It's here, it's FINALLY here! Battlefield V finally has RTX functions enabled through DirectX Ray Tracing, unlike another game that launched much earlier and has yet to be updated to support Ray Traced Shadows, looking at you Shadow of the Tomb Raider. But, with the October 1809 Update for Windows 10 finally being re-released in mid November we now have the ability to run DXR, and EA/DICE were ready to let it rip with Battlefield V. While we already knew that Real Time Ray Tracing (RTRT) would be insanely computationally intensive and has not yet been possible on a single graphics card, much less at all in games, the question weighing on a lot of minds is how demanding? That's something we're exploring today in this article. So let's get down to it but it is very important to note that even EA/DICE list this as experimental, and that's why the conclusion is not quite conclusive.
Testing Methodology
Testing this wasn't exactly the easiest as I had to search for a part of the game that was not only repeatable but also fairly heavy in reflective surfaces, or else there wouldn't be much point. Single player was not really that helpful, the performance varied wildly in each of the Stories, so we actually settled for a run through Rotterdam in the Conquest Multiplayer mode. Getting stars to align for a single run was less than fun if I might add. But at the end of it, we managed to get fairly consistent runs through the path traced on the below map in a scenario that starts off modest and ends pretty heavy in reflections. The Scaling Test was interesting because, thanks to Brad Chacos of PC World, we know that testing the Ultra, High, Medium, and Low presets for Ray Tracing presents an interesting bug. It starts at Ultra and in order for it to change, other than Low, you have to change the setting to Low then move off of it for it to register a change. So we actually started our testing at Low and moved up from there.
There was a bit of debate for the RTX Off vs RTX On and whether or not to have the RTX Off numbers represented by DX11 performance or not. We ultimately decided to have the RTX Off testing done in DX12 because of, in part to, the increased performance in DX12 through the latest patch for DXR and it's a more direct comparison of Off/On testing since DXR requires DX12 and cannot function in DX11.
X370 Test Bench
Components | |
---|---|
CPU | Ryzen 7 1700 4GHz |
Memory | 16GB G.Skill Flare X DDR4 3200 |
Motherboard | MSI X370 XPower Gaming Titanium |
Storage | Adata SU800 128GB 2TB Seagate SSHD |
PSU | Cooler Master V1200 Platinum |
Graphics Cards Tested
GPU | Architecture | Core Count | Clock Speed | Memory Capacity | Memory Speed |
---|---|---|---|---|---|
NVIDIA RTX 2080ti | Turing | 4352 | 1350/1635 | 11GB GDDR6 | 14Gbps |
NVIDIA RTX 2080 FE | Turing | 2944 | 1515/1800 | 8GB GDDR6 | 14Gbps |
DXR Level Scaling
Thankfully the option to use DXR is a bit more robust than just On/Off and that gives us the ability to see how it scales across settings when they actually apply as we discussed above. There is a definite benefit of going from Ultra down to Low in the performance that you get, but there is a visual hit as well. Low may impact the reflections on the gun, but it's still much better looking than Off.
RTX 2080ti
These tests were run at both Ultra Preset and DXR Ultra Preset.
RTX 2080
These tests were run at both Ultra Preset and DXR Ultra Preset.
[Un]Conclusion
Well, that's a start. This is where I normally would wrap up my final thoughts and opinions on performance, but it's kind of hard right now. Sure this is DXR and it is working, but it's also the very first implementation of it and to be quite frank, it's pretty neat. I know a lot of people are downing on it because of the performance cut, it's huge so I get it, I would have a hard time cranking everything to ultra myself and only being able to do so at 1080p is a real bummer. But, thankfully there are quality options for DXR itself. Anyone saying they can't tell the difference of it on vs off must be playing in areas with very little reflective surfaces. The argument would be better suited in a Low vs Ultra comparison, there is a difference indeed, but when the action gets cranked I couldn't tell if I was running DXR on Low or Ultra, but I could always tell if it was on or off.
The good news for RTX owners is the ability to tweak settings to their hearts' desire to achieve much better performance than what I've found so far. Taking the RTX 2080 and moving to 1440p with DXR set to Low gave about a 60 FPS experience at that higher resolution. Moving past 1440p with either card was not a pleasant experience. We don't currently have an RTX 2070 GPU on hand so we won't be able to speak to what those expectations would or should be. If you own one and this game, feel free to leave us an idea of what you're experience is like down in the comment section.
Perhaps I expected worse, so I wasn't surprised by the numbers but that doesn't mean I'm ecstatic about it. How much you notice the change in the graphics depends entirely on the area and level of reflective surfaces. But I can say, for myself at least, it's something you don't notice near as much as when you turn it back off and go back to using the screen space reflections. This right here is a fine example of why DLSS is something that should be partnered with DXR for a much better overall experience, the launch of one without the other has clearly left the community in a bit of an odd spot. Some people are excited about new technology while others are either outraged or laughing at the performance penalty of running real time ray tracing. Where do you stand? I, for one, look forward to how the rest of RTX pans out as now we see it exist in the world.