Nvidia’s Turing architecture promised to bring the holy grail of real-time ray tracing to desktop graphics hardware by blending the best of rasterization and ray tracing into a new form of hybrid rendering platform, but that revolution took longer to arrive than the company probably hoped for. Although GeForce RTX cards have been available to gamers for a couple of months, Microsoft only just finished exorcising a troublesome file-deletion bug that inadvertently delayed the release of the essential-for-RTX Windows 10 October 2018 Update. On top of that, developer DICE and publisher EA pushed back the release of the headlining RTX title, Battlefield V, from October 19 to November 20. Gamers who sprung for Nvidia RTX-ready hardware have understandably been getting antsy for the day when they could finally fire up their cards’ RT cores and trace some rays for realistic takes on effects like reflections, ambient occlusion, global illumination, and more.
The GeForce RTX 2080 and RTX 2080 Ti, powered by the ray-tracing-accelerating Turing architecture
Last week, the support structure that the RTX hardware-and-software stack needed to work finally began to fall into place. As we noted above, Microsoft made the Windows 10 October 2018 update available again after a long period of review and retesting, and DICE began opening up early access to Battlefield V for those who had paid for the privilege. Nvidia also released a corresponding driver to get Turing cards’ ray-tracing hardware up and running in that game. We ponied up for the Deluxe Edition of Battlefield V to get into the game early, and we’ve been working hard over the weekend to see just how RTX effects look—and perform—ahead of that game’s release to a wider audience tomorrow.
As we hinted at above, titles that incorporate RTX effects can use Turing cards’ ray-tracing hardware in a variety of ways. In the case of Battlefield V, DICE has opted to use the technology to improve the quality of certain reflections on cars, glass, small puddles of water, stone flooring, and more. The best way to appreciate these reflections is in motion, so I’ve created two 4K gameplay videos to try and show the differences.
To test the performance and appearance of RTX effects, we walked through a portion of the game’s “Tirailleur” single-player mission with an abundance of standing water for light and objects to reflect from. To establish a baseline for what to look for in our RTX-enabled walk through the woods, I traveled through the scene with RTX off and Battlefield V‘s settings otherwise left on the ultra preset.
I’ve stopped in a few places in the video above to show some of the ways that the game’s default reflections fall short of realism. You’ll especially want to note how leaves passing over puddles create distracting flickering between dark and light tones in the water, as well as comet-trail-like artifacts in some cases. To my eye, there’s a lot happening in these reflections, but they often fall short of realism in ways that break the sense of immersion to the watchful eye.
With RTX on at the game’s low preset, the appearance of reflections in these puddles becomes more subtle, and to my eye, more convincing. Part of that may be because the RTX reflection algorithm doesn’t seem to try to account for the leaves blowing across the puddles of water in the scene, as the default reflections do, simply because casting that many rays on fast-moving objects may be too much to ask at the moment. Still, the fainter images of trees, structures, and such that do get reflected seem more true to life, in my estimation, and there’s much less of the jarring alternation between quite light and quite dark values of tone that we can sometimes see in Battlefield V‘s default reflection rendering.
Despite the photorealism promised by ray tracing, RTX reflections aren’t free of artifacts of their own. As we approach and move up the hill past the guard shack in this video, you might see some of the puddles exhibit what appears to be noise. These artifacts appear as “sparkles” or cross-shaped areas of bright pixels whose source isn’t immediately evident. I didn’t see this kind of noise too often while putting my nose in this section of the game, but it was evident enough when it did appear that I was reminded of the fact that I was looking at a screen. Whether this noise is something that can be worked out with future software updates and improved use of RTX resources will remain to be seen, but at the moment, it is a reminder that we haven’t quite grasped the holy grail of computer graphics even in the limited context of hybrid rendering.
Now that we have a basic idea of how RTX effects look in practice in Battlefield V, let’s see what those improved reflections ask of the hardware used to generate them.
Our testing methods
If you’re new to The Tech Report, we don’t benchmark games like most other sites on the web. Instead of throwing out a simple FPS average (or even average and minimum FPS figures)—numbers that tell us only the broadest strokes of what it’s like to play a game on a particular graphics card. We can go much deeper. We capture the amount of time it takes the graphics card to render each and every frame of animation before slicing and dicing those numbers with our own custom-built tools. We call this method Inside the Second, and we think it’s the industry standard for quantifying graphics performance. Accept no substitutes.
