ATI Radeon X800 XT Platinum Edition

Nvidia GeForce 6800 killer?

Funny as it sounds, but it could be the low power consumption that makes the X800 a success. When you consider that the Nvidia 6800 Ultra draws a massive amount of power, it's not really an option in a small form-factor system. But the Radeon X800 XT will happily sit in an SFF box, and still push big frame numbers in all the latest games. Add to this that the Radeon X800 family of cards are all single-slot solutions, and it's clear that anyone wanting cutting-edge performance in a small system case is likely to go down the ATI route.

The decision to use GDDR 3 memory has cured any potential memory bottlenecks, allowing frequencies up to 800MHz (1.6GHz effective). The need for faster memory isn't something that's escaped Nvidia either, which is why you'll see GDDR 3 memory on the GeForce 6800 cards as well, running similarly high frequencies.

Physically, the Radeon X800 Pro looks pretty much identical to the Radeon 9800 XT. This is no bad thing, since it will obviously fit in almost any PC you care to throw it at. We probably will see a few custom designs, like we did from Asus with the 9800 XT, but most board vendors are likely to stick to the reference design.

As I've already mentioned, the latest generation hardware from both ATI and Nvidia look pretty similar (clock speeds aside). But, there is one major difference between the Radeon X800 and the GeForce 6800, and that's Shader Model 3.0 support. Now, this is where the marketing battle will be fought between ATI and Nvidia. Obviously Nvidia will be pushing Shader Model 3.0 support quite heavily since ATI doesn't have it, but how big an issue is this really?

Well there's no denying that there are some interesting features in Shader Model 3.0 that could potentially make developers' lives a little easier. For one, Pixel Shader 3.0 removes the 32 instruction limit, allowing for more complicated effects. Also, Vertex Shader 3.0 has the ability to create multiple instances of the same model with very little overhead, thus allowing more detailed environments to be created without causing too much performance degradation.

As far as Pixel Shader 3.0 goes, ATI argues that the same or very similar effects can be produced using Pixel Shader 2.0 and that's probably true. Where it does become an issue is when a developer wants to use more instructions than PS 2.0 will allow. For me, though, the best argument is not whether the same effects can be produced using PS 2.0 and PS 3.0, but rather whether we'll see many games using PS 3.0 to its full potential any time soon. It's all very well having hardware that supports the latest technology, but if nothing makes use of that technology is it worth having it? But then the other side of the coin is that it's better to have a feature and not need it, than to need a feature and not have it.

The Shader Model 3.0 debate is going to be a long one, and only time will tell whether ATI or Nvidia has got it right. Ultimately, it's in the hands of the game developers, so we'll just have to wait and see.