A week ago we raised some concerns over the quality of images produced by the Geforce 6800 Ultra in 3Dmark03. We forwarded a note of our concerns to Futuremark and Nvidia, Futuremark have today provided us with a response. Before we look at that response, it may be useful to read the original article which is located here: http://www.driverheaven.net/articles/driverIQ/ The short version of our concerns is that the 6800 Ultra was rendering an image that differed considerably from the images produced by 3dmark reference image. We were concerned that these changes were the result of optimisations which may not provide comparable results from the 6800 Ultra. Here is the response Futuremark have provided: â€œ We have studied carefully this issue and contacted NVIDIA and Microsoft regarding this. We re-tested the drivers and the result is the same as before: There are no 3DMark03 specific optimizations in Forceware 60.72 drivers for GeForce 6800 and they indeed fulfill 3DMark03 optimization guidelines with GeForce 6800. The explanation for the rendering differences is GeForce 6800's new isotropic filtering algorithm: It differs from that of the reference rasterizer. There is an analysis of this available at Digit Life (http://www.digit-life.com/articles2/gffx/nv40-part1-d.html). Based on all information we have, it seems that this algorithm is applied consistently the same way regardless of the application, i.e. it is not application specific and it is not targeting 3DMark specifically. We also contacted Microsoft about the difference between the rendering of the reference rasterizer and NV40: "The DX9 reference rasterizer does not produce an ideal result for level-of-detail computation for isotropic filtering, the algorithm used in NV40 produces a higher quality result. Remember, our API is constantly evolving as is graphics hardware, we will continue to improve all aspects of our API including the reference rasterizer." Now, the most interesting question is the performance effect. As regards 3DMark03, the performance effect due to this generic optimization is insignificant and should not cause incomparability of the results. â€œ So what does it all mean? Firstly the news that the rendering differences do not impact performance is good and you can be confident that any 3Dmark03 scores you have seen posted in 6800 Ultra reviews are valid for comparison with cards which are available, or soon to be released. The other side to the statement is the comment that the filtering method Nvidia are using is not specifically targeting 3Dmark03. To better understand how this filtering method impacts image quality, here is a sample of a 3Dmark frame from game test 2. The image on top is the reference image, the image on the bottom is the Nvidia 6800 image. You will notice the change in image quality from one image to the other in the red circled areas. This is the type of IQ issue you are likely to see across all applications due to the NV40 rendering method. How significant or insignificant this is should be left up to each individual. You should also take into account the impact of these changes will vasy from game to game. Thanks for reading, and we hope that our original article and this follow up have better explained the rendering of the NV40. NOTE: I'd like to thank Tero Sarkkinen at Futuremark for the time and effort put into the investigations and response.