You are here
Home > TECHNOLOGY > Ray Tracing & DLSS with the GeForce RTX 3080

Ray Tracing & DLSS with the GeForce RTX 3080

With Nvidia’s GeForce RTX 3080 out the door, it is time we take a deep dive into ray tracing and DLSS efficiency as soon as once more. We have already taken a complete look into how the cardboard performs usually and gave you a teaser of what to anticipate with RTX efficiency, however at present we’re going to point out you simply how fast the RTX 3080 is for these particular results in a wider vary of video games.

We’ll be inserting explicit give attention to evaluating the RTX 3080 in opposition to the RTX 2080 Ti simply because that was the quickest GPU you can get earlier than for ray tracing, however for good measure additionally, you will see outcomes for the RTX 2080 and 2070. As you’d count on, we plan to debate RTX on versus RTX off, which is the primary speaking level with regards to ray tracing specifically.

First, let’s undergo a few of Nvidia’s claims concerning the RTX 3080’s RTX capabilities and among the adjustments they’ve made to the Ampere structure versus Turing. The 2 huge ones come within the type of up to date, next-generation RT and Tensor cores…

On the RT aspect, Nvidia says that Ampere is utilizing a second-generation RT core, which packs three most important enhancements. The primary is double the ray-triangle intersection throughput, this isn’t the one factor that an RT core can do and is barely a part of the ray tracing image, nevertheless it’s the half that’s acquired a velocity up.

Not every little thing is quicker on this RT core versus Turing, however we do get a giant enchancment to ray-triangle intersection. Ampere additionally helps improved concurrent processing, and introduces help for {hardware} accelerated ray-traced movement blur.

Ampere additionally brings us as much as a third-generation Tensor core design, with Turing being the second-gen and Volta the first-gen. Whereas the Ampere SM options half the tensor cores per SM, every tensor core is quicker, so the general SM delivers the identical efficiency at minimal, or 2x the efficiency in some workloads in comparison with Turing, relying on the info kind. There’s additionally help for brand new information sorts with Ampere.

Whereas these are type of the low-level adjustments Nvidia has made to Ampere, there are questions marks as to how these enhancements will translate into real-world sport efficiency. In Nvidia’s launch presentation, they spent a variety of time evaluating the RTX 3080 to the RTX 2080 or RTX 2080 Super in ray traced workloads, displaying a 1.5 to 1.8x enchancment between these two components. However these good points aren’t simply coming from the 2nd-gen RT core and Third-gen Tensor core design, the RTX 3080 additionally merely options extra RT cores, so we’ll unpack at present whether or not these ray tracing enhancements are all the way down to a brute pressure method of including extra RT cores, or whether or not the RT cores themselves are quicker.

All our benchmark information was captured utilizing a GPU check rig comprising a Ryzen 9 3950X and 32GB of low latency DDR4-3200 reminiscence in a twin rank configuration.

Benchmarks

We’ll begin with the oldest ray tracing sport in our line-up, Metro Exodus. That is an attention-grabbing one as a result of the sport solely helps DLSS, not DLSS 2.0, so the picture high quality you acquire from the function is lackluster to place it flippantly. What we’re all in favour of right here is whether or not the RTX 3080 is a succesful card for native decision ray tracing, as a result of we will’t think about it would be best to play with DLSS enabled on this title.

There’s good things and unhealthy stuff to see in Metro. The great is that the RTX 3080 is quicker than the RTX 2080 Ti for ray tracing at 4K. We’re seeing a 30% efficiency enchancment, which makes the sport considerably playable at 4K utilizing the Extremely preset with ray tracing.

Playing cards just like the RTX 2080 and particularly the 2070 look very sluggish compared, definitely these GPUs are nowhere close to quick sufficient for 4K gaming in Metro with every little thing turned on.

Nonetheless, there’s nonetheless an unlimited efficiency distinction between ray tracing enabled and disabled. With the RTX 3080, there’s a 61 % efficiency drop with RT enabled, which is equal to the efficiency drop seen whenever you activate ray tracing with the RTX 2080 Ti. This implies that whereas the 3080 is quicker for ray tracing general, that’s down to easily being a quicker card; the RT cores aren’t any higher at accelerating ray tracing on this sport, because the efficiency delta between on and off is similar between Ampere and Turing.

At 1440p we begin to run into among the limitations of the Ryzen 9 3950X check system, however the takeaway is essentially the identical.

There’s nonetheless a big efficiency distinction between RT on and off, nonetheless the RTX 3080 is just quicker general. Whereas the 2080 Ti was extra of a 60 FPS card at 1440p with ray tracing, the 3080 begins to push up into greater refresh charges, which is a welcome enchancment. It additionally demolishes the RTX 2080 at 1440p with a 92% efficiency increase with RT enabled.

