Looking at the architecture and comparing it to other hardware is great, but how do these limitations on paper show up in the real world? That is what doing some simple benchmarking will show. We continue to stick to the core of game engines to see what is happening. The so called DX10 titles are FUD for the time being, as they are merely DX9 ports and even DX9 console to DX9 PC to DX10 PC ports (such as Lost Planet). You can look at them as much as you want, but until many more DX10-centric games appear, you cannot put too much stock in them. However, the value and entry level cards aren't playing.
There is an eerie issue in the wings that I haven't heard much talk about among gamers. Nvidia has invested time, effort and money into exclusivity, with titles coming out the rest of this year. What does that mean? Sometimes that simply means that a title will just have a logo on a box, but it could be as bad as limiting the access of ATI to coding optimizations right out of the box. Normally this depends on the developer: if a developer is looking for more support and the financial benefits of hooking up with a cash rich IHV, then a lockout can be likely. Large developers and publishers have much more power and are not swayed by IHV offers. Either way, this kind of practice is not conducive for the industry and especially for gamers. If the IHV wants to help optimize games for their hardware, great! Games should look, feel, and play well out of the box. However, if the edge is to use some form of prohibitive practice then that is not cool.
Coming back to the topic at hand, we stayed with Windows XP for this article. Most of you are using it, and frankly, performance is still better under XP. We have much more in store for this topic in subsequent articles. In fact, there are two articles in progress and a third in development.