Today we’ll be comparing AMD vs Nvidia, but rather than looking at the latest $1000 or $700 or even $400 cards based on the latest chipsets, we’ll instead be paying attention to a pair of models that have been around for a while. There are a few reasons for this. First of all, we haven’t reviewed video cards in quite a while (for the reasons why, I suggest you read this blog post from last week). Secondly, when new models are released, the older ones often become available at a discounted price, or you can find them used as people dump their older models in favour of the latest hype.
Most of all, we just want to find out if a $200 video card has what it takes to power the most common gaming resolution according to Steam’s hardware survey – 1920×1080. around 1/4 of all gamers with Steam installed are using this resolution, with a total of about 96% of them using lower resolutions. In other words, if you are one of those 96%, you would probably be wasting your money buying a $1000 GeForce 690. The question is, is $200 enough? Let’s find out!
AMD vs Nvidia – The Contenders
For the purpose of this review, we’ll be sticking with one manufacturer – MSI. They produce video cards that perform better than stock, and have other features worth considering. The models are:
AMD Radeon HD6870 – MSI R6870 Hawk
Nvidia GeForce GTX 560 – MSI N560GTX/OC Twin Frozr II
Both these cards are available now for around $180-185 after a mail in rebate. The “Hawk” is actually a slightly souped up model with a few extra features we’ll talk about on the next page. Note that the GTX 560 is not a “Ti” model.
Our Testing Methods
I should make a quick note of our testing methods. As mentioned in this blog post from last week, it was the work of Scott Wasson over at The Tech Report that inspired me to start reviewing video cards once again. I had been having trouble justifying it for a long time, since I truly believe that “average FPS” is a rather meaningless benchmark. You can see why in the post we made, as well as Scott’s thorough explanation in his original article.
So we will be using the method he developed, of looking at frame time latency, as well as frames per second. We will be testing six PC games which are not completely crippled console ports. Settings used were those that would be considered high quality graphics. Advanced DX11 features and Antialiasing are used whenever possible, In other words, we are not going for pure framerate here, but a more engrossing experience. If you don’t mind a “console like” experience on your PC, a five year old DX9 card is all you would need, because the equivalent setting in most of these games would be 720×1366 with just about every advanced feature disabled.
The games we’ll be using are:
- Batman: Arkham City
- Battlefield 3
- Crysis 2
- F1 2011
- Elder Scrolls V: Skyrim
- Witcher 2
In every case, we tested actual gameplay, rather than canned benchmarks. We will show you a screenshot of each level used, and the settings used, so you know what to expect if you’d like to replicate the tests. Drivers were left at default settings, and Vsync was disabled in each game.
In all cases, a 60 second playthrough was recorded with Fraps, with the third consecutive playthrough used for the benchmark results (to ensure nothing strange happening with the test system, or playthrough mistakes).
We will also be looking at media encoding, power consumption, and OpenCL computing. We’ll talk more about these tests as we approach them in the review. For now, here is our test system’s specifications:
|CPU||Intel Core i7 3770K (Review)|
Intel Z77 Chipset
AMD: Catalyst 12.4
Nvidia: ForceWare 296.10
|Memory||8GB Crucial Ballistix Extreme 1866 MHz (Review)|
|Hard Drive||OCZ Vertex 3 240GB (Review)|
|OS||Windows 7 Professional x64 SP1|
First we’ll take a look at the two cards from MSI before moving on with performance results: