SimCity Efficiency, Benchmarked

I can remember two computer games that made me so interested in personal computers. The original Command & Conquer was the first around 1995. Under the venerable MS-DOS, at the ripe old age of 9, I spent a lot of time playing this game on our pokey HP powered i486.

I discovered SimCity 2000 shortly afterwards. The first SimCity title, released in 1989, was before my time, so I never played or watched the original. At the time, SimCity 2000 was amazing, it was extremely detailed and provided endless hours of gameplay. About five years later SimCity 3000 was released (1999) and again I spent much of my childhood playing it.

I never got into SimCity 4 (2003) for reasons I can't remember. I know I played it, but for some reason it didn't appeal to me like the previous two titles. Then came the SimCity societies and at that point I thought my days of enjoying the SimCity games were over and had been for almost a decade.

But when Maxis announced last year that a sixth installment in the SimCity franchise was coming, the hair on my neck stood on end. The announcement revealed that it was a dramatic overhaul of previous titles with full 3D graphics, online multiplayer gameplay, a new engine, and several new features and gameplay changes.

A year later, like so many others, I pre-ordered the game and waited for it to be available for download. Unfortunately, like everyone else, when the game became available and I was finally able to download it, I was no longer able to play.

As you've probably heard over the past few weeks, the game requires an internet connection, which means there is no offline mode. This in itself is very annoying, but it's much worse when the servers you're supposed to play on can't handle the demand and lock you out.

It took me several days to try, as did the thousands of outraged fans. Since we wanted to test SimCity, I really had to jump in and figure out how we were going to test the game. Fortunately, things improved on Sunday and for the next three days I set out to set up our test environment.

Test methodology

Usually when benchmarking a first person shooter it's just a matter of the game finding a good portion of the game to test until we find a section that is quite challenging. This usually takes an hour or two of game time, and then we can run the test in full. It is similar when we test real-time strategy games such as StarCraft II. In this case we decided to play a 4v4 game, record the replay and use that for benchmarking.

With SimCity, however, things were far more complex and time consuming. Since the progress of the game is saved on EA servers, it is not possible to download and use the saved game of another giant city. While it is possible to load the leaderboard into SimCity, see who has the biggest city, and check it out, we couldn't use it for testing either, since it's a live city that is being played and itself therefore the test environment is forever changed and barely controlled enough.

There are some pre-built cities like the one used in the Summer Shoals tutorial, but with fewer than 4000 residents, this isn't exactly the most demanding test environment. So we made a city with a population of half a million Sims and three more cities like on the map.

When testing StarCraft II, some readers were upset that we were testing a large 8-player card, claiming that they only play 1 on 1 and therefore perform better. That's fine, but we wanted to show what it takes to play the game in its most demanding state so you never run into performance issues.

Back to SimCity, this is a slightly different situation as all regions are the same size. Some maps have more regions than others, but they are all 2×2 kilometers long (comparable to the medium size of SimCity 4). For testing, we loaded one of our custom cities (the same each time) and increased the game speed to the max, as that's how I always play anyway. Once that was done, we started a 60 second test with Fraps, during which time we zoomed in and out several times as we scrolled through town.

As usual, we tested with three different resolutions: 1680 x 1050, 1920 x 1200 and 2560 x 1600. The game was tested with two quality configurations, which we call maximum and medium. Normally we'd test three different quality settings, but there was virtually no difference between "max" and "high" so we scrapped the latter.

  • HIS Radeon HD 7970 GHz (3072 MB)
  • HIS Radeon HD 7970 (3072 MB)
  • HIS Radeon HD 7950 Boost (3072 MB)
  • HIS Radeon HD 7950 (3072 MB)
  • HIS Radeon HD 7870 (2048 MB)
  • HIS Radeon HD 7850 (2048 MB)
  • HIS Radeon HD 7770 (1024 MB)
  • HIS Radeon HD 7750 (1024 MB)
  • HIS Radeon HD 6970 (2048 MB)
  • HIS Radeon HD 6870 (1024 MB)
  • HIS Radeon HD 6850 (1024 MB)
  • HIS Radeon HD 6790 (1024 MB)
  • HIS Radeon HD 6770 (1024 MB)
  • HIS Radeon HD 6750 (1024 MB)
  • HIS Radeon HD 5870 (2048 MB)
  • Gigabyte GeForce GTX Titan (6144 MB)
  • Gigabyte GeForce GTX 680 (2048 MB)
  • Gigabyte GeForce GTX 670 (2048 MB)
  • Gainward GeForce GTX 660 Ti (2048 MB)
  • Gigabyte GeForce GTX 660 (2048 MB)
  • Gigabyte GeForce GTX 650 Ti (2048 MB)
  • Gigabyte GeForce GTX 580 (1536 MB)
  • Gigabyte GeForce GTX 560 Ti (1024 MB)
  • Gigabyte GeForce GTX 560 (1024 MB)
  • Gigabyte GeForce GTX 550 Ti (1024 MB)
  • Gigabyte GeForce GTX 480 (1536 MB)
  • Gigabyte GeForce GTX 460 (1024 MB)
  • Intel Core i7-3960X Extreme Edition (3.30 GHz)
  • x4 4 GB G.Skill DDR3-1600 (CAS 8-8-8-20)
  • Gigabyte G1 Assassin2 (Intel X79)
  • OCZ ZX series 1250w
  • Crucial m4 512 GB (SATA 6 Gbit / s)
  • Microsoft Windows 7 SP1 64-bit
  • Nvidia Forceware 314.14
  • AMD Catalyst 13.2 (Beta 7)

Leave a Reply

Your email address will not be published. Required fields are marked *