Bjorn said:
I find it rather interesting that they've gone trough all this trouble and still didn't mention anything about image quality in the whole article. (unless i'm missing something).
Anand has recently developed a new strategy for IQ evaluations among products. Whereas his former approach was to publish long, complete, somewhat exhaustive reviews which included IQ examinations, and which generally garnered him high marks for effort if not content, of late he has taken to "multiple-part" hardware reviews which seem to always postpone IQ evaluations until part II or part III of his review, which may differ from Part I chronologically by weeks or even months when published.
My theory about that is that it has become inconvenient for Anand to do IQ and performance reviews at the same time because Anand's primary sponsors object to them being done at the same time. What happens is that Anand does "performance" in Part I, which is based on a set of highly optimized drivers which suffer from an IQ perspective, which provides his sponsor's products with glowing performance numbers. In the latter parts of the review dealing with IQ which appear much later, the IHV sponsor has had time to work on the IQ in subsequent driver sets so that it is now possible for Anand to say, when he publishes his IQ evaluations weeks or months later, that "As you can see, from what I hear there is no difference in IQ between product X and product Y," and Anand will publish "screen shots" which he purports to prove such assertions.
But since "Part II" (or III, as the case may be) deals with IQ, instead of performance as measured by frame-rate bar charts, Anand simply links to "Part 1" out of Part II (or III) for his performance numbers in Part II, and so directly relates the IQ revealed in Part II to the performance he revealed in Part 1, some weeks or months earlier. This allows Anand to be kind to his sponsor and test his sponsor's products with separate driver sets for performance and IQ, respectively, and thus remain "fair" to his sponsor. After all, as Anand says, "From what I'm hearing, the idea that different drivers produce differences in IQ and performance when run on the same hardware is but a myth, because IQ and performance are entirely different things, and so it's only natural to use different drivers when comparing the two."
Yes, indeedy, from what I'm hearing that makes all the sense in the world...