't much of a premium here.
But let's agree on one thing. Smithfield is not the ideal dual core. However, it does win some price/performance arguments, and Intel will sell millions of them this year in spite of the technical arguments against it.
Re: So you are right, AMD is doing barely OK financially with significant lead in performance in 2 out of 3 segments, and if Intel catches up in 2 of those 3, things will be more difficult for them. But the reason has nothing to do with anything in the first paragraph of your post here
No, I think there is a point in what I said if you look at it carefully. First, do we agree yet that Toledo had an R&D budget? That it took more than an insignificant amount of money to design? If so, then we are looking at a scenario where the X2 isn't going to make money for AMD (that is, recover their R&D investment) until some time next year, when it will more than compensate. I believe they made the right decision to go dual core at 90nm. It's not too much more than they can handle, and it gives them one hell of a competitive boon. Quad core, on the other hand, is a different story.
Like I said before, 90nm quad core is much larger and hotter than AMD can handle. If they aim for a low volume launch, they are never going to ramp the thing to the point where it can make money without seriously compromising their other products. Furthermore, I think a 65nm dual core product is too risky, because it will be a brand new process for AMD. They *might* be able to do it, but it could lead to massive delays. A second generation 65nm product, some time in 2007, makes a lot more sense, which brings me back to the point I made in my original post.
Re: Come on, a dispute was about which code name refers to which concept. It is the concept of QC 32 bit Yonah derivative that I was talking about all along, and I said nothing negative or AMD favorable thing about Merom derived server parts (CSI or not).
Come on... -> Joe.
Intel publicly put Whitefield on their MP server roadmap for 2007 at the last IDF, and here you are speculating that it's a 32-bit-only quad core Yonah. Surely you can see how this doesn't make sense....
I don't know where you are coming up with this "one in the universe of chips" stuff. I am talking about 100% of Intel's dual core products, the core called Smithfield, ever heard of it? It also runs on the 945 chipset with commodity DDR2 DRAM, so there isn't much of a premium here.
But let's agree on one thing. Smithfield is not the ideal dual core. However, it does win some price/performance arguments, and Intel will sell millions of them this year in spite of the technical arguments against it.
Re: So you are right, AMD is doing barely OK financially with significant lead in performance in 2 out of 3 segments, and if Intel catches up in 2 of those 3, things will be more difficult for them. But the reason has nothing to do with anything in the first paragraph of your post here
No, I think there is a point in what I said if you look at it carefully. First, do we agree yet that Toledo had an R&D budget? That it took more than an insignificant amount of money to design? If so, then we are looking at a scenario where the X2 isn't going to make money for AMD (that is, recover their R&D investment) until some time next year, when it will more than compensate. I believe they made the right decision to go dual core at 90nm. It's not too much more than they can handle, and it gives them one hell of a competitive boon. Quad core, on the other hand, is a different story.
Like I said before, 90nm quad core is much larger and hotter than AMD can handle. If they aim for a low volume launch, they are never going to ramp the thing to the point where it can make money without seriously compromising their other products. Furthermore, I think a 65nm dual core product is too risky, because it will be a brand new process for AMD. They *might* be able to do it, but it could lead to massive delays. A second generation 65nm product, some time in 2007, makes a lot more sense, which brings me back to the point I made in my original post.
Re: Come on, a dispute was about which code name refers to which concept. It is the concept of QC 32 bit Yonah derivative that I was talking about all along, and I said nothing negative or AMD favorable thing about Merom derived server parts (CSI or not).
Come on... -> Joe.
Intel publicly put Whitefield on their MP server roadmap for 2007 at the last IDF, and here you are speculating that it's a 32-bit-only quad core Yonah. Surely you can see how this doesn't make sense....