Shadow of the Tomb Raider consists of ray traced shadows, however sticks to an older implementation of DLSS, so these outcomes are solely proven right here for curiosity’s sake, DLSS is unhealthy right here, too, so it’s not value utilizing.

At native 4K utilizing the very best high quality settings, the RTX 3080 will get nearer to being a 4K 60 FPS card with ray tracing enabled. A 34% efficiency enchancment on the 2080 Ti is first rate, barely greater than the 30 % acquire we get on this title with out RT enabled. There’s little question the 3080 is far quicker than the 2080, we’re seeing 76% greater efficiency and that is key as a result of the 2080 was not highly effective sufficient for ray tracing at 4K.

There’s nonetheless a giant hole between RT on and off. The RTX 3080 sees a 41% drop to border fee with RT enabled, in comparison with 43% with the RTX 2080 Ti, so the 3080 is about 5% quicker at accelerating ray tracing in SoTR. That’s not a big enchancment although, so that you’ll nonetheless have to decide on between excessive body charges, or ray tracing on this sport.

At 1440p, the RTX 3080 is 35% quicker than the 2080 Ti with ray tracing enabled, in comparison with 28% quicker with RT disabled. Each playing cards see a ~40% drop to efficiency with ray tracing enabled. At this decision, it’s a alternative between maxing out a 144Hz monitor with the 3080, or enjoying with first rate however decrease body charges with ray tracing.

Management is the place issues begin to get rather more attention-grabbing, as a result of now DLSS 2.Zero turns into an element.

Simply taking a look at native 4K gaming with and with out ray tracing to start with, the RTX 3080 is 44% quicker than the RTX 2080 Ti utilizing the max high quality settings, which incorporates each ray traced reflections and ambient occlusion. It is a bigger margin than the earlier two video games, and with a 125% lead on the RTX 2080, the 3080 is a beast for 4K ray tracing despite the fact that the general body fee is a bit low. That is the primary card that makes RT playable on this sport at this decision.

The 3080 sees a 36% drop to efficiency when ray tracing is enabled, versus 39% on the Turing-based 2080 Ti. This implies the 3080 is about eight % higher at accelerating ray tracing on this title.

However, you’re much more more likely to be enjoying Management with DLSS 2.Zero enabled, as that gives close to equal picture high quality to native 4K. Right here all Nvidia playing cards examined are literally quicker with ray tracing and DLSS enabled, in comparison with operating at a local decision with out ray tracing, which is a perfect situation the place DLSS can be utilized to compensate for the added results.

Nonetheless, in our opinion that’s far too simplistic of a comparability. As you can too use DLSS with out ray tracing, the extra correct exploration is DLSS with out ray tracing versus DLSS with ray tracing, as a result of we don’t actually see why you’d run at a local decision on this sport. Most Nvidia GPU house owners might be selecting between the excessive body charges with DLSS, or the picture high quality provided with DLSS + ray tracing. When DLSS works this properly, it is a given you need to flip that on first, then tweak the opposite settings to hit the efficiency you might be after.

When taking a look at DLSS outcomes, the RTX 3080 is 40% quicker than the 2080 Ti with each ray tracing and DLSS enabled, and 85% quicker than the 2080. The 3080 sees a 34% hit to border fee whenever you allow ray tracing with DLSS already enabled, whereas the 2080 Ti sees a 37% hit. That is just like the RT on vs off numbers at native decision.

The RTX 3080 was 71% quicker when enabling DLSS with out ray tracing, and 75% with ray tracing. The RTX 2080 Ti was 73% quicker with out ray tracing, and 80% with ray tracing. That is utilizing the very best high quality DLSS possibility at this decision.

There’s not an excessive amount of else to say when operating Management at 1440p. The efficiency hit resulting from enabling ray tracing is rather less extreme at this decision at round 30%, and general efficiency is sort of spectacular with the RTX 3080.

The RTX 2080 was solely a 1440p 60 FPS card with ray tracing and DLSS enabled, whereas the RTX 3080 is a real 1440p excessive refresh possibility.

Subsequent up is Wolfenstein: Youngblood and what we now have listed here are two separate units of knowledge for the RTX 3080. The primary set is with the sport’s present code, and the opposite, labelled “permit async current”, is efficiency with an upcoming patch that’s designed to higher make the most of Ampere’s asynchronous compute capabilities (notably for ray tracing).

With out the enhancements to asynchronous compute, the RTX 3080 provides a 37% enchancment to efficiency over the RTX 2080 when each DLSS 2.Zero and ray traced reflections are enabled. However that will increase to a a lot bigger 54% when permit async current is enabled on the RTX 3080.

We additionally see over 2x the efficiency of the RTX 2080, taking respectable 4K 60 FPS efficiency proper up into the realms of excessive refresh fee 4K.