What’s more, we don’t rely on canned in-game benchmarks—routines that may not be representative of performance in actual gameplay—to gather our test data. Instead of clicking a button and getting a potentially misleading result from those pre-baked benches, we go through the laborious work of seeking out test scenarios that are typical of what one might actually encounter in a game. Thanks to our use of manual data-collection tools, we can go pretty much anywhere and test pretty much anything we want in a given title.
Most of the frame-time data you’ll see on the following pages were captured with OCAT, a software utility that uses data from the Event Timers for Windows API to tell us when critical events happen in the graphics pipeline. We perform each test run at least three times and take the median of those runs where applicable to arrive at a final result. Where OCAT didn’t suit our needs, we relied on the PresentMon utility.
Here are our in-game settings for Battlefield V. Although it’s not captured in the screenshots below, we varied the game’s “DXR Raytrace Reflections Quality” setting, as well as the resolution, to achieve our desired test configs.
Battlefield V is a work in progress (recall that it launches to the general public tomorrow, November 20). As early versions of Battlefield V can apparently have issues saving and applying settings for the DXR Raytrace Reflections Quality parameter, our results should not be compared to sites that did not manually verify the DXR Raytrace Reflections Quality settings they were using. We verified the settings we chose were being saved in the configuration file by changing the parameter with a text editor, saving the file before launching the game, and checking that our change was reflected in the game settings menu.
As ever, we did our best to deliver clean benchmark numbers. Each test was run at least three times, and we took the median of each result. Our test system was configured like so:
|Processor||Intel Core i9-9980XE|
|Motherboard||Asus Prime X299-Deluxe II|
|Memory size||32 GB (4x 8 GB)|
|Memory type||G.Skill Trident Z DDR4-3200|
|Memory timings||14-14-14-34 2T|
|Storage||Intel 750 Series 400 GB NVMe SSD (OS)
Corsair Force LE 960 GB SATA SSD (games)
|Power supply||Seasonic Prime Platinum 1000 W|
|OS||Windows 10 Pro with October 2018 Update (ver. 1809)|
We used the following graphics cards for our testing, as well:
|Graphics card||Graphics driver||Boost clock speed (nominal)||Memory data rate (per pin)|
|Nvidia GeForce RTX 2080 Ti Founders Edition||GeForce
|1635 MHz||14 Gbps|
|Gigabyte GeForce RTX 2080 Gaming OC 8G||1815 MHz|
|Asus ROG Strix GeForce RTX 2070 O8G||1815 MHz|
Thanks to Intel, Corsair, Gigabyte, G.Skill, and Asus for helping to outfit our test rigs with some of the finest hardware available. Nvidia, Gigabyte, and Asus supplied the graphics cards we used for testing, as well. Have a gander at our fine Asus motherboard before it got buried beneath a pile of graphics cards and a CPU cooler:
And a look at our spiffy Gigabyte GeForce RTX 2080, seen in the background here:
And our Asus ROG Strix GeForce RTX 2070, which just landed in the TR labs:
With those formalities out of the way, let’s get to testing.
Battlefield V (1920×1080)
Our first and least-demanding round of testing suggests that while real-time ray tracing may be a reality in Nvidia’s new hybrid-rendering vision, it’s quite intensive for Turing graphics cards to accelerate. Moving from RTX off to BFV‘s low RTX preset halves average frame rates and doubles 99th-percentile frame times for all of the cards on the bench. 99th-percentile frame times remain well-controlled, though, and that fact squares well with the fact that BFV feels plenty smooth to play at this resolution on all of the graphics cards we tested, even if frame rates aren’t as fluid-feeling as they are with RTX off.
It’s only when we start to crank the RTX preset higher that the lesser pair of Turing cards on the bench begins to really show signs of strain. Let’s have a look at how long these cards had to spend working on the most demanding frames they had to render over the course of our test run using our advanced metrics.
These “time spent beyond X” graphs are meant to show “badness,” those instances where animation may be less than fluid—or at least less than perfect. The formulas behind these graphs add up the amount of time our graphics card spends beyond certain frame-time thresholds, each with an important implication for gaming smoothness. Recall that our graphics-card tests all consist of one-minute test runs and that 1000 ms equals one second to fully appreciate this data.
The 50-ms threshold is the most notable one, since it corresponds to a 20-FPS average. We figure if you’re not rendering any faster than 20 FPS, even for a moment, then the user is likely to perceive a slowdown. 33 ms correlates to 30 FPS, or a 30-Hz refresh rate. Go lower than that with vsync on, and you’re into the bad voodoo of quantization slowdowns. 16.7 ms correlates to 60 FPS, that golden mark that we’d like to achieve (or surpass) for each and every frame.
