GPU Gameplay Performance Review

GPU Gameplay Performance Review

On October 11th, 2011 AMDs Bulldozer CPUs were launched. We looked at Desktop Performance as well as Gameplay Performance on the AMD FX8150 CPU. Our initial Gameplay Performance Review of the AMD FX8150 focused on performance with a singleGPU at 1080p. Our results in that configuration indicated that performance was similar between the CPUs compared, except in Civilization V which showed a major advantage toward the Intel CPUs.

The goal of this evaluation today is to expand upon gaming performance with the AMD FX8150 CPU using multipleGPUs and higher resolutions. In our initial evaluation, we used a single Radeon HD 6970 GPU. This time, we are going to be using NVIDIAs GeForce GTX 580, but we are using more than just one. We will be using two for standard SLI, and we are also including three in a TriSLI configuration. On top of that, we are going to be testing in a tripledisplay NV Surround setup that lets us test up to a resolution of 5760x1200, well beyond the 1080p we tested before. To add to that, we are including more games, this time seven games are included, and we had wanted to include an eighth, but we had difficulty getting it work as we will explain later.

The idea of going with two and three GPUs is that as the GPU count increases, the CPU must be able to keep up and provide enough performance for SLI to remain efficient. We have chosen specifically to use the GeForce GTX 580 for this because we have tested with absolute certainty in the past that CPU performance affects GTX 580 SLI and TriSLI more than it does Radeon HD 6970 CFX and TriFire. Therefore, since we know AMD FX8150 isnt exactly up to par already with Sandy Bridge from our prior testing, we think we might see some large differences in performance today.

We are only going to be comparing AMD FX8150 to an Intel i5 2500K today. In our previous testing, we found that there are no gaming performance differences between the 2500K and 2600K when overclocked to 4.8GHz. Weve already seen how even the 2500K is not much of a challenge for the AMD FX8150 in previous testing. Therefore, we started testing with the 2500K for this evaluation, and with the results we received, we saw no need to do any testing with the 2600K. Both CPUs are overclocked as high as stably possible, the AMD FX8150 is running at 4.6GHz and the Intel i5 2500K is running at 4.8GHz. We have found these to be easy overclocks to attain on these processors using moderate water cooler solutions.