We finally got into Battlefield V for the first time this week, and of course we spent more time benchmarking than gameplay. So we have a couple of results for you … which we call preview because we're testing the & # 39; Open Beta & # 39; version of the game. The whole thing will be published on all major platforms in two months.
The purpose of the open beta is to test technical aspects such as server stability, latency and matchmaking as well as gameplay elements such as weapon balancing.
For those who pre-ordered the game – and we never recommend you pre-order until reviews are published – but the purpose of our pre-order is to let you know how well the game is doing in its current form or not. In any case, those who pre-ordered the game now have early access, while those who don't have access get access two days later.
<iframe allowfullscreen = "" frameborder = "0" height = "390" src = "https://www.youtube.com/embed/UZW4cPUIVf4?rel=0&showinfo=0&modestbranding=1&vq=hd720&autohide=1&autoplay=1" width = 560 "load =" lazy "srcdoc ="
We won't waste time explaining what Battlefield V is or giving an introduction. We just assume that by the time you read this, you will know what it's about (let alone the amazing RTX-enhanced graphics that we can't test yet). We'll just say that Battlefield V, like Battlefield 1 from 2016 and Battlefield 4 from 2013 – yes, the naming is confusing – uses the same third iteration of the Frostbite engine. Countless other quality games have also made use of the engine over the years.
Battlefield V's PC beta lets players customize the quality and filtering of textures, lighting, effects, post-processing, mesh, terrain, undergrowth, anti-aliasing, and environmental occlusion. The beta also includes options for resolution scaling, UI scaling, FOV sliders, motion blur, lens distortion, chromatic aberration, vignette and more.
As expected, limited maps and game modes are available in open beta. However, this does not matter for tests as we would only use one card anyway. Getting into a game was pretty difficult since all the servers were constantly full, but we always managed to get where we needed to be to test.
We chose Grand Operations mode on the Narvik map and had to load the games until we ended up in the right team in the right section of the game so we could do the right pass for each GPU. It was messy, but we managed to test almost a dozen latest-generation GPUs with three resolutions. We also need to go through some quality scaling data.
We use our GPU test bench, which is powered by a Core i7-8700K with 5 GHz and 16 GB DDR4-3400 memory. The WHQL driver 399.07 was used for the GeForce GPUs and the Radeon Adrenalin 18.8.2 driver for AMD.
Before we get into testing too much, we decided to figure out how to compare the DX11 and DX12 APIs in this new title. In this way, if AMD worked best with DX12 and Nvidia with DX11, we would test the GPUs of all companies in the future. However, we quickly found that DirectX 11 was the way to go for both teams.
When using DX12, the game stuttered heavily on both AMD and Nvidia GPUs. This is something we saw when Battlefield 1 was released. We suspect that this is not entirely surprising. Developer EA DICE clearly does not prioritize the low-level API. For the rest of this article, we will be testing with DirectX 11.
First we got the 1080p results with the default of ultra quality and here the RX 580 was up to 14% faster than the GTX 1060 6 GB, the Radeon graphics card also kept over 60 fps in our test. Vega 56 was also 8% faster than the GTX 1070 at the average frame rate, but suffered from a larger discrepancy between the average and the 1% low result, which made it 5% slower for frame time performance.
This weaker than expected frame rate performance was found in both the Vega 64 models and both GTX 1070 Ti models, although the average frame rate was exceeded at 1080. Some drivers still work from AMD, but no surprises since we're only testing the beta version of the game.
The GTX 1080 Ti also sees a pretty big difference between the frame time and the result of the average frame rate. This means that we are likely to encounter a system bottleneck here as the low 1% has broken the 100 fps barrier.
When we switch to 1440p, we see that the gap between the 1% low and the average frame rate for the 1080 Ti has narrowed, probably due to the fact that we are more tied to the GPU here and this is the pressure from the previous one performance-limiting component relieved. maybe the CPU or the memory.
Again, we see that the Vega series is somewhat inadequate in terms of frame time performance, which is a shame as we see competitive average frame rate results. Hopefully AMD can fix this before the games are officially released.
At 1440p, the RX 580 and GTX 1060 can deliver playable performance, although the RX 580 was significantly better. However, if you are looking for around 60 fps for the 1% low result, the GTX 1070 is for you, and for a smooth game with these resolutions this is really all you need.
With the extreme 4K resolution, you either have to lower the quality settings or at least bring a GTX 1080 Ti, but even then the performance wasn't amazing, playable, just not amazing.
Here's a quick look at how the presets are compared to four of the previously tested GPUs. Please note that we had to manually configure the preset each time because for some reason the developer decided to embed the vsync option in the presets. For some of them it is forced, very strange, and we hope that they will fix it before the release.
In any case, the results are somewhat unexpected. Switching from Ultra to High has only increased the performance of the RX 580 and GTX 1060 by around 8-9%. We saw a slightly larger 11% increase in the GTX 1080 and a much larger increase of 23% in the Vega 64, although the frame time performance was still lower than that of the GTX 1080 with the same quality settings.
Then we saw an increase of around 40% in all the GPUs tested, and now Vega 64 and the GTX 1080 were able to maintain over 100 fps at all times during our test.
Battlefield 1 was one of the first games to really kill the quad-core CPU, 64-player multiplayer Battlefield 1. As expected, Battlefield V isn't easier for quad cores. We have regularly seen my Core i7-8700K test system, in which all cores clocked at 5 GHz have a load of more than 70% and sometimes have a peak value of over 80%. Good luck with four threads.
VRAM usage was not that extreme. With a GTX 1080 Ti, only 3.4 GB were allocated at 1080p and around 3.8 GB at 1440p. This explains why the 3 GB 1060 did so well at 1080p. You should make sure your gaming rig has 16 GB of RAM since memory usage exceeds 8 GB. Again, we saw the same situation with Battlefield 1, where 8 GB of memory would stutter when playing the multiplayer part of the game.
The good news is that the game looks great and works very well on mid-range hardware. Even the 3 GB GTX 1060 offered a solid gaming experience at 1080p. We didn't have time to test other lower-priced GPUs like the GTX 1050 Ti and the RX 570. We'll save the big GPU bash for 50 when the game is done.
At the moment, however, it's nice to know that those who rock a GTX 1060 or RX 580 can look forward to a lot. Even at 1440p the game can be enjoyed in all its glory with Vega 56 or a GTX 1070 and we expect things to only improve from here unless you use ray tracing 🙂
As we found, you can reduce the quality settings to increase frame rates, although for the most part, switching from ultra to high didn't bring much of an advantage. For this you have to go to medium. It is difficult to predict whether we will see major performance improvements until the game is released, whether current performance is what you could call "beta", whether the display drivers will improve things, or whether it will. Another variable is that multiplayer is less consistent than single player, although an average of three runs is given.
The stuttering DirectX 12 performance was unexpected, but not entirely surprising, since we found exactly this problem when we first tested Battlefield 1 in 2016. We were hoping the developer would be more capable, but we are certain that we & # 39; I will see that this will be addressed soon. For those of you who are wondering, we still have to try multi-GPU technology like SLI, but that along with other CPU tests (8600K, 8400 and then a number of Ryzen CPUs) will wait for the incoming CPU tests that we carry out I cook for the official release of the game.