Test System and Benchmarks
Star Wars Battlefront seems to be a fairly enjoyable game that imparts enough of a difference from historical Battlefield titles to be a truly unique experience. The sound effects, animations and everything are generally very well done. Though there are only a few modes and only four total planets to play on, the complexity of the maps is sufficient that it looks great but also may very well play great on nearly all configurations, given proper consideration to keeping settings turned to appropriate levels.
Unfortunately this difference in gameplay doesn't mean that people of all types are going to enjoy it, and thus the lines have been drawn rather clearly in the mud. But we already know the dissent among gamers regarding Battlefront and how it might be fairly disappointing, regardless of whether we enjoyed it here. But how does it actually perform? That's perhaps the most important question.
Star Wars Battlefront performs well on anything, even lower spec GPU's.
There’s definitely a different feel to it, and it’s more than just a simple ‘update’ or re-skinning to Battlefield. In fact, though it’s casual and doesn’t have quite the variety of even other multiplayer shooters, it still remains a fun game to play. It’s casual and very entertaining. The ambiance and attention to detail is exquisite.
But how does this return to the Star Wars universe actually run on modern hardware? Being based on Frostbite, the natural inclination is to criticize the netcode and perhaps wonder how Mantle performs. But Mantle isn’t an option, and the netcode is actually well implemented, with the improvements from Battlefield 4 implemented to good effect. The actual performance is also rather obviously assumed from previous Frostbite 3 games, which have traditionally ran quite well on a variety of hardware.
Star Wars Battlefront is based on the Frostbite 3 engine, and an evolution of that as well. It’s a natural evolution that includes a swathe of higher resolution textures, a generous increase in mesh quality, more complex and natural lighting effects with better use of ambient occlusion and softer more natural looking shadows.
Specifically there are three things that have been improved for this iteration. One of those is the inclusions of compute based ambient occlusion, which is a technique that analyzes the scene for areas that should receive less ambient light. The ambient occlusion technique used in previous Frostbite games has now been improved, optimized and moved from pixel shaders to compute shaders.
Another improved technology is that of displacement mapped tessellated terrain. This is where the various terrains in Star Wars Battlefront are able to be highly detailed, using a combination of high resolution textures and geometry that is hardware tessellated and then displacement-mapped. The degree of tessellation that’s actually executed is intelligently determined based on the roughness of the terrain and the distance to the camera.
Last but not least is the inclusion of physically based rendering, or the important aspects of correctly modeling and simulating how light interacts with the surfaces and materials seen in the game.
Star Wars Battlefront will be played on the Endor Supremacy due to it’s complexity and rich use of shader-based effects like those listed above, especially tessellated terrain. The benchmark was ran at the Ultra preset with 2x MSAA enabled. Testing was done on the following system:
| Component | Selection |
|---|---|
| CPU | Intel Core i5-6600K |
| Motherboard | ASRock Z170 Extreme 4 |
| Power Supply | EVGA SuperNOVA 1300 G2 |
| SDD | SanDisk Extreme II 120GB |
| Storage Disk | Seagate 2TB |
| Memory | 16GB Crucial Ballistix DDR4 2400/ DDR4 2666 |
| Monitor | Dell P2715Q |
| Video Cards | Sapphire R9 Fury Tri-X, Sapphire R9 Nano, ASUS R9 380X Strix, ASUS R9 290X DCII Gigabyte R7 370, NVIDIA GeForce Titan X Stock, EVGA GTX 970 ACX 2.0, EVGA GTX 960 (Single Fan) |
| Operating System | Window 10 64-Bit |
| Drivers | NVIDIA - 359.00 AMD - Crimson |
Thanks to ASRock for providing the Z170 Extreme 4 for our test bench!
To start off with we'll test with a resolution that's certainly very much obtainable with a single card, as we'll see. It's possible to have a smooth and playable experience.
4K is a struggle for the lower-end cards at Ultra, though putting the settings at medium or even low can help to increase framerates just enough to make it somewhat enjoyable. Though most do indeed enjoy having everything turned on at 4K due to the increased resolution, otherwise it just doesn't seem "right". Overall a strong showing by NVIDIA and AMD as well.
1440P is far more acceptable a resolution for more GPU's, becoming even faster with the more mid-range cards, with the 380X and GTX 970 able to keep consistent and good framerates. The Titan X and Fury are stretching their legs a bit more, with the Titan X holding a clear lead in this particular case.
When we switch to a lower, less CPU dependent resolution, we see the Titan X (and thus it's not a stretch to see the 980 Ti do the same) achieve an even larger lead. But is this due to more VRAM or a more efficient use of shaders?
We'll take a look at how VRAM is used at 2160P and 1440P to see what worst case scenarios hold, and if VRAM is used efficiently across the board, or if the extra within the Titan X is squandered for whatever reason.
Quite clearly VRAM is not an issue, or not simply used up as it becomes available. Though it's very close to 4GB on all cards, and slightly higher on the Titan X, it isn't indicative of poor memory management, because there is no stuttering as would be a common symptom of swapping between VRAM and system RAM, or even the SSD or HDD. Not terrible considering assets are likely somewhat large in nature, and quite nice looking.
It had been indicated that faster system RAM might provide a faster experience overall, even dramatically. So we also tested that in what limited way we could. With only a slightly faster DDR4 RAM speed at our disposal, we put them in to see if Battlefront ran any faster as a result. It should be noted that this doesn't indicate whether even faster RAM has an effect, but at least should show if any difference has an effect on framerate.
The results are perhaps indicating the normal differences that can occur between runs. Perhaps faster RAM, not at our disposal at the moment, could make a difference, but this small of an increase wasn't significant enough to truly have an effect on gameplay. RAM, then won't make your game run faster than getting a faster GPU.
Conclusion
The obvious conclusion from this analysis is that the Titan X is fast in Star Wars Battlefront. But the less obvious is that perhaps the 980 Ti can performance nearly as well for not nearly as high a price. VRAM isn't necessarily a limiting factor, so even 6GB from a 980 Ti would be more than sufficient.
But depending on just what settings you absolutely must have turned on, it can run rather quickly on nearly any kind of GPU, from the low-end to the very high-end. Just don't expect to play at 4K on an R7 370 or a GTX 960. But you can still expect it to run quick if you're willing to tone it done a bit.
It looks good, too, even at the medium preset. Endor is lush and full of activity that's not directly the cause of the conflict you're a part of, and it's a rather interesting and fun level to play. But take the time to look around, and it's actually quite fantastic looking, and a testament to the versatility of the Frostbite engine. It might not be on the absolute cutting edge, but it still looks good.
What kind of GPU should you get? The one that your budget can afford. But really, AMD and NVIDIA are both great, and so your personal preference is all the deciding factor you need, aside from your budget.
Contents
Follow Wccftech on Google to get more of our news coverage in your feeds.
