The credit crunch may be forcing us to make the most of what we've got and it seems graphics chip manufacturers are no less affected than the rest of us. It's been some time since either nVidia or ATI came up with a truly radical new design.
The GTX 295 is really two GTX 260s in one, while this, the PNY nVidia GeForce GTX 285, is only a slightly warmed-up version of the GTX 280. It's smaller too, and consists of the 280's 65nm GT200 architecture compressed into a smaller 55nm package. The leaner dimensions allow the chip to be run at higher speeds without overheating, so it's an extremely cost-effective method of creating a new chip. The result is that the PNY nVidia GeForce GTX 285 isn't as expensive as the MSI nVidia N280GTX-T2D1G-OC also reviewed this month.
The PNY nVidia GeForce GTX 285 isn't as fast as the Asus ENGTX285 TOP we reviewed last month, and sticks closer to the stock figures put out by nVidia. In fact, the 648MHz core clock speed is actually 2MHz slower than the 650MHz offered by MSI's overclocked version of the 280. But the PNY nVidia GeForce GTX 285 still prevails in the end, and its effective memory clock speed hits 2,484MHz - a good 184MHz up on the MSI's 280, but 116MHz down on the Asus version.
Get free games downloads. Visit PC Advisor's dedicated Games website to download hundreds of the latest titles, to read gaming news and game reviews and to pick up tips and discuss your favourite games in the popular PC Advisor Games forum
This is, however, a very decent general-purpose card. Its switch to dual 6-pin power connectors makes the PNY nVidia GeForce GTX 285 ideal for multi-card systems. And its hardware is sufficient to make good use of those potentially exciting developments that nVidia has implemented across the whole of its GTX 200 range - CUDA and PhysX.
CUDA will make it easier for programmers to tap into the PNY nVidia GeForce GTX 285's sophisticated hardware and add some rather attractive bells and whistles to future games, while PhysX takes the technology pioneered by AGEIA and allows coders to build more complex physics calculations. Both technologies rely on programmers implementing their features. And, while they may well count for something in the future, there's still little on the market to make CUDA and AGEIA compelling reasons for buying nVidia.