Oh, that's great. So they managed to pack the damn thing onto one PCB, so you're giving Nvidia the win because it beats the 4870 X2 in "most benchmarks."

Except, you didn't really review the facts. The fact is, these new cards are being sold at exactly the same ... the older dual-PCB parts. Further, that price is $125 more than the MSRP on the 4870 X2 ($399), and more than the MSRP of two 4890 cards ($199 each), not to mention two GTX 275 cards ($219 each)!

Any "savings" from the new single-PCB design are being pocketed by the manufacturer, so there's really nothing new here to write a review about, is there? The GTX 295 will continue to be bypassed by enthusiasts who want better bang for their buck, and only a handful of fools will drop the $525 necessary to own this sucker.

And if you don't look at MSRP, but look at the actualy prices... the picture is even sadder for nVidia. Cheapest GTX 295 (Single PCB) - $489.99 for XFX on ZZF ( http://www.zipzoomfly.com...770&prodlist=celebros )
Cheapest HD4890 - $185 for HIS on NE ( http://www.newegg.com/Pro...76&Tpk=HIS%20H489F1GP )
Cheapest GTX275 - $205 for Sparkle on NE ( http://www.newegg.com/Pro...=SPARKLE%20SXX275896D3-VP )

This is just by looking at the first 2 pages on froogle for each of the items. One can often find HD4890 for around $160 shipped with MIR. I've seen GTX275 get as low as $190 shipped.

And this crippled, failure design of GTX295... well, it's still overpriced POC for 14yo with no brains and rich parents.

Sorry, but this is a fail. Good night.

It sounds like the author(s) have an never ending bias for all things Nvidia.

Terribly written article.

The authors of this article have always provided what I thought to be a fair and balanced review. However, it appears to me that some fine tuning maybe needed for future reviews.

Therefore, it is my suggestion (and that's all it really is) to forgo the use of IQ enhancing features such as Adapative AA: Quality and Antialiasing - Transparency: Multisampling in favor for how the developers intended the game to look based on in game settings. The use of those control panel settings are a always a welcomed addition. However, since there are no IQ comparisons being offered the use of them do appear counterintuitive when performance (among other aspects) are implied.

This would allow us (the readers) to gauge what the performance of those games really are and, if we want to use features such as AdAA or TrAA it would be with the understanding that there would be a performance penalty. Everything else in the testbed still looks fine to me.

Hey, why do you always test the ATI cards with Adaptive Anti-Aliasing: On/Quality and the NVIDIA boards with Multisampling only?????? You just make a big dissavantage for ATI cards from the start when using AA!!!!!!