Benchmarking with Fraps: A Basic Example
Before beginning any benchmark, a key must be setup in Fraps that will begin and end the benchmarking utility. On the Frames per Second (fps) tab the user must assign a benchmarking hotkey. This hotkey may be any of the keys found on the standard Windows keyboard but it is advised that the user be certain no other function both in the game and in Windows, is assigned to this key. Good candidates include the “scroll lock", “pause", and any of the upper F-keys (F9 through F12) as these are usually not assigned to other functions. In the screenshot of Fraps above, I have arbitrarily assigned the “w" key as the benchmarking hotkey.
For this basic example, I have decided to capture just 60 seconds of gameplay data so I have checked the “stop benchmark automatically after" checkbox and I have assigned a value of 60 seconds to this option. I have also checked the “FPS" option under the three detailed benchmark statistics options available at the bottom of the Fraps window.
I have decided to run a benchmark while playing Crysis Warhead, a recently-released first-person shooter from Crytek that is generally known to excessively tax a gamer’s computer. When I am ready to start the benchmark in-game, I will hit the “w" hotkey and continue playing as usual while Fraps gathers the data automatically. I am playing with the following general video settings:
Video Settings: Mainstream (synonymous with medium settings)
It is beyond the scope of this article to discuss the meaning of all these settings. I normally play the single-player portion of Crysis Warhead at these settings. I chose a particularly high-poly count area of the first mission in the game. After running the benchmark, the results can be found in the “c:/Programs Files/Fraps/benchmarks" directory or wherever the user has installed fraps. The resulting file can be opened in a spreadsheet for viewing and further analysis.
The results from this example test show that the maximum fps I got was 94 and the minimum was 34. This reveals that under the settings stated above, my computer is generating enough fps to allow smooth gameplay under resource-taxing conditions in Crysis Warhead. Had the results shown an inadequate minimum fps, I could lower the in-game video settings and run another benchmark. In addition, since the raw data may be opened in a spreadsheet, further statistics such as mean, median, mode, geometric mean, standard deviation, etc. could easily be calculated and used to create a custom benchmarking profile.