In this comparison, you can see the increase in the overall 3D performance as the CPU performance increasesThe E2160 at 1800MHz sets the reference point for percentage increases: A rating of 120% for example is only 20% faster than this reference point. The percentage value is calculated using the overall result achieved in the benchmark suite at all tested resolutions, by a combination of graphics card and CPU.

The Geforce 9800 GTX responds very well to the clocking speed and performance class of the CPU. Overclocking the FSB by 33% on the E2160 results in 20% more performance overall for the Geforce 9800 GTX. There is only a difference of 2% to 7% among the E6750, the Q6600 standard rate and the X6800EE.

The Geforce 8800 GTS OC responds even better to the CPU performance, thanks to the higher clocking rates for shader and GPU. A weaker and lower clocked CPU wastes at least 35% of overall 3D performance with the G92 graphics chip.

A Geforce 8800 GT OC is as fast as an unclocked Geforce 8800 GTS 512. The streamlined graphics chip has only 112 stream processors instead of 128, which has an immediate effect on the maximum potential increase in performance. It is 3-5% lower than for the Geforce 8800 GTS 512 OC.

Where the overclocked E2160 is able to obtain a 16% performance increase from the Geforce 9600 GT, the benefits sink to 3-4% for the more expensive CPUs, when the values are compared using the Geforce 8800 GT OC.

A more powerful CPU increases the overall results of the Geforce 7950 GT by 11%. The overclocked Q6600 at 3.2 GHz loses points in this graphics card class to the E6750 at 2.67 GHz, or the X6800EE at 2.94 GHz.

The Geforce 6800 GT barely profits from the better CPU; in fact, when used in conjunction with a quad core, it actually loses performance. These results arise because the Geforce 6 reaches its limits with maximum graphics quality and high test resolutions. With lower resolutions and graphics quality, the CPU is able to achieve more.