To best demonstrate the performance of these powerful graphics cards, it’s useful to look at our three strictest graphs. 8.3 ms corresponds to 120 FPS, the lower end of what we’d consider a high-refresh-rate monitor. We’ve recently begun including an even more demanding 6.94-ms mark that corresponds to the 144-Hz maximum rate typical of today’s high-refresh-rate gaming displays.
Happily, none of our graphics cards accumulate any time spent working longer than 50 ms on any frame, and even the RTX 2070 on the medium RTX preset only puts up a vanishing few milliseconds at the 33.3-ms mark. If we saw larger accumulations of time spent beyond those thresholds, we’d expect choppy gameplay that would badly tarnish any gains in fidelity from RTX reflections.
At the 16.7-ms mark, however, we can start to see some signs of strain. For the low RTX preset, the RTX 2080 Ti impresses with only a few unnoticeable milliseconds put up here, and the RTX 2080 puts up a similarly impressive performance. The RTX 2070 spends just four seconds of our test run working on tough frames that need longer than 16.7 ms to render, too—a plenty credible result for the most attainable Turing card so far. Even as we crank the RTX preset to high or ultra on the RTX 2080 Ti, the card has little trouble delivering a pleasant 60-FPS experience. The same can’t be said for the RTX 2080, though, and the RTX 2070 groans mightily even under the load of the medium preset.
For those looking for higher frame rates yet with RTX effects, the RTX 2080 Ti is the only game in town, but only at the low preset. The card spends a reasonable eight seconds of our one-minute test run working on tough frames that would drop the instantaneous frame rate below 90 FPS. Use any other Turing card or start cranking RTX presets, though, and an experience more in the neighborhood of 60 FPS should be as high as you set your sights.
Our first taste of Turing performance with RTX reflections enabled puts a steep price on the extra visual fidelity that architecture can provide through ray tracing. Let’s see if any of these cards can maintain a playable experience at 2560×1440.
Battlefield V (2560×1440)
As we up our resolution to 2560×1440, the same basic formula we just saw holds for performance with the lowest level of RTX reflections enabled: take the average frame rate with RTX off and divide by two, or take the 99th-percentile frame time and double it.
Depending on your personal threshold of playability, the 99th-percentile frame times our Turing trio puts up suggest smooth frame delivery remains the order of the day at 2560×1440 with RTX effects set to low. All three cards stay inside of the 33.3-ms mark that indicates frame rates are holding above 30 FPS, and the RTX 2080 and RTX 2080 Ti stay well within that margin. That figure is also worthy of note to Nvidia graphics-card owners because it’s the minimum frame rate at which G-Sync monitors can continue to do their thing with GeForces.
Crank the RTX effects preset to medium at this resolution, however, and only the RTX 2080 Ti maintains smooth sailing. This preset overwhelms the RTX 2080 and RTX 2070, dropping average frame rates to 30 FPS at best and spiking 99th-percentile frame times well past what we would consider playable. Using the high or ultra RTX presets pushes even the RTX 2080 Ti to the limit if we consider 33.3 ms a playable 99th-percentile frame time, and the RTX 2080 is well and truly out of gas. (We didn’t test the RTX 2070 at the highest RTX presets for that reason).
Our time-spent-beyond-50-ms chart captures a few milliseconds from one massive frame-time spike late in the RTX 2080 Ti’s run, but our other cards’ noses remain clean. Not quite so at the 33.3-ms mark, where the RTX 2070 spends eight seconds working on tough frames that drop instantaneous frame rates below 30 FPS when it’s asked to cope with the medium RTX preset. The RTX 2080 starts to lose its cool at the RTX medium, high, and ultra presets here, as well. Still, the low RTX preset lets all three of our cards at least stay off this important chart.
Flip over to the 16.7-ms mark, however, and all of our cards put at least some time up on the board with RTX on. The RTX 2080 Ti posts a hardly worrisome three seconds or so spent working on tough frames that would drop instantaneous frame rates below 60 FPS under the low preset, but upping the RTX quality to medium more or less doubles that figure. The RTX 2080 spends a quite-noticeable 11 seconds past the mark with RTX low, however, and the RTX 2070 spends another six seconds in trouble yet.
Cranking RTX presets higher imperils all of our cards to varying degrees if a consistent 60 FPS is the goal. The RTX 2080 Ti suffers a massive increase in time spent beyond 16.7 ms with the high or ultra presets in use, and the RTX 2080 at RTX medium spends half the period of our test run under 60 FPS. Cranking that preset higher on the RTX 2080 suggests there’s not much further left for that card to fall, while the RTX 2070 brings up the back of the pack even with the medium preset. To achieve even a consistent 60 FPS with RTX on at all at 2560×1440, get ready to empty your wallet.
