1. The eligibility test was whether the chip was sampling by the end of 2002. NV30 unquestionably passes this test, and it's by far not the only winner to get in under this "technicality".
I still think this is an odd criterion for a "best of 2002" award, but it does make a certain degree of sense considering uPR's readership, namely microprocessor design engineers. The design engineer's job is done when the chip is sampling; when the product based on the chip is commercially available is based on a completely different set of factors. As the award is for design engineering, in a journal targeted at design engineers, using sampling-date as the cutoff is not necessarily a bad choice. (Although it still seems strange to refer to most of the winners as as "2002 chips".)
2. Peter Glaskowsky
is one of the most respected and influential computer industry analysts in the world. This isn't a matter of opinion, it's a fact. And judging by his
bio page, he focuses particularly on graphics/media technology (I'd only run into his name in connection with PC platform/CPU related stuff before), so it's probably also an undisputible fact that he is the most respected analyst of the 3d industry. It may be impolitic of him to come out and state it like that in the nvnews.net forums, but it's not conceited, because it's true.
On second thought, the entire idea of him debating the choice in the nvnews.net forum was impolitic and rather bizarre, and I assumed it was a hoax at first but after comparing his posts there with his
defunct personal website I realized he's just that goofy.
3. It's still a bad choice, and almost certainly a result of being biased by the hype. I don't know that much about Peter Glaskowsky, but I do recall him relentlessly playing up Merced (Itanium 1) like a pathetic Intel sycophant...except that
Intel itself had by that point dismissed Merced as nothing more than a pilot architecture to seed ISV development for Itanium 2.
Point is, Glaskowsky has at least some history of being swayed by hype (and I'm willing to bet there are plenty more examples). This is probably what makes him such a successful analyst. That's a serious comment, BTW: you rarely go wrong betting on establishment players who hype their products successfully. That means they have the best marketing, and probably the best brand associations among potential customers. Not to mention the resources to eventually make good on their overhyped product with a successful followup.
Glaskowsky was completely wrong to go gaga over the ridiculous Merced, but he can cover it up now by pretending he was just sold on IA64 as a whole, which has produced the fastest currently available CPU on the planet, and will likely continue to hold that disctinction for most of the forseeable future. Of course that's because Intel's 64-bit competitors all laid down and died in the face of the hype, except for Sun (which was the least competitive in the first place) and IBM (whose long term commitment to fighting Itanium is questionable). Obviously the same won't happen with ATI and Nvidia, but if NV35 turns out to be a balanced and successful product which significantly outperforms R350 (and it very well might), this award may not look in hindsight as stupid as it is in reality.
4. But it is stupid in reality. Blech.