Thanks TheSwede, there certainly is some interesting stuff there!
LiamC, excuse me, I didn't see your post till just now. Tea is asleep on the job, same as usual. I'll make some enquiries anout the nForce II Epox boards as soon as I get back from my holidays. (Yes, I'm having a couple of weeks off.)
All: OK, it's very early days yet, but from what I read there it rather looks as though the nForce 2 and 166FSB Athlon could be one of those magic sweep-all-before-it combinations. Something in the same vein as:
OPTi 895 & DX/4
VP-X and 6x86MX
MVP3 and K6-2
BX and Celeron-A
KT-133A and Thunderbird C
KT-266A and XP
nForce 2 and Athlon RT (or whatever stupid letters it gets)
If it lives up to the hype it could be another one. But let us all remember that nVidia are just the best in the business when it comes to hype. Remember the huge wave of expectation that accompanied the original nForce chipset? It was going to stomp it's way right to the top of the tree, be the hottest-selling thing since the 386DX-40, and the unquestioned performance champion. And, of course, it flopped. Well, OK, they sold a few, but it was a few. I have seen exactly one of them, which came to us because the owner couldn't figure out how to make sense of the IDE drivers. Neither could I. Far as I know, he's still running his burner in PIO mode.
nVidia's troubles with the nForce 1 were (a) they didn't get it out into the marketplace: very few board makers bothered with it, and those that did often didn't bother bringing it to the more obscure corners of the world (like Australia), and (2) it didn't have a clear and sensible market position. It was way too dear to compete with the SiS535 or that long-forgotten and rather poxy ALI thing that ASUS seemed to like so much, and still too dear to compete with the KT-266A higher up the food chain. The only thing it was on a par with price-wise was the AMD 760, and the people that bought 760s were, in the main, quality-conscious higher-end buyers who had no use for a board with built-in video.
Now you could argue that the on-board video of the nForce was pretty good, or that it had an AGP slot and there was nothing to stop you buying an nForce board and a Gforce III if you wanted to - but why would anyone in their right mind do that when you could buy an AMD 760 for a good deal less and have peace of mind in the early days, or (a little later) a KT-266A for much less and have a marginally faster board as well? And people who buy on-board video buy one thing and one thing above all: cheap.
What it came down to was that, for the same money, you could have an nForce board (with good on-board sound and mid-range on-board video) or a KT-266A with low-end on-board sound and a real, stand-alone video card and some change, or spend just a tiny bit more and have a decent stand-alone soundcard too. We added the nForce boards to our pricelist, gave them a reasonable plug, came up with situations where they made some sense - and no-one was interested. They needed to be within about US$10 of the KT-266A boards if they were to to be a realistic retail proposition, or US$25 at the most.
Net result: we didn't sell any at all. Oh, I could have pushed them harder, but why should I? With KT-266As selling like crazy, what was the point?
And now we get the same exact approach, or so it seems. A mega hype-fest for an integrated board. Well, hats off to nVidia for trying something different, but they have to have something special if this product is to be a goer. I can see it being a winner if, and only if, (a) it performs somewhere within cooee of the hype, and (b) it really is demonstrably and meaningfully faster than the competition from VIA and SiS. Then they can charge plenty for it, and people will pay up. At this stage it looks as though it will be the fastest chipset on the planet - but so did the nForce 1. Early indications seem to be that it will be a KT-400 killer, just as the nForce 1 was set to be an ALI and KT-266 killer ..... but if VIA do a KT-400A that really cooks, then for nVidia, it's back to video cards again.