Whenever you’re willing to buy a new PC or upgrade your graphics card, there will be a colleague, neighbour or friend who’ll tell you that the CPU is going to be a bottleneck if you take a really strong graphics card. Truth be told, these claims are not without basis in fact, but lately, we’ve been under the impression that the issue has been overblown by far. In other words, how many people do you know who’ve actually tested a certain graphics card model with several CPUs and can back up their claims in fact? Certainly, there is a difference between various CPU generations, but how does a Core i3 cope with applications other than the CPU-intensive ones compared to Core i5? Is the price difference of 50€ a clever investment if the only thing you care about as far as performance goes is the number of frames per second?
In order to answer this question, we’ve devised a detailed plan. Firstly, we needed CPUs of the same architecture and from the same generation/family, so that we get results as realistic as possible. Furthermore, it would be unfair to only cover CPUs of one manufacturer, as that would leave out too many users from the applicability of the test. Afterwards, in order to get exact and precise data on performance jumps with different CPUs, we had to obtain one representative from each market niche, so to speak. The resulting conclusion was that we needed four CPUs from each manufacturer, covering the entire range of the latest the CPU market has to offer. Representing Intel, we have the freshly presented Pentium G840, over Core i3 2100 and the very popular Core i5 2500K, all the way up to the serious Core i7 2600K. Ergo, Intel’s entire palette. The same goes for AMD: their strongest representative, namely Phenom II X6 1100T, then down to the mid-class with Phenom II X4 980, and finally the dual-core tandem of Phenom II X2 560 and Athlon II X2 650. You know what this all means - a hellish round of testing.
