GeforceFX=surprisingly slow

[BlackMonk]

I well remember P4 specs... first generation P4's got beat by Athlons... now look where we are a little while later?

So you mean that the GFFX has a longer pipeline and will be able to scale to higher clockspeeds?  Or are you attributing the P4's current performance to software optimization?

I'd take issue with the P4/software route since many of the P4 optimizations also benefit the Athlon and give it a performance boost as well.  Perhaps not as much as the P4, though.

If you take a P4 at 2 GHz and an Athlon at 2 GHz, clock-for-clock, and ran the same program, the Athlon would perform better.  Sure, software optimization will help the P4 a bit, but... I think the fact that the P4 is something like, what, 600 MHz faster? is more of a factor than anything else.

Since the GFFX isn't exactly positioned as the first "GPU" of a family of chips designed to scale up to 2 GHz, uh, I don't see how the analogy holds any water.

---------

Enron?  Diversification?  Actual products?  Blah blah blah.

http://finance.yahoo.com/q?s=NVDA&d=t

Looks to me like nVidia's motherboard chipset efforts, broad product range (high-end, mid-range, budget, portable), and R&D (creation of Cg and attempting to get it adopted, migration to .13 micron process) has really paid off.

I mean, a year ago their stock was almost 7 times as valuable.  Granted, the stock market is insane, but this means that investors have waning faith in nVidia's ability to provide a competitive product--diversification or not.

And now consumers are also questioning nVidia.

http://www.hardocp.com/article.html?art=NDIxLDY=

I don't know where there was a mention of 300 MHz overclocking or room for improvement.  The sample only went up 30 MHz or so.  They'd NEED to clock it up another 300 MHz the way the current sample is performing.

This is a "reference" board with premature drivers.  I fully expect the card to perform better with the actual retail products and after a few driver revisions.  However, this is a really crappy start.

Oh, and whoever linked to Tom's... that site is generally regarded as biased and bowing to nVidia.  It might not be the best place to cite praise for the GFFX.  

[BlackMonk]

Focusing on a single product doesn’t guarantee survival (refer to 3DFX ).

However, there appears to have been a direct correlation between nVidia's delays in their process shrink and their next flashship product and a lack of consumer and investor confidence.

I'm not sure what your example with 3dfx is supposed to show... they had a great product in the Voodoo2.  Then they had delays with their follow-up products and the products didn't perform as well as people were expecting.

Are you trying to imply that ATI is going to buy and gut nVidia?  That's the only thing I see in a reference to 3dfx (RIP).

[BlackMonk]

Can that be said for the first release of Radon?

The Radeon has sucked for a while.  Only when ATI got threatened by nVidia's growing market presence with OEMs did they decide to try and change their image by initiating their "Catalyst" driver program and rapid betas.

And look, it's mainly on the Windows platform--where they get the most exposure.  Still don't have features from older ATI cards working on the Mac.  Same problem with nVidia, too.

ATI and the Radeons have done well.  The hardware is capable and the drivers have markedly improved.  And they also have successfully been playing the PR game.  Now it's time for nVidia to turn this disappointing first impression (after tons of delays) around into something positive.

Even if 3 months down the line the drivers are giving 30% speed improvement, nVidia needs to do something sooner rather than later or they'll fall out of the public's eye.  Thats' just my opinion, though, so who knows.  We'll all have to wait and see.