World in Conflict

The first game we're going to look at represents RTS (Real Time Strategy) games. World in Conflict is undoubtedly one of the most impressive looking RTS games available. The unit and terrain detail is simply breathtaking for this genre of game. I got my first peak at World in Conflict two years ago at E3 in Vegas. The demonstration of tanks rolling over hills while bombers zoom past while beautiful special effects drop napalm, is breathtaking (not that I like napalm, I'm complimenting the visual effects!).

There are two different ways developers make RTS games. Some restrict the amount of units you can have via gameplay methods and others have an open free-for-all where the only limitation is the raw power of your computer. The latest style RTS games have always been of interest to me, as they truly test the capabilities of your PC. It's cool to have the ability to create five hundred units to trump your enemy but having the game perform like a slideshow gets boring very quickly.

World in Conflict takes the liberty to control more of your actions and restrict your gameplay style more than any other RTS game I've seen before. They've had to do this to allow the level of detail found in the game, to run on mainstream PCs. I have read countless reports of gamers being upset when the game would run like a slideshow, even with the most basic graphics settings. There may be a few more complaints about how they have tried to change the classic working structure of RTS games, in the way of resource collecting and building. Regardless of my impression of the gameplay, the game is a good place to start the benchmarking.

There are two sets of benchmarks I've run for World in Conflict. One uses the built in time demo style benchmark which you can run from the configuration control panel. The other is an actual real gameplay test. I played through a big chunk of the game, looking for a decent level to test on - a level that was structured enough that I could reproduce similar style gameplay time and time again.

Playtest

When I got to level 11, Aftermath, I discovered a unique level with lots of units, with a set series of goals to achieve. This allows for a structured style gameplay that I could reproduce. It also has a lot of activity with combat and unit movement.

Fraps was run for 160 seconds during gameplay, during which I achieved the initial goals of the mission. Fraps recorded the framerate during every second of that gameplay and then calculated the minimum, maximum and average from the test. We took the the minimum, maximum and average and put them into a simple to understand graph for you. I feel that all three figures are important.

With the quad core installed, I adjusted the graphics settings to the highest possible level in which the game was perfectly smooth, which was "Very High". Then when I came to test dual core, I left the settings at the same level. So the figures you see here are run with identical settings. I may have been able to tweak individual modules in the custom settings but I thought having the settings at a default option would be easier for people to relate to. As we're testing on a Windows XP 64bit operating system, these tests are run in DirectX 9.0 mode and not DirectX 10.

Although it may seem like the dual core was faster than the quad core, I must say that every time I tested I got a variance of 3-5 fps on the average scores and there was really no performance difference to the gameplay, no matter how many times I tested. The results you see here are just the last set of results I ran on each CPU. So the quad core did absolutely nothing to increase performance in gameplay.

World in Conflict playtest - dual core.

Looking at the screenshot of the CPU utilization, you'll notice that the dual core CPU ran at a maximum of 55-56° C. Although there was a spike at the start on the second core, both cores seem to run anywhere between 65-85% utilization. At no stage was the dual core bottlenecked.

World in Conflict playtest - quad core.

This image might explain why there was no performance difference. It looks like the first two cores were doing literally nothing during the test. This is one of the major issues with performance in dual and quad core systems. If an application is not written to take advantage of the extra cores then you will see no performance increase. The graph speaks for itself. The quad core extreme edition also ran a little hotter than the dual core, peaking between 72-74° C.

Official Benchmark

The official benchmark showed very similar results. The only major difference was at the minimum level. The benchmark does things that you don't actually see in the game. It uses camera angles that you would never see. The minimum fps on the dual core was 16, where as the quad core only went as low as 22. This slump in frame rates was during a split second of a massive explosion which takes out an entire town during the benchmark. Frame rates very quickly pick back up on both CPUs and ran smoothly for the rest of the benchmark. A difference of 4fps for the average score with the official benchmark is nothing in this level of gaming. I severely doubt you'd see any type of difference if you watched the two tests playing on the systems side by side.

World in Conflict official benchmark CPU utilization graph - dual core.

World in Conflict official benchmark CPU utilization graph - quad core.

Looking at the CPU graphs, there's no doubt that the dual core was working a lot harder than the quad core during this benchmark. Although the dual core was not working at 100% load, it was very close to it. The quad core on the other hand, looks like it was sitting back taking it easy, with plenty of power in reserve.