Earlier this month, we tested the graphics performance in Battlefield V, covering a variety of graphics cards with 1080p, 1440p and 4K. We also looked at real-time ray tracing performance. For these tests, we used the single player campaign to measure performance because it's easy to record accurate data and the single and multiplayer parts of the game are visually very similar. However, there is a big difference in CPU demand between the two modes, and the 64-player multiplayer mode hurts lower-end processors.
To test CPU performance, we use Battlefield V's 64-player conquest mode, which brings with it a number of challenges for detailed testing. Nevertheless, we took up the task. The fluctuations between runs can be very large compared to our Battlefield V single player GPU benchmark, which often delivers the same average frame rates and 1% low results.
With high-end hardware, a deviation of more than a few frames is rare, although testing in multiplayer mode often found a difference of up to ~ 10 fps for the average frame rate. We neglected extreme outliers and tested much more often than usual to determine the most accurate performance. The results presented in this article are based on an average of 6 runs instead of 3.
If you have 30 players nearby, the frame rates are much lower than if the action takes place in the distance. So making sure that the same things going on around the player character for each round was a serious challenge and very time consuming. For the test, we used the Narvik card in conquest mode for 64 players for about 60 seconds with each run.
We use DirectX 11 for most benchmarks because it offers more consistent frame-time performance. We know that many players claim that DX12 is solid and even better than DX11, but if you don't use a low-end CPU, DX11 offers a better experience. However, to get this discussion out of the way, here are the DX12 numbers for those interested. We will not discuss these results in detail as we will focus on the slightly more powerful DX11.
Here are all the CPUs that have been retested with the DX11 API, and as you can see, Ryzen cops hit a bit under these test conditions. The Core i5-8400 is roughly equivalent to the Ryzen 7 2700X, while the older 7700K is faster. We also see processors like 8600K, 8700K, 9600K, etc. that easily outperform the best AMD offering. At well over 60 fps on a consistent basis, Ryzen was able to achieve smooth performance regardless, but we'll move on to some cheaper tests in a moment.
It is interesting to see how the once powerful Core i5-7600K fights here. The Ryzen 5 2400G was comparable to the 7600K and we don't see that often, if at all. Then we see how the Ryzen 3 2200G comfortably beats the Core i3-7350K and the Pentium G5400. Basically, everything dual-core is destroyed by the 2200G quad-core, even if you have hyper-threading.
So a decent performance from AMD at the bottom, but not great at the top. What does Ryzen have to offer if you are not using an RTX 2080 Ti with 1080p and deactivated ray tracing function?
Here is a comparison between the Ryzen 7 2700X and the Core i9-9900K. Above in the graphic we see the previous RTX 2080 Ti result at 1080p. Here, the Intel CPU offered a performance increase of 16% for the average frame rate and 18% for the 1% low, which is a decent performance advantage from Intel.
The move to the RTX 2080 hasn't changed much, and we're still mostly tied to the 1080p CPU. Here the 9900K was 11% faster for the average frame rate and 15% faster for the result of the frame time. For those looking for maximum performance at 1080p, the 9900K seems to be the way to go.
However, if you use a medium-range GPU like the RTX 2070, you will get a performance similar to that of the 2700X or 9900K. Here, the Intel CPU was 2% faster for the average frame rate and 5% faster for the 1% low, which is a negligible difference.
When using a GPU that is slower than the RTX 2070, no difference is found between these two CPUs when using the ultra quality settings at 1080p, and we see that this is the case with the GTX 1070. But what if you want to play at 1440p and that resolution? seems to be more suitable for all four GPUs tested here.
Since we're limited to the GPU, at 1440p we see very little difference between the 2700X and 9900K with the GTX 1070, RTX 2070 and even the RTX 2080. The 9900K still offered a superior gaming experience at 1440p if that RTX 2080 Ti was 13% faster on average with a 21% higher frame time result.
Assuming you've bought the 2080 Ti for smooth 4K gaming, the choice of CPU doesn't seem to play a big role in this game when comparing high-end AMD and Intel chips. Even with the RTX 2080 Ti, both CPUs enabled the same level of performance and achieved an average of 80 fps at a 1% low of 65 fps.
Battlefield V can be played on quad cores, but you can expect frequent frame dips, which results in less consistent frame rates. We've largely noticed that the older Core i5-7600K works pretty well with the AAA titles released in 2018, but for the multiplayer part of Battlefield V, you should avoid the big 64-player battles.
This also means that for a smooth, consistent gameplay, the Core i3 series along with the quad-core Ryzen CPUs is a depreciation. If you are willing to accept regular stuttering and immersion below 60 fps, you can of course still get by with these processors.
If we only built a PC to play Battlefield V at the highest possible frame rates without paying attention to anything else, we would probably get the Core i5-8600K. For just $ 10 more, the Ryzen 7 2700 offers you a better upgrade path across the board, a cooler, and better value for all computing. If you're using an RTX 2070 or slower with the ultra quality settings, you're bound to the GPU and it doesn't matter which processor you use. Alternatively, if you work with high or even medium settings for maximum frame rates, you will find a greater advantage if you use the Intel CPU with higher clock speed and lower latency.
Overall, we think Ryzen looks a little disappointing at 1080p where the game isn't limited to the GPU, although it's impressive considering the clock speed deficit that the 2700X is only 15% slower than the 9900K. Note that both AMD and Intel CPUs can be overclocked for higher performance and storage times can be adjusted manually, which is beneficial on both platforms. In general, AMD is better at optimizing memory, while Intel benefits more from core overclocking.