A couple of weeks in the past we did a horrible factor. A horrible, biased, and albeit unforgivable factor. At least that is what Steve was instructed by well-informed of us over on Reddit… we benchmarked the RTX 4070 Ti utilizing FSR, in 6 of the 50 video games we examined. That’s proper, we used FSR. Even although truly one sport enabled FSR by default, so we guess we are able to solely be immediately blamed for that horrible factor we did in 5 of the 50 video games examined, however that is no excuse.
Sarcasm apart, what we thought was an affordable and honest strategy to tackling a problem, kicked off a violent controversy that escalated to abused ranges over on Reddit with numerous customers claiming issues they merely have little or no information of. But the excellent news is, we are able to now set them straight, which we’re positive will not assist to keep away from this kind of factor sooner or later, however it should present the remainder of players with some helpful benchmark information.
A little bit of again story first. When we lately in contrast the GeForce RTX 4070 Ti and Radeon RX 7900 XT face to face in 50 video games, we determined to incorporate a variety of ray tracing-enabled outcomes, as video games utilizing the added results have continued to achieve floor and change into an essential facet of GPU efficiency. Despite that truth that each time we ballot your curiosity in ray tracing, the info usually comes again trying very bleak for the expertise – looks as if most of you continue to do not discover the visible improve important sufficient to warrant the large efficiency hit, however that is a separate matter.
In our check, there have been a number of video games like Cyberpunk 2077 which are literally a superb exhibiting for ray tracing, and on this particular instance we noticed body charges dip properly beneath 30 fps on the RTX 4070 Ti at 4K, however with upscaling we may stand up over 40 fps. Since should not anticipating anybody to sport at 17 fps, we felt the upscaling information could be extra helpful, even when it isn’t really exhibiting 4K efficiency.
For a greater illustration of picture high quality comparisons, take a look at the HUB video beneath:
The subsequent choice we confronted was: which upscaling expertise to make use of? The choices included testing GeForce RTX GPUs with DLSS and every part else with FSR, or to maintain it as apples-to-apples as attainable, so check every part utilizing FSR.
Given we’re attempting to precisely examine frames per second efficiency, we at all times wish to hold the testing as comparable as attainable. In truth, our essential concern was to not present AMD with a possible efficiency benefit. We know that DLSS is the superior upscaling expertise, usually offering higher picture high quality, and we additionally know that when it comes to FPS efficiency GeForce GPUs are likely to carry out a lot the identical utilizing DLSS and FSR.
In different phrases, we did not wish to run assessments on any sport the place FSR regarded noticeably worse than DLSS, whereas additionally offering better body charge efficiency to the Radeon GPU. So, to be able to keep away from all these potential points, we felt testing each GPUs with the identical open-source upscaling expertise was the most suitable choice.
After the RTX 4070 Ti vs Radeon 7900 XT shootout went reside, we noticed many feedback complaining about how biased the check was as we did not use DLSS, regardless of mentioning within the article that when it comes to picture high quality DLSS is superior to FSR, and the explanation that we used FSR for benchmarking to maintain the comparability as apples to apples as attainable.
On the HUB neighborhood, I made a neighborhood put up, addressing among the misconceptions across the subject and ran a ballot. Thankfully, the vast majority of viewers agreed with our choice, besides the outcomes had been pretty cut up. This neighborhood put up then led to a number of Twitter threads, primarily calling us out for utilizing FSR on GeForce GPUs. It was troubling to see low effort posts with no proof or any actual info getting upvoted, whereas these explaining either side of the story had been closely downvoted.
The major argument (or relatively, assumption) from those that opposed our alternative to make use of FSR believed that DLSS is one thing Nvidia’s accelerating utilizing their Tensor cores and subsequently DLSS can be quicker than FSR on GeForce RTX GPUs. Mind you, there isn’t any proof to help this past Nvidia’s personal advertising materials. A fast search on YouTube for stuff like “RTX 3080 FSR vs DLSS” or some other common GeForce GPU, and you will find loads of content material evaluating varied video games with the 2 upscaling strategies, exhibiting just about an identical body charges.
