17 comments

» MEDIA (8) VIDEO: Enter Commander OgreFade Enormous Explosions!! The Testing Rig The Performance Test Pop-up Performance Test Data Log The Test Results Data The Benchmarking Results Bargraph My Final Video Options Configuration Screen

Supreme Commander is brought to us by THQ and Gas Powered Games via Chris Taylor of Total Annihilation fame. As far as reviews go Supreme Commander is gathering its fair share of Kudos. This is not for nostalgia sake but because the game itself is very technically sound and entertaining. I have not seen a more robust and beautiful RTS. The sheer scale of this game is enormous, with battles involving thousands of units and positively massive explosions. Some of the humongous units are so big they cannot be transported, they have to walk across the open sea and other units are tiny specks in their wake. All endorsements aside, a major concern among potential consumers is the performance of the game. Simply put, Supreme Commander places a supreme demand on PC hardware.

My intention is to find the results of some benchmarking and testing to find out exactly what some minor tweaking and upgrading can do for performance.

Behold!

Minimum Specifications:
Operating System: Windows XP Service Pack 2 or Windows Vista
Processor: 1.8 GHz Intel or equivalent AMD processor
RAM: 512 MB
Video Card: 128MB video RAM or greater, with DirectX 9 and Shader 2.0 support
Hard Disk: Space 8 GB
Internet connection: 56.6 Kbps modem

Recommended/Optimal specifications:
Operating System: Windows XP Service Pack 2 or Windows Vista
Processor: 3.0 GHz Intel or equivalent AMD processor or better a multi-core processor
RAM: 1 GB or more
Video Card: Nvidia 6800 or better; ATI X800 XL, ATI X1300 or better
Hard Disk: Space 10 GB or more
Internet connection: Cable/DSL speed or better

These specifications are rather steep as far as recommendations go, and as we’ll see, even good midrange hardware is going to have a rather difficult time climbing the mountain that Supreme Commander is standing on.

The Test Setup
My main machine:
OS: Windows Vista Ultimate.
Processor: AMD 64 FX 4200+ Dual Core @2.2GHZ
Motherboard: Asus A8N-SLI Deluxe
Memory: 1 GB of DDR400 3200 RAM
Hard Disk: 200 GB SATA Drive.
Video: 256 MB X800 GTO.
Desktop Resolution: 1600x1050
Sound: Onboard
Network: Onboard
3dMark06: 1683

Supreme Commander comes with a performance test integrated in the game if you know how to access it. This is my primary means of measuring performance in Supreme Commander. To access the performance test simply make a copy of the desktop shortcut for Supreme Commander rename it and add “/map perftest” to the target line. For example:

"C:Program FilesTHQGas Powered GamesSupreme CommanderbinSupremeCommander.exe" /map PerfTest

Click on this shortcut to start a demo that provides a score and a log in the /bin folder. Digging into the logs produces a lot of data; I relied heavily on the composite score and the average FPS for the benchmarking demo. The included performance test is a nice way of completing benchmark testing in a reliable manner that can be replicated.

And so it begins…
I started with the Performance Test on my PC in unpatched, unmodified form, with background applications running. This netted me a score of 204.6 with an average of 7.582 FPS.
Horrified with what I was seeing, I struck out into the world trying to find the cheapest piece of hardware to get myself some immersive gaming, rather than swearing at the slow choppy experience I was getting. After looking at the most common options, (a faster CPU, a hulked up video card, or more RAM,) I decided the least expensive path, (and the one that keeps me on my wife’s good side,) was to add additional RAM.

The plot thickens…
Gas Powered Games recently released a patch that brought Supreme Commander to version 1.0.3220. I then ran the performance test with the same video configuration as the previous test. This rated resulted in a score of 5632 with an average of 10.858 FPS. As you can see there is a discrepancy in the way the ‘perftest’ works. The previous score was scored on a ‘lower is better’ scale, whereas the latest patch has changed this to a ‘higher is better’ scale. The average FPS increased significantly with this patch, but the original score is now meaningless.

I ran some repeat tests or control tests to make sure that the integrated Performance Test was actual performing a replicable sequence each time. I found that the results were very acceptable within 0.113% for scoring and 0.3% for FPS. Then I ran a total of 35 tests.
See the attached graphic for the actual testing.

Lessons learned…
During the testing, I did a lot of research to see what else I could do to squeeze a little more out of my black box. It takes a long time to run these benchmarks so I had to have plenty of time. As a rule, be sure your hard disk is defragmented. When beginning a game, turn off all background applications, including AIM, Yahoo!, the Vista Sidebar etc. Update all drivers, (include video, sound, and chipset drivers). Ensure your version of Directx is the most current. Use windows update to bring your operating system current and then turn off windows update. Install any additional hot-fix and/or optimizer for your processor if you have more than one core, then patch the Supreme Commander itself to its most current version.

After preparing your PC, if the performance is still lacking start making modifications to the games settings. The first setting to lower or turn off is the complexity of the shadows. The next setting on the chopping block would be the background image. Follow that by lowering of the detail, texture, and fidelity as your preferences desire. I started with fidelity, moved on to texture settings, and then finished by lowering detail. Next, because the main play window zooms so well, I would turn off the mini-map. If all else fails, start lowering the resolution to a level that makes the game playable. The final setting that can help make the game more playable, (but with a serious consequence,) is disabling the sound.


The Results are in… and I lose…
The baseline for this testing was 7.582 FPS. With an initial patched Supreme Commander benchmark score of 5632. The fastest settings rated 27.084 FPS, with a score of 14299. This is an enormous increase of 357%. The score increased by over 254%.
The addition of RAM did have some significant external effects; Supreme Commander loads much faster, and my system in general is more responsive. Unfortunately for me, the in-game results were mediocre at best with a 5.8% increase in performance test score and a 2.5% increase in FPS on High Fidelity. The additional ram did not show an increase in FPS when simply changing the overall fidelity settings on Low and Medium Fidelity. This is shocking! The upgrade did nothing! When looking at the custom settings tests, some of the tests did benefit from the additional RAM, but they only show a 1.5% increase in performance.
Granted in real world terms, with the optimal settings, this game was only barely breaking the 25 FPS mark. I found more substantial increases in FPS by simply turning off the in-game mini-map. The benchmarking shows that on Low Fidelity there’s a 4.8% increase, a Medium Fidelity increase of 16.5% and finally a High Fidelity increase of 6.5%. So there you have it, save $90.00 and go close the silly mini-map instead of buying more RAM.

I found that the best balance of settings on my system was:

Score 11338 FPS 20.697
Fidelity Preset: Custom
Background Image: On
Fidelity: Medium
Shadow Fidelity: Off
Anti-Aliasing: 2
Texture Detail: Medium
Level of Detail: Medium
Vertical Sync: Off
Resolution: 1024x768
No mini-map

My biggest disappointment was that I could not find a better way to play at a higher resolution with medium settings. The sounds and graphics are so good that once you see them you want to see them all the time and the lower settings lack the flavor once savored at High Fidelity. Despite that fact, I find that now that I’ve slightly optimized my system configuration and selected some new settings I am now more fully immersed in this great game. I hope that everyone can benefit from some of the increases in performance and save some money. Especially if your only reason to buy more RAM was Supreme Commander. Now I’m going back out to the doghouse where I deserve to be after running the same benchmark over 35 times and driving my wife nuts.