Battlefield V (3840×2160)
Given the results on the preceding pages, it should come as no shock that 4K gaming with RTX on is an incredibly taxing workload. Going by 99th-percentile frame times, only the RTX 2080 Ti can stay on the right side of playability at all, and only then by a whisker with BFV‘s low RTX preset. The RTX 2080 and RTX 2070 are already on the wrong side of the 33.3-ms line with the low RTX preset, and cranking the RTX 2080 Ti to medium serves only to tank its performance.
Our time-spent-beyond-X graphs help to put only a slightly better face on the RTX 2080 Ti’s performance with the low RTX preset. The uber-Turing puts only a vanishing 30 ms on the board at the 33.3-ms mark, suggesting that gamers with “cinematic” tastes might be able to enjoy ray-traced Battlefield V vistas with a consistent 30 FPS. Even the RTX 2080 spends only two seconds, but really, that’s not good enough for a game that demands precise mouse movement like Battlefield V does. Meanwhile, the RTX 2070 nopes out at the 33.3-ms mark even with the low RTX preset at 4K, while the RTX 2080 Ti at medium puts all the time on the board we need to call its work a chug-fest. It seems 4K gaming with RTX effects is going to be the next hurdle to clear for Nvidia’s ray-tracing products of the future.
After staring at way too many puddles for way too many hours over the past couple of days, I can confidently say that Nvidia’s RTX reflections are often more lifelike compared to the base reflection technique DICE has implemented for Battlefield V. With RTX on, I saw far fewer jarring artifacts from reflections that gave away the fact that I was looking at a game. A few cases of distractingly noisy puddles aside, I got plenty of glimpses and glances of lifelike reflections that made me feel more immersed in the simulated French forests I snuck through for our tests.
Those sometimes-superior reflections come at an incredible performance cost, however. Even at their lowest settings tier, RTX reflections cut Battlefield V‘s average frame rates in half on every GeForce RTX card released so far at 1920×1080, although all three cards at least have some wiggle room to increase RTX presets at that resolution—plenty, in fact, in the case of the RTX 2080 and RTX 2080 Ti.
At least in our test area and at 2560×1440, the RTX 2070 hangs on to a tolerable 99th-percentile frame time at the low RTX preset, and the RTX 2080 stays well within the 33.3-ms mark critical to keeping G-Sync monitors happy. Turning up the RTX effects from there results in unplayable average frame rates and 99th-percentile frame times for our lesser RTX cards, leaving only the RTX 2080 Ti standing. 4K gaming with RTX on isn’t practical without an RTX 2080 Ti, and even then, that card isn’t offering anything more than bare-minimum performance for interactive gaming at the low RTX preset.
While the RTX 2070 will certainly let gamers dip their toes into what RTX can do, I felt the best balance of visual richness and reflections came from playing at 2560×1440 with the low RTX preset on the RTX 2080 or the medium preset on the RTX 2080 Ti. I honestly didn’t see a ton of on-screen differences from turning Battlefield V‘s RTX slider to high or ultra in our test area, but at 2560×1440, those settings murdered performance on the RTX 2080 and left the RTX 2080 Ti at the ragged edge of what even G-Sync can save. If you do want to turn those knobs to 11 for its own sake, the RTX 2080 Ti is unsurprisingly your best bet.
Despite the enormous hit to average frame rates, it’s commendable that RTX effects don’t result in horribly uneven or jerky frame delivery from Battlefield V. So long as gamers with GeForce RTX cards can tolerate 1920×1080 or 2560×1440 gaming and less-fluid frame rates overall, our results suggest they’ll still be able to enjoy consistently-paced frames that help maintain a sense of immersion.
On the whole, I don’t think gamers who already have high-end graphics cards with performance comparable to an RTX card should ditch their current pixel-pusher and go buy one of Nvidia’s latest on account of Battlefield V‘s reflections alone. We’re in the early days of this technology, and the effects I witnessed probably have room for improvement as developers get their coding fingers around real-time ray tracing and the hybrid rendering approach that Nvdia’s Turing cards require.
Still, if you were planning to upgrade to an RTX 2070, RTX 2080, or RTX 2080 Ti to begin with, you’ll be happy to have RTX reflections in your load-out when you fire up Battlefield V, and after this first taste of the tech, I’m eager to see what other studios can do with Nvidia’s hybrid-rendering tool kit for shadows, global illumination, ambient occlusion, and more. I’m hopeful it won’t be a long wait to find out.