Submitted 2 weeks ago by inclementimmigrant@lemmy.world to games@lemmy.world
The very bad news is, Nvidia says turning DLSS 5 on will currently reduce performance by 50-60%. And that’s across the board—no matter the game.
Looking at the basic performance charts Nvidia has provided us so far, the ultra-high frame rate figures are based on 6x Multi Frame Generation.
Let this sink in for a moment.
I want to make sure I understand this right. The 6X here means that for every one real frame generated by the GPU, the next six frames that follow will be AI generated?
I’m pretty sure its 1 real frame every X frames so 6x would be 1 real, 5 generated.
Isn’t the input lag going to be insane if the game runs at, say, 10 fps? This would have to come with some kind of new render pipeline where the simulation can somehow still run at a higher framerate.
Exactly (we almost posted at the same time, where I explicitly talked about this in the other reply, lol, also with 10 fps as example. :D). There is a reason why Nvidia does not share the latency. I have no idea how this is legal.
Yea, nevermind the fucking visual issues
Yeah, it will be the frame time of 10fps (100ms), plus however long it takes to generate the frames. Plus input, display and all the other latencies.
Note that at 10 FPS you will have a lot less information between frames, so it’s gonna look real bad. It only starts looking presentable at like 45-60 fps.
Are you saying that the 50-60% lower framerate is with with 6x framegen on?
Yes, but also it does not matter if framegen is on or off. Using DLSS 5 to change the faces is a 50 to 60% hit in performance. Lets say the 5090 gets about 60 fps in average (at 4k, Ultra Settings and RayTracing on) as an example. That didn’t look good to Nvidia, so they boosted those numbers with framegen on at 6x times, to get an average framerate display of 370 fps. Which also increases latency BTW, which Nvidia does not disclose here, as it would look bad.
Each those diagrams and numbers Nvidia put on screenshots, look down at the bottom it says Multi Frame Gen 6x. It’s small, most people would miss.
I don’t really get how frame generation works, but at what point are you extrapolating so much that it gets out of sync with the gameplay? Surely there’s a limit…
The game can pass movement vectors so they gen image can adapt.
Looking at the basic performance charts Nvidia has provided us so far, the ultra-high frame rate figures are based on 6x Multi Frame Generation. So, while the little RTX 5060 can churn out a 258 fps result in NBA 2K27 at 1080p Ultra settings with ray tracing and DLSS 5 enabled, the game underneath will actually be running at around 43 fps.
I can feel the molasses from here.
Just generate more fake frames with a second 5090!
Nvidia’s new SLAI technology!
Stop supporting Nvidia. If you don’t see the writing on the wall, you’re part of the problem.
Inb4 the flak from fanboys who think that a $5 trillion corporation is in their best interest: they certainly aren’t in your interests as a gamer, and they’re making money (in part) off of those who have forgotten what gaming is about.
Also inb4: yes, losing some frames to not have to deal with Nvidia’s bullshit is worth it.
Inb4^2: AMD is not a saint. But whataboutism has never fixed anything.
PhysX, G-Sync, HairFX, etc.
The writing was on the wall for a long time already.
I will never get how Nvidia got THAT popular amongst gamers. Like sure, Nvidia always had some benefits in certain usecases. But how many people were actually capturing game footage as a job and needed the best hw encoders? How many gamers also used their gaming rig also as a workstation for things where CUDA is essential? A lot! For sure. But not >90% market share a lot.
I honestly blame AMD for consistently missing every opportunity to actually be competitive in the space. So much so that they’ve spawned a phrase. “AMD never misses an opportunity to miss an opportunity”.
It’s not hard for Nvidia to capture the market when AMD doesn’t compete at all for anything other than the low end market except in a blue moon. Their name is synonymous with mid range at best cards. If you want high end you literally only have one choice and I say that as someone who owns a 7900 XTX. Even that card, which was designed to be competitive with the 4080, comes with a whole lot of compromises that make the card only competitive in pure raster performance, Linux compatibility, and literally nothing else.
You mentioned some of the technologies that Nvidia cards have. You’re right that not everyone needs every feature but being on the other end of that spectrum also sucks. Not having compatible hardware encoders means you take a performance hit every time you screen share your game in discord. Not having CUDA means your card is basically ONLY good for video games (yes I know rocm exists but trying to use it is a frustrating experience compared to CUDA). Having Ray tracing performance that’s so far behind the competition means you’ll always be making a sacrifice in either quality or performance. Choosing AMD means consistently watching Nvidia create new technologies and waiting for AMD to reinvent them a generation later to lackluster support.
I will never get how Nvidia got THAT popular amongst gamers
It was the only gfx card maker that doesn’t sucks, and the only real competitor is AMD and they kinda sucked. Games made back then are mostly optimised for nvidia. And then they have shadowplay which capture footage from x second/minute ago with a press of a button, which popularised sharing funny clips. AMD only get better like a few years ago, then Intel enter the race. Today Nvidia doesn’t really offer much advantage for most people because of the price. If chinese gfx card maker catch up, Nvidia will probably have to pivot to AI company to survive.
I believe HairFX is Unity’s version of AMD’s TressFX, so not an Nvidia tech in this case.
Exactly for those reasons the last gaming notebook I bought (used) was all AMD. Comes with some caveats but it’s better than supporting Nvidia.