We have, in fact, completed our personal in-house testing. With no precise proof to help complainers’ claims, we did what we at all times do, and acquired benchmarking. So let’s go check out that…
Benchmarks
Here’s a have a look at DLSS high quality mode vs. FSR high quality mode in Dead Space. Please observe we have solely used the respective ‘high quality’ modes for all our testing. As you may see on this instance, there isn’t any body charge efficiency distinction between the 2 upscaling strategies, the RTX 4070 Ti delivered 141 fps utilizing both expertise at 1440p and 76 fps at 4K.
As for scaling, at 4K the Radeon 7900 XT noticed a forty five% improve when utilizing FSR, whereas the GeForce RTX 4070 Ti noticed a considerably extra substantial 73% improve.
It’s the identical story in Call of Duty: Modern Warfare II and this information might be simply verified by anybody with a Radeon and GeForce GPU as we’re utilizing the built-in benchmark.
The RTX 4070 Ti offered the identical outcome utilizing FSR and DLSS at each examined resolutions.
As for scaling, the 4070 Ti noticed a 38% improve when utilizing FSR at 1440p and the 7900 XT a 34% improve. Then at 4K each the 4070 Ti and 7900 XT noticed a 39% improve.
Next we’ve The Witcher 3, the place the RTX 4070 Ti was 2% to 4% quicker when utilizing FSR against DLSS. Not precisely a big distinction, however on this instance FSR was quicker than DLSS.
The Radeon 7900 XT noticed a 49% improve when utilizing FSR at 1440p and 74% at 4K, whereas the 4070 Ti noticed a 49% improve at 1440p after which 69% at 4K, so a barely bigger uplift for the Radeon GPU at 4K.
Dying Light 2 Stay Human was examined utilizing the top quality ray tracing mannequin as this permits DX12.
Here the RTX 4070 Ti noticed the very same stage of efficiency utilizing both FSR or DLSS. Scaling was additionally barely higher for the Radeon GPU at each examined resolutions.
Hogwarts Legacy additionally noticed the identical stage of efficiency for the 4070 Ti utilizing both FSR or DLSS with not more than a 2% efficiency discrepancy.
Hitman 3 additionally noticed no change in efficiency when utilizing both upscaling methodology with the RTX 4070 Ti, each delivered 71 fps at 1440p after which 38-39 fps at 4K.
The efficiency uplift when utilizing FSR was additionally comparable for each GPUs.
Warhammer 40,000: Darktide runs barely higher when utilizing DLSS, the RTX 4070 Ti was 4% quicker at 1440p and a pair of% quicker at 4K. No massive margins, however that is the primary sport the place DLSS is repeatedly a number of frames quicker than FSR.
We additionally discovered DLSS to be a number of frames quicker in Death Stranding, although on this case we’re speaking a couple of 1-3% distinction in efficiency, which we do not discover statistically important.
The F1 22 outcomes are stunning, right here FSR was 11% quicker than DLSS for the RTX 4070 Ti at 1440p and 14% quicker at 4K, and these outcomes had been recorded utilizing the built-in benchmark with the extremely top quality preset, and primarily based on a 3-run common.
In this instance, testing the 4070 Ti utilizing DLSS would place it at an obstacle on a benchmark graph. That stated, whereas we did check F1 22 with ray tracing enabled in our massive face to face benchmark, we did not use upscaling of any type because the body charges had been already fairly excessive, a minimum of at 1440p.
Forza Horizon 5 has a really typical set of benchmark outcomes for us. In brief, FSR and DLSS offered an identical body charges. Oddly although, FSR appears damaged on the Radeon 7900 XT on this title, as we noticed no efficiency uplift when utilizing it. Forza Horizon 5 has confirmed to be a really poor title for RDNA 3, which we suspect is expounded to a driver situation.
Moving on to Spider-Man Remastered, we’ve yet one more instance the place FSR and DLSS offered an identical outcomes at each examined resolutions.
Atomic Heart, like F1 22, did see the 4070 Ti performing higher with FSR against DLSS, on this sport we’re taking a look at a 4% uplift at 1440p and a extra substantial 8% improve at 4K.