Whereas elevated asynchronous capabilities are good, there’s nonetheless that efficiency drop you may count on between operating the sport simply with DLSS, and with DLSS plus ray tracing.

The RTX 2080 Ti noticed a 36% drop to border fee, the RTX 3080 with out the brand new sport patch noticed a 32% drop, and with the brand new patch we noticed a 29% drop. Nonetheless, that makes the RTX 3080 24% higher at accelerating each ray tracing and DLSS than the RTX 2080 Ti, which is a wholesome acquire.

The good points aren’t as important at 1440p on this title, with the 2080 Ti and 3080 offering comparable acceleration capabilities.

General, the RTX 3080 is a big 41% quicker than the 2080 Ti with ray tracing, DLSS 2.0, and “permit async current” enabled, however, we’re not getting the 54% to acquire we noticed at 4K.

Fortnite is without doubt one of the newest titles to incorporate ray tracing help, and it really options fairly quite a few settings together with shadows, reflections, ambient occlusion and world illumination. This comes at a hefty value, particularly at 4K. Even with DLSS enabled, if you happen to activate all RTX results with an RTX 3080 GPU, you’re going through a discount from 135 FPS, down to only 42 FPS. The sport is unplayable with ray tracing and no DLSS at this decision.

In the event you determine to solely use reflections on the utmost setting – with reflections being probably the most noticeable impact within the sport – the efficiency hit is much less savage, dropping down to only 70 FPS.

Nonetheless, with the 3080, turning on reflections to most will halve your body fee, when you’ll see only a third of the body fee with all results enabled. The identical is seen with the RTX 2080 Ti, a minimum of with solely reflections enabled. A drop from 104 FPS with DLSS enabled, to 51 FPS with epic RT reflections, is a large hit. Nonetheless, the RTX 3080 is 37 % quicker than the 2080 Ti at ray tracing reflections on this title, in order that’s a strong outcome for this new era. The RTX 3080 can be 80 % forward of the 2080.

At 1440p, the distinction between RT on and off is much more extreme than it was at 4K. The RTX 3080 is sweet for 212 FPS with DLSS enabled and no ray tracing, however drops under 100 FPS with epic reflections, and 76 FPS with all ray tracing results.

For a quick paced sport, that’s an unlimited efficiency hit. Whereas the RTX 3080 is quicker than different RTX GPUs on this title, in our opinion the efficiency stage with ray tracing would not warrant turning it on over enjoying at tremendous excessive body charges by merely benefiting from DLSS 2.0.

We’re going to wrap up with a take a look at titles that simply embrace DLSS 2.0, so we will see how the RTX 3080 fares when ray tracing will not be within the image.

In Loss of life Stranding, the RTX 3080 receives a wholesome efficiency increase from DLSS 2.Zero at 4K, with 31% greater body charges. In the meantime, the 2080 Ti sees a 33% increase, and the 2080 will get a 38% increase. We all know that DLSS efficiency good points are higher at decrease resolutions primarily based on our earlier investigations, so this means that each Turing and Ampere have comparable DLSS acceleration capabilities and we’ve seen that from earlier charts, too.

What this implies in apply is that with DLSS enabled, the RTX 3080 is 27% quicker than the 2080 Ti in Loss of life Stranding, whereas with out DLSS, the 3080 is 29% quicker. You’re nonetheless getting that class-above efficiency with the 3080, native 4K efficiency is just like what the 2080 Ti can do at 4K with DLSS which is spectacular.

At 1440p we run into extra limitations with what DLSS can do, with the 3080 there’s only a 7% efficiency acquire, in comparison with 21% with the 2080 Ti. In the long run, the 3080 finally ends up simply 6% forward of the 2080 Ti. This can be associated to a attainable bug with Loss of life Stranding and DLSS with decrease resolutions, at 1080p efficiency is definitely worse with DLSS enabled, however solely with Ampere. It’s attainable this might be patched or improved.

F1 2020 was just lately up to date to incorporate DLSS 2.Zero help. At 4K utilizing the Extremely Top quality preset, the 3080 can profit from DLSS simply because it did on Loss of life Stranding, with a 20 % efficiency acquire. This permits it to be 34 % quicker than the 2080 Ti with DLSS enabled, though the acceleration good points are comparable for DLSS between Turing and Ampere.

At 1440p, each the RTX 2080 Ti and RTX 3080 ship a ~9% efficiency acquire when enabling DLSS. We’re seeing very excessive body charges right here, though it is clear DLSS sees diminishing returns the upper the body charges are. Not that I believe anybody will complain about getting 200 FPS at 1440p with this new Ampere GPU.

What We Discovered

The GeForce RTX 3080 is the quickest GPU available on the market for ray tracing, and the efficiency on supply brings correct ray tracing capabilities to a brand new tier of use circumstances (sure, the RTX 3090 is even quicker, however actually we do not see many players shopping for it over the 3080).

