The Inquirer-Home

Nvidia 270/290 in deep trouble

GT200 based GX2 dead as well
Tue Nov 04 2008, 06:21

IT LOOKS LIKE Nvidia's run of bad luck is continuing. The latest word is that the GT200b is in for another respin, that would be the third. This means the 270 and 290 are effectively delayed until February, and the GX2 part is quite dead.

If you recall, the original plan was to have them out for the Nehalem launch on the 17th, an ambitious goal considering the timing on the last respin. Since that spin failed, it is a moot point, but word is that NV will put something out on the market.

Since this is (was?) a simple optical shrink, there are going to be no new features, so that leaves better power, higher clocks, and lower prices as an option. Since most if not all GT200 based parts are selling at little if any profit – we hear 260s are under water, 280s barely above – don't look for price being the point the wheel-o-semicoductor-planning stops on.

That leaves power and speed. Unfortunately, the shrink is a mess and not working out well at all, so neither of those features panned out. Where does that leave us? With a shrunken GT200 that doesn't really perform any better or worse than the current part, but gives NV a little better breathing room. That is the current best case.

With that in mind, NV is probably going to slipstream 55nm parts into the mix, likely without much fanfare. Anything more would gut the eventual 270/290 launch, if they can get the shrink done right, and that is a big if right now. The other tactic, and don't put it past them, is to launch an upclocked 55nm part to try and gain back the halo.

It will be interesting if the company tries this, it can't make them in commercial quantities, so it would be a press + 200 for Newegg only part. The downside would be to eat Christmas sales, the upside would be to stroke their collective egos. Given the egos involved, this is a tough one to call from the peanut gallery.

That brings us to the GX2/dual card. Suppliers tell us that it is quite dead. We thought it was an impossible thing to pull off when we first analysed the part, and it looks like several universal physical constants agreed with us. Thermodynamics is a bitch.

So in the end, NV has what it has now, PR stunts aside, until February. If it hot lots a few wafers, and don't look for TSMC to be doing it favors right now, that would put it at September 1 + 20 weeks.

Throw in a little time for testing, not that the concept seems to matter much at NV, and you are into February.

No wonder ATI cards are on allocation. µ

 

Share this:

blog comments powered by Disqus
Advertisement
Subscribe to INQ newsletters

Sign up for INQbot – a weekly roundup of the best from the INQ

Advertisement
INQ Poll

Heartbleed bug discovered in OpenSSL

Have you reacted to Heartbleed?