They’ve not been for the gamers for a long time
haha you make it sound like any GPU will be affordable in the first place
i miss when GPUs just rendered the game, and that was all.
None of this upscaling, AI graphics altering bullshit that is there ultimately to distract from the impotent promises of 4k gaming, which is unnecessarily ridiculous to begin with.
While I do agree, upscaling is nice when your primary monitor is 4k, because playing 1080 on a 4k monitor looks like a pixel art game.
This is kinda weird since 4K monitors have no trouble drawing at 1080p and don’t look worse than 1080 monitors doing so.
And I mean you bought the 4k monitor, so for it not to be a good choice for gaming needs to be addressed. Can’t have that money go to waste. Not your fault you fell for the marketing again.
At least movies will look good on it.
4k is only ridiculous if you play games seated at a desk. Videogames are still ultimately a livingroom toy at their heart, and there are people who’ve graduated from game consoles to PCs and still want to play on the couch. That means distance from your TV, which means the TV has to be pretty large. Lower resolutions on a TV that size do not look good.
The relationship between screen size and distance to viewer decides at what resolution you won’t be able to notice more detail at a higher resolution. 4K on a desktop monitor can make the same amount of sense (or not) as on a TV (or phone, or any other device with a display really).
Which is bull, because we’ve been playing games from our couches for like 40 years and there were beautiful, amazing, wonderful graphics long before the 4k propaganda.
Ummmm 4k is way less important at couch distance compared to desk distance. Most of us don’t have TVs that occupy the same field of view from the couch as a monitor does at the desk. For example I have a 32 inch monitor at my desk 1.5 feet away and a 65 inch TV about 10 feet away. Sure the TV has 4x the viewing area but my viewing distance is more than 4x so the picture is not as crisp.
For fucks sakes, I don’t want “realistic” graphics, I want high and smooth FPS, they have been calling their games “realistic” for the past 20 years. Just make the game with a unique art style and optimise it.
Every single “realistic graphics” game looks shit after 5 years while Mario is over here still looking great decades after release.
No the games have to look like real life don’t you get it? Nevermind that fortnite and Minecraft, the two biggest games out there have cartoonish graphics and literal pixelated cubes no! We need to be able to see the reflection in a drop of sweat! And more AI!
Games should have “style” and “performance”? Bah. What poppycock I say, another billion to the AAAA games! For the investors! /s
I’m playing 10 year old games that go for realistic graphics, and they look just fine.
They don’t look like a film, they look like generated graphics. It’s not really a problem for immersion anyway.
Yeah, Same here honestly
Doing a shitty job of something at exorbitant cost? Sounds like "AI" to me!
total polar opposite of its original purpose.
Generating stock value?
that too, but I w as referring to enabling shittily coded games to run on midrange hardware
But it’ll keep the hype cycle going for a little bit longer! Line must go up!
Wait, what is even the point of using an upscaling technology if it makes your game run slower than without it? Why wouldn’t I just run native res at that point? Is their only point that “it looks better” in their minds?
You can make black people look white in your games.
Wonder if it could also make female protagonists look like men, might be useful for the sort of people who complain that a game is woke when there is a women
Don’t forget about making Cloud, a Japanese inspired character, look like Scarlett Johansson.
Well, gonna be real gay when that transgender protagonist is now in a romantic scene with a dude.
What’s the point of running 4k over 1080p?
Yes. It does make the game look better at the cost of framerate…
???
I think you’re missing the point of the question here. DLSS was meant to “fake” performance, i.e. you could have the game run at an upscaled 4k, when your PC can’t do native 4k. Now it’s reducing the performance at those upscaled 4k again, so you could just run native 4k at the same FPS.
Yes, their only point is that it looks better “in their minds”, as you say. This feature in particular isn’t meant to improve performance.
Personally I think the more subtle application of it can look pretty good, but it seems hard to control and does more harm than good. And that is even ignoring the massive performance impact that almost guarantees it to be unusable. Nvidia published some performance numbers that looked impressive, but that was using pretty agressive upscaling and 6X frame gen which is stupid because for frame gen to be decent you need a high base frame rate, which they didn’t get on anything other than the 5090 and arguably the 5080, or by pairing the GPU with a lower resolution that doesn’t really make sense for them (like you wouldn’t normally buy a 5070Ti for 1080p).
So it’s closer to an ML shader than actual supersampling?
As someone who works with 3D software everything about this trend annoys me to no end. I don‘t benefit from any of this. Instead hardware becomes unaffordable. And before anyone mentions AMD, last time I checked Nvidia still had more bang for your buck in Blender despite everything.
I pray my current hardware survives for a long time because it‘s looking grim out there.
If only a certain country and their president didn’t cause the prices of GPU and RAM to skyrocket, maybe it’d be more affordable. That’s not AIs fault, that’s capitalisms fault. Inb4 China comes out with RTX GPUs for a fraction of the price and leads to a democratization of AI technology. But “it’s unaffordable” is how all new tech feats always start, that’s how the shiniest new tech has always been. Were people complaining about this when Half Life 2 was being released with all its DX9 bells and whistles? No, because a brand new “512mb GPU” at that time was cutting edge and less affordable, but at least people were being paid enough to make that expense. Now that capitalism has destroyed our economy, people are blaming the prices and not the capitalism. It’s silly.
Upscaling was a mistake, I always turn it off and end up with higher or at least the same frame rates. