Here, I'll avoid any suggestion of lying...
I'll come right out an assert it. Does "Top to bottom, stack of 10 SKUs" family launch ring a bell?
There are plenty of places to find that particular quote paraphrased by ATI/AMD personnel and that is but one example. I'm not pro-IHV with regard to any of them, but I know BS when I smell it. Please don't even imply that AMD hasn't gilded the lily, - or to put it more plainly - *hoodwinked and duped* with the R600 launch. Easy litmus test : was it true (red) or was it blue (blue).
They strung people along that the R600 would rock...and it didn''t. Thus questioning their truthfulness makes sense. You don't buy a chicken without testing it from the same guy that sold you a rabbit two days ago thinking you wouldn't catch on.
I read the same quotes you did--exactly the same ones--and I read them exactly the same way that you did--but I'm having a great deal of trouble trying to understand how you're equating R600's launch to date with Barcelona as AnandTech claimed to have "tested" it? That's what AnandTech was talking about here, and so was I. What I said about R600 was incidental and meant simply that the R600 story isn't over--it's just beginning. That is my opinion.
I'll assert something equally: anybody at all who has looked at and touched a pre-production Barcelona sample somewhere, which is running in an engineering-sample test environment, and who further claims to have "run at least one", I say again, "one test" on it--and then further claims to have come to some ultimate conclusions about
what Barcelona products will be when they ship is lying through his teeth...
To put it more politely, he's a bloomin' idiot...
I simply cannot count over the years how many, many derisive, critical, asinine "stories" I've read that were written by people who thought so much of themselves that they could take brief, teeny-tiny looks at pre-production hardware running in engineering-sample type environments and predict the ultimate fate of those products when they actually shipped in final form. Really, *anybody* who has been around and involved with the software-hardware scene for just the past decade can say the same thing with absolute candor. This axiom has applied to just about every major hardware and software product release I can recall.
Now that I've had to again explain my position relative to premature conclusions being reached on the basis of tiny glimpses of pre-production hardware--namely, that such conclusions very often prove themselves foolish when those products actually ship--I'd like to conclude with adding my perspective on R600 and what AMD has said about it to date...
Back in early 2002, nVidia started lying through its teeth about its upcoming ".13 micron" nV3x gpu family--it started--oh, at least a year--before nVidia actually shipped nV3x. After nV3x shipped, the lying continued at a rapid and intense pace--that included lots and lots of things I simply have no interest in reminding you of at the moment. When did nVidia *stop lying* about nV3x? Answer: when nVidia shipped nV4x, that's when. Since then, even JHH has publicly stated the nV3x was a "mistake" quote, unquote. Point I'm trying to make is that whatever mistakes we all saw nVidia make with nV3x, it did not stop the company from turning itself around beginning with nV4x, did it?
Relative to R600, which has been released all of a couple of *months* now, it's still far too early to know what sort of ultimate product that R600 is, or even will be, in the coming months. That's why I think it's way, way too early to count the R600 chickens that have hatched, simply because so few of them have hatched to date. Do you think that two years from now AMD is going to make statements to the effect that R600 was a mistake? My own opinion is that it is still far too early in R600's lifespan to know whether in two years AMD will be apologizing for it or boasting about it. Time will tell.
One last point here. Does anyone remember nV25? What I remember very clearly about it was reading several pre-release accounts written about pre-production samples that people had seen running in engineering-sample test environments--these reports talked about "abysmal clock speeds" and "frequent crashing" and so on and so on--and pretty much flatly stated that nV25 was basically a piece of &hit....
Well, it so happens that I bought an nV25 product after it shipped in final form, and I experienced *none* of the nightmarish things people had written about it as they observed it its very early, pre-production forms. I remember very similar things written about the 3dfx V3 and V5 products immediately after they shipped--and when it comes to the V3 I'll
never forget as long as I live the horrible, horrible screen shots that AnandTech (speak of the devil) posted on site purporting to accurately represent the IQ produced by the V3.
Those screen shots were so ugly that my V2 easily appeared superior. Imagine my surprise upon actually purchasing a V3 and discovering that not only was nothing AnandTech had written about the product actually representative of it, but that up to that point in time the V3 was clearly *the best* 3d card I'd ever owned, and I'd owned a lot of them...
Later, of course, it came out that AnandTech's screen grabbing wasn't grabbing the V3's post filtering, and that was why the screen shots he posted looked nothing at all like the IQ the V3 produced from the start onscreen. The fact that a lot of damage had already been done never seemed to bother Anand to the extent that I certainly felt it should have.
OK, that's just a *sample* of the kind of stuff I'm talking about. Here's the thing: is it possible that when it ships Barcelona will suck? Of course it is--just like it's possible that Intel's 45nm mass production plans at present will implode just like Intel's 90nm Prescott production plans imploded. But what all of us should do, I'm convinced, is
wait until the final product ships before deciding what's what. Trying to make conclusive judgements about products based only on brief glimpses of pre-production products running in sample test environments is probably the best way that I know of to wind up with egg on your face. If Barcelona indeed sucks, or if Intel's 45nm mass-production plans are fairy tales for children, there'll be plenty of time to say that
when these products ship in final form, won't there?
What surprises me most about some of these responses is that they were necessary in the first place. I had thought most of this to be common knowledge.
