There's probably no single PC gaming title that drives more PC component sales than Battlefield, and with Battlefield 1, the latest installment in the series, Battlefield is back in style. This is the series that inspired us many years ago to try to squeeze every last bit of performance we could out of our gaming PCs. And the good news is that the folks at DICE, the developer of BF1 and every BF title that came before it, are committed to making PC games that run beautifully on modern hardware, and can even provide a decent experience on older gear. The combination of amazing graphics, serious graphics tweaking options, and unbeatable online gameplay make Battlefield a standard by which all other PC games must be judged.
While this is not a game review, we'll say that based on our many hours in the game so far, it's pretty clear that Battlefield 1 is a well-designed and well-polished game. After years of playing in modern times, we were a bit anxious about a rewind to the World War I era. Would the vehicle-based combat be as dynamic? And what about the range of weaponry available? Well, rest assured that this game is very bit as exciting as Battlefield games set in the present (or in the future for that matter), and one of the reasons for that is that the game is grounded in actual events. This makes the single-player missions particularly compelling, as they are telling stories that are grounded in reality. Sometimes, it feels like you're watching a serious WWI action movie. Similarly, the multi-player combat feels quite balanced, with none of the overpowered vehicles found in previous entrants in the series. And as for the range of classes and weapons, there's practically a science-fiction feel to the game, in that the real-life weapons modeled in the game almost seem other-worldly. It's amazing (and terrifying) to realize that over 100 years ago, military combat already utilized such destructive weapons of war. In a sense, BF1 is a cautionary tale that nations have failed to heed over and over again: war's destructive power rarely leads to a lasting peace.
We performed our benchmarks on two test systems. First up, our Z170-based platform, using two different quad-core CPUs:
- CPU #1: Intel Core i5-6600K (overclocked to 4.4GHz)
- CPU #2: Intel Core i7-6700K (overclocked to 4.4GHz)
- Motherboard: Gigabyte GA-Z170X-Gaming 6
- RAM: Corsair 2x8GB Vengeance LPX DDR4-3000
- Solid-State Drive: Samsung 850 Evo M.2 500GB
- Solid-State Drive #2: Crucial MX200 1TB
- Power Supply: EVGA Supernova GS 850W
- Case: Phanteks Evolv
- CPU Cooler: Noctua NH-U14S
- Operating System: Windows 10
Secondly, our X99-based platform, using an octo-core CPU
- CPU: Intel Core i7-6900K (overclocked to 4.3GHz)
- Motherboard: Asus X99-Pro/USB3.1
- RAM: Corsair 4x8GB Vengeance LPX DDR4-3200
- SSD #1: Samsung 950 Pro M.2 512GB
- SSD #2: Samsung 850 Evo 1TB
- Case: SilverStone Primera PM01
- Power Supply: EVGA Supernova 1000 PS
- CPU Cooler: Corsair Hydro H100i v2
- Operating System: Windows 10
Here are the nine video cards we tested (with links provided to all cards that are still sold):
- Sapphire Radeon R9 270X 2GB Dual-X
- Sapphire Radeon R9 290 4GB Tri-X
- Sapphire Radeon Fury 4GB Nitro
- EVGA GeForce GTX 980 4GB ACX
- EVGA GeForce GTX 980 Ti 6GB ACX
- Asus GeForce GTX 1070 8GB Founders Edition
- EVGA GeForce GTX 1080 SC 8GB
- Nvidia Titan X Pascal 12GB
- 2x EVGA GeForce GTX 1080 SC 8GB with EVGA PRO SLI Bridge HB
Note that while some of these cards came pre-overclocked from the factory, we ran all cards at their reference clocks to keep things as fair as possible. We tested with a range of video cards representing what we think our readers are using today. Now, that does mean that AMD tends to cluster at the lower-end, while Nvidia is at the high-end, but that's more or less how things break down. The eagle-eyed among you will note that we don't have AMD's latest card, the RX 480, represented in our collection, but we hope you'll rest easy knowing that we included the significantly faster R9 Fury, which is still the fastest air-cooled card AMD has ever released. We think AMD fans will be pleased with the performance numbers it puts up in this test.
Among our cards, we have models with 2GB, 4GB, 6GB, 8GB, and even 12GB represented. We'll be providing some basic VRAM utilization numbers to help readers determine how much is really enough. This article will also serve as our unofficial "SLI vs. Titan X Pascal" shootout. There's a lot of folks out there who believe that SLI isn't worth using, and that if you're building a seriously-high-end gaming rig, the Titan X Pascal is the superior choice to GTX 1080 SLI for around the same price. Well, we'll let you decide for yourselves after seeing our extensive benchmarks.
We also want to make clear that we purchase all of our CPU and video card samples at retail. We do this to avoid any potential bias that may result from accepting a certain vendor's products in a highly-competitive market. If we love a CPU or GPU, we'll tell you, and if we think it's a dog, we'll tell you that as well. If you support this type of testing, please consider using any of the product links we provide throughout articles on this site the next time you are looking to purchase a tech-related item.
While we don't typically run benchmarks on early release software (due to the many updates that are sure to follow), we bet a lot of shoppers are heading into the holiday season eyeing what kind of system will give them the best BF1 experience. Well, we're going to figure that all out for you here. Of course, in the process, we got a bit burned, as DICE released a new game patch on October 24th, three days after the game's release and half way through our testing. Likewise, Nvidia surreptitiously dropped a replacement for its "game-ready" driver (375.63) on October 23rd, days after the original BF1 game-ready driver (375.57) appeared on its website. That left us having to re-test with the new driver, delaying this article by several days. That's life on the cutting edge! But we want to bring you both the freshest and most accurate test benchmarks, so we hope the delay is worth it to our readers. AMD's most recent driver, Crimson Edition 16.10.2 Hotfix, was released on October 19, 2016, in anticipation of Battlefield 1, and did not change during our testing.
While there's a lot of discussion right now about the DX12 option within Battlefield 1, we stuck with DX11 for two reasons. First, it's easier to test with, as the tried-and-true FRAPS application, which record average and minimum framerates, only works under DX11. Options for recording such information under DX12 are much more labor-intensive. Secondly, the DX12 code path in BF1 is typically slower on all hardware than the DX11 code path. This is starting to become a pattern in a lot of games, and it means that for now, DX12 is still a work in progress. That does not mean it doesn't hold a lot of promise; indeed, we're sure that DX12 will become the de facto standard soon enough. It will just take developers time to figure out how best to make use of its potential.
To collect our data, we repeated three 30-second in-game scenarios for every test we ran, trying to follow as similar a path as possible. Note that in our multi-player benchmarks, there are factors we have no control over, such as the activity of other players on screen, as well as the point during the match at which we're testing. Due to dynamic weather and destruction, a map will look (and play) completely different at the beginning of the round than it does at the end. The map we chose, "Empire's Edge", changes drastically, with buildings crumbling to the ground and a dense fog rolling in. There's just no way this can be controlled for in benchmarks on live servers. The only recourse would be to only benchmark at the very beginning of a round, which for obvious reasons is an impossibility when testing around 50 different hardware configurations! In short, there's a pretty wide margin of error when it comes to multi-player testing, but we think you'll still be able to discern patterns that make clear which hardware is best at various pricepoints.
So, without further ado, let's turn to our CPU benchmarks, which will determine which platform we'll be using to get the most out of our stable of video cards!