Nvidia has primarily achieved this by brute-forcing the design: relative to the RTX 2080, there’s extra SMs, far more CUDA cores, extra reminiscence bandwidth, and extra RT cores. This makes the RTX 3080 a lot quicker card than the RTX 2080 and RTX 2080 Ti general, and by extension, this implies the cardboard can be quicker in video games with ray tracing enabled.

This permits for 2 main upgrades in comparison with Nvidia’s earlier flagship GPU. The massive one is 4K gaming with ray tracing enabled. In a variety of titles, the RTX 2080 Ti was barely quick sufficient for 4K 30 FPS ray tracing, particularly in video games that didn’t help DLSS 2.0. No PC gamer desires to play at 30 FPS — not to mention these spending tons of {dollars} in a GPU. This has modified with the RTX 3080, which regularly approaches 60 FPS and positively delivers very playable efficiency in graphically beautiful video games like Management with DLSS and even Metro Exodus without DLSS.

The RTX 3080 can be a way more successful ray tracing possibility at 1440p. The RTX 2080 Ti was largely a 60 to 80 FPS card at 1440p with ray tracing in probably the most graphically intensive video games. The RTX 3080 pushes up into the realms of excessive refresh fee 1440p with ray tracing, and if you happen to coming from the RTX 2080 which struggled at this decision, the 3080 might be an evening and day improve.

We have been much less impressed by the acceleration offered by the 2nd-gen RT cores and Third-gen Tensor cores. There’s nonetheless a big efficiency hit when enabling ray tracing, with a 30 % drop in body fee being a reasonably typical outcome. So when you would possibly get say 65 FPS with ray tracing enabled, usually that very same sport will run at over 100 FPS without it.

With Ampere, there’s nonetheless that alternative between the visible results of ray tracing, or efficiency. In some video games like Management, ray tracing gives sufficient of a visible improve that you could be need to settle for the efficiency hit. In different video games, the improve will not be as important, so taking a 40-50 FPS hit may be exhausting to swallow. That’s no totally different from earlier generations, it’s simply that the 3080 is a quicker GPU general.

Our exams were additionally present that ray tracing acceleration within the RTX 3080 isn’t overly higher than the 2080 Ti’s, with lower than a 10% velocity as much as RT acceleration separating the 2. In a best-case situation like Wolfenstein: Youngblood at 4K which acquired an Ampere-specific patch to enhance efficiency, the RTX 3080 is 25% higher at ray tracing acceleration. Ideally, we would see a 50% and even 100% enchancment to RT acceleration earlier than the RTX on and off hole feels extra acceptable with at present’s results.

As a result of Ampere is barely a minor enchancment over Turing for acceleration, there’s nonetheless query marks over whether or not this card might be enough for the subsequent few years of ray tracing. Most of at present’s video games use just one or two results and use them sparingly. If video games begin going all out on ray tracing results – like what we see with Fortnite – the hole between RTX on and off will develop considerably. Extra acceleration with extra highly effective RT cores might be required to maintain up. In fact, that is all hypothesis concerning the future, so we’ll must see what occurs with the subsequent few years of video games.

As for DLSS 2.0, it will probably definitely present a large efficiency uplift on the similar visible high quality and in the very best of circumstances, can absolutely compensate for the losses when enabling ray tracing. We merely wished it was current in an increasing number of titles. DLSS 2.Zero would not appear to harness Third-gen Tensor cores to ship much more efficiency than earlier playing cards, a minimum of on present implementations.

With the launch of the RTX 3080, the dialogue around ray tracing could be very totally different than it was with the RTX 20 collection. Beforehand, we slew Nvidia for overpromising and beneath delivering contemplating ray tracing and DLSS have been two main options they used to promote first-gen RTX GPUs. Nvidia hoped the poor worth of 2080 versus the 1080 Ti can be overshadowed by how superb ray tracing is and the way it was a must-buy function. Two years later, there are nonetheless few video games that help ray tracing, even fewer with ray tracing at launch, and Nvidia needed to completely overhaul DLSS to make it helpful.

With the RTX 3080, we’re in a very totally different place. The cardboard provides a lot better worth than the earlier era, because of critical good points to common sports efficiency and decrease costs. This makes ray tracing extra of a bonus functionthe best way it ought to have been earlier than. It’s a lot simpler to miss the small number of video games with ray tracing when in all the opposite video games you might be getting 20 to 30% extra efficiency than a 2080 Ti for $500 much less.

From right here, we expect it’ll be clean crusing for ray tracing. The {hardware} is extra succesful, we’ll see extra video games with higher ray tracing results, new era consoles will help it, and we’ll additionally see competitors from AMD, and finally Intel GPUs on the PC.

Buying Shortcuts:
Top
Translate »