Benchmarking Standards

Some may think that UMark may never be a benchmarking standard for Unreal® Tournament 200x, because it's openness to configurability lets the users run pretty much any kind of test they want. This kind of flexibility places the user in their own benchmarking realm, which strays away from standard testing. This can be argued as to be true because you can't let standards have so many variables, but let's get some of the facts straight about UMark. The following explains how the standard Unreal® Tournament 200x Benchmarking elements have been built into this program.

When you choose "High Performace" or "High Image Quality", you are telling UMark to use the official standard .ini files from the UT2003\Benchmark directory that came installed with UT2003. When "High Performace" is selected, it uses "MinDetail.ini" and "MinDetailUser.ini" configuration files to run the tests with. Respectively, when "High Image Quality" is selected, it uses "MaxDetail.ini" and "MaxDetailUser.ini". These configuration files have not been edited, tweaked or tampered with, so they reamain standard. While UT2004 didn't include standard .ini files for benchmarking, UMark has it's own files for UT2004, in which will hopefully become the de facto standard for UT2004 benchmarking. Users are still welcome to use non-standard .ini files with the "Custom Ini File" option.

You'll notice in the UT200x\Benchmark directory, there are several batch files (included in the offical UT200x retail release) that run benchmarks on UT200x maps. These batch files each run a different map, and there are some for botmatches, and some for flybys. Now, in UMark, there are preset map lists that include the same maps as all the batch files.

UMark is not built upon it's own benchmarking engine, it uses UT200x's benchmarking engine to run the tests. Running benchmarks with UMark is just like running UT2003.exe or UT2004.exe with benchmarking options passed to it, but without having to type out complicated and lenghtly command line options or write batch files.

Choose any resolution(s) you want, and any detail you want. Leave the number of bots at the default (which is 12). Show the batch controls and call up the "Preset Tests" dialog. Select the preset test called "Unreal Tournament 200x Botmatch". Click the "Benchmark" button, and we now are running a standard UT200x benchmark, which is equivalent to running each of the official botmatch .bat files in the UT200x\Benchmark directory.

Although UMark is very botmatch-centric, it still is capable of running fylby benchmarks. One reason why support for flyby benchmarks aren't parallel with botmatch support, is that a limited number of maps support flyby benchmarking properly. Another reason is because flyby's aren't as accurate as botmatches when comparing your framerates to "real" gaming. To run UMark to imitate the official fylby batch tests, simply follow the instructions above, but with 2 differences: Choose 0 bots, and select the "Unreal Tournament 200x Flyby" preset test.

Now the advantages of using UMark over benchmark.exe or the batch files are very clear and simple. UMark is highly configurable, yet easier to use, and it runs the same benchmark. Another plus is the results management system that even keeps track of detailed information about each benchmark (more comprehensive than the offical tools provided with UT200x), paired with a comparison engine for bar graphs and an online scores database. It is an efficient, graphical way of running standard and custom benchmarks on Unreal® Tournament 2003 and Unreal® Tournament 2004.