The RTX 4070 Ti would have solely matched the 7900 XT if we had examined and in contrast utilizing DLSS, however in our graphs it was proven to be 7% quicker when utilizing the identical upscaling methodology.
Last up we’ve Cyberpunk 2077, which we examined twice: one with RT disabled after which with the RT preset. First, utilizing the usual extremely preset we discover no distinction between FSR and DLSS with the 4070 Ti, efficiency is principally an identical.
With RT enabled we discover the identical: FSR and DLSS end in the identical stage of efficiency for the 4070 Ti. So from a efficiency standpoint, it would not matter which upscaling methodology you employ, the body charges would be the similar, although we extremely advocate GeForce customers go for DLSS because it usually offers higher picture high quality.
What We Learned
Well, that was a variety of drama and false accusations over nothing now wasn’t it? Generally talking, FSR and DLSS ship the same efficiency increase for GeForce GPUs, although we did discover a number of video games the place FSR was notably quicker, which is essential to notice for benchmark functions. If we had examined the Radeon with FSR and the GeForce with DLSS and put that in the identical graph, the Nvidia GPU would have been at an obstacle.
Let’s be as clear as we are able to utilizing the English language: we’re in no way suggesting that FSR is best than DLSS in these video games the place body charges had been increased. On the opposite, the picture high quality of DLSS is nearly assured to be superior, subsequently GeForce homeowners would wish to use Nvidia’s proprietary upscaling expertise. But for the aim of a benchmark comparability, we wished to offer you apples to apples information, and we expect we have now confirmed that we did simply that.
Some should still wish to argue, “if I’m going to buy an RTX 4070 Ti, I’m going to use DLSS, so FSR benchmarks are completely pointless for me,” and we noticed loads of folks making that argument. But as different customers identified, overview retailers akin to ourselves are taking a look at relative efficiency, which is why we frequently check the best way we do.
If you assume we’re attempting to please everybody right here, we’re not, that is an inconceivable aim, relatively we’re attempting to do what’s proper. We know if we do not use any upscaling folks will nonetheless scream AMD shill, as a result of DLSS is best than FSR, and they’ll have satisfied themselves we’re attempting to faux DLSS would not exist, regardless of discussing it and testing over the previous months and years.
We’re damned if we use DLSS for GeForce and FSR for Radeon, as FSR typically would not look pretty much as good and will present AMD with a efficiency benefit, and as we have simply discovered. Meanwhile we’re damned if we attempt to make an apples to apples comparability by utilizing the identical upscaling approach for all GPUs, as we present in our latest 4070 Ti vs 7900 XT content material.
Ultimately, after cautious consideration, we have come to the conclusion that for giant head-to-head GPU benchmarks we can’t use upscaling in any respect, one thing many others additionally advised. Again, this does not remedy the difficulty, folks will nonetheless cry biased, as a result of their very own private biases imply you are both with them or in opposition to them in supporting the company they love (?), but it surely does remedy an apparent situation for us, we’re truly testing on the decision we declare to be testing.
Testing at 4K with DLSS or FSR is not truly testing at 4K, and if we’ll declare to be benchmarking at 1440p and 4K, we must always actually be sure we’re rendering at these resolutions. So there, downside solved. As for product opinions, like a day-one overview of the GeForce RTX 4070 Ti, we’ll nonetheless embrace DLSS benchmarks there, for the precise part of the overview that appears at upscaling.
It could be good if we may have higher discussions about this kind of stuff, simply because somebody has a unique opinion does not imply they’re immediately morally corrupt, they simply have a unique opinion, they may even be fallacious, however screaming shill is not going to assist remedy the issue, and it actually would not strengthen your place. It’s additionally good to have some supporting proof, as a result of replying on advertising materials from a beloved company is not an excellent technique both. Anyway, meals for thought, and that is the place we’ll finish this one.
Shopping Shortcuts:
- Nvidia GeForce RTX 4070 Ti on Amazon
- AMD Radeon RX 7900 XT on Amazon
- Nvidia GeForce RTX 4090 on Amazon
- AMD Radeon RX 7900 XTX on Amazon
- Nvidia GeForce RTX 4080 on Amazon
