Anemone,
With regards to "plug and play", that's been the situation for over 20 years in terms of PCs. It is nothing new. How many PC clones were there, after all? Too many to name.
I am not talking about building them, which is simple, but in supporting them. In building new lines of production, in supporting staff to back up machines, in training sales staff to know the differences, in warehousing more models, in acquiring more chipsets, etc... In dealing with another company and having staff that interfaces with them and engineers assigned to that line, etc...
It isn't trivial, but not overwhelming either. Working with AMD based machines is a lot more difficult than working with Intel based machines, because of the incredible level of support Intel gives to their customers and the customers of their customers. This is not speculation, I know this from first hand experience. Intel is there any time we call them, or anytime Dell calls them. Size does matter. Of course, Dell could offer excellent support for AMD based machines without AMD being quite as helpful as Intel, but again you have to factor in that cost.
With regards to risk taking, I would take the exact opposite position of yours. I would say it is riskier to NOT use AMD chips than it would be to use them. I think the risk is going with a more efficient model that offers less selection. In retrospect, it works very well when Intel has a competitive product, but even Dell has not escaped the wreckage from the Pentium 4. The other thing is, the Pentium 4 may have helped AMD's reputation improve so much that the perception of quality AMD offers could force Dell to accept it even if Conroe, et al, are better than AMD's chips.
From the reports I have seen from industry insiders, there seems to be little doubt this will be a better processor than the Athlon 64. In a discussion, Linus Torvalds, who is very unbiased and obviously pretty bright, responded to me that he expected this processor to have "best in class" integer performance. His explanation for the fourth pipeline stage was to compensate for the slower L2 cache timings, but I'm not sure I buy that since the Yonah already has the slower cache timings with only three pipes. Plus, up until now, the fourth pipeline was useless and a slower L2 cache is not going to make it any more useful. But, the consensus is that this will be better than anything in the x86 world in integer, but no one seems as sure in floating point.
I am not saying you are wrong, I actually agree with your approach of wait and see, because I've been badly surprised a few times. But, in this case, it seems that most knowledgeable people agree it should be a very, very good processor, and typically this close to release the information gets pretty accurate. But, I agree, nothing is certain until we see it, so we'll leave it at that.
With regards to the mobile market, it may be fastest growing, but I'm not sure rubbish like Turion will help Dell too much since Intel has a superior product currently. Let me ask you, do you think a lot of people say no to Dell because they want to buy a Turion based notebook and Dell does not offer one? I doubt this is a high number, although maybe the demented consumer favors this approach. I do know that people look elsewhere for desktop and server machines because of AMD processors, and I think this number is much greater.
One term I want to clarify is "spectacular". I really doubt the Conroe will be "spectacular" as you put it, so I think we agree on that. I think it will be overall a very good processor, well-balanced but not spectacular in any sense like the Pentium 4 was. I guess the fourth integer pipeline could be spectacular in one sense, but I doubt it will be perceived that way. What would you think if it offered say 10% more integer performance, roughly equal floating point, around 10% less power use and was around 10% smaller than the Athlon 64? It really wouldn't be spectacular, or would it? It would be a great product though, I think we could agree on that. So, I don't think it has to be too spectacular, because Intel has created spectacular failures like the Pentium 4 and Itanium. So, I think they'll just focus on more basic, and effective design for this processor. I sure hope so.
[Posted by: TA152H | Date: 01/30/06 04:42:06 PM]
Make Home Page
Add to Favorites
Print article
Send link to friend
Digg this
Post to del.icio.us
Post to Slashdot
