Ali Re...<<<<<<<<<<<<It does not matter how many. What does matter that the
particular segment holds a performance crown. It makes
people feel like belonging to the elite class even
if they own a crappy-performing<<
What percentage of computers are bought by just such customers. Corporations reportedly buy over 70% of the computers, and how many IT managers care about the secretary's need for speed. In the old days, IT managers could justify high performance for almost everyone, as time is money, but now, I would bet no more than 10% of corporate computers justify the extra money.
We will need to see this "undisputed lead".
I am sorry if I misinterpretated what you were saying when you said this. . In the PC business, performance sells, and will be.
Either MHz-based, which is easy to communicate to
buying public, or true performance, which must have
a clear undisputed lead<<<<< The clear undisputed lead I thought you were talking about was in IPC, as in, I think computer OEMs will go back to rating computers by MHZ unless there is a clear difference in IPC in AMD's favor. And I feel Hammer will do just that and force AMD based computers ,at the least, to be performance rated.
Two cores would require a bigger die too, would not they?<<<<
True, but the $100.000 question is if you gain more performance with 2 cores with an x amount of bigger die, than putting in x amount of cache. IBM, in the power4 series went with 4 cpus, and has a big lead over Itanium.
Maybe it should, but it is not what is happening.<<<<<<
Oh its happening, maybe not at the rate you would like, but it is happening. AMD increased its market share up to 22% by many accounts, which is a huge market share increase, percentage wise, and that is without much of corporate sales yet. Marketing, advertising, etc, can often slow the change rate down, but usually, it happens eventually. Patience.
<<<<<I found it strange that nobody picked up on my thoughts
about inherent die size limitations and therefore
about limited shrink scalability...<<<<<<<
I thought I did, and I said that the small die theory isn't just about a small die as opposed to efficiency. Certainly, if the die needs to be bigger, than AMD could easily increase the die size, by adding cache for instance. However, Intel's big die theory, if I understand it correctly, dictates a big die over an efficient die , in order to starve its competitors of resources, and take advantage of Intels lead in processing. That is what I disagree with. A big efficient die is fine, a big ineffcient die is bad, because it forces the consumer to pay extra, just so Intel can maintain its monopoly. Any company which has itself as the prime benefactor, will eventually find itself as its only benefactor.
We have not
seen any evidence that the commodization is happening.
Quite to the contrary, former CPU design houses are
going belly up, only two remains standing in the x-86.<<<
When I read that, I tried to look up commodization, but my dictionary seems to have come out before that was a word. At any rate commodization to me means being common, and generalized, as opposed to specialized. And isn't that exactly what has happened. The whole world uses x86 computers design by just two houses, and they are being sold by the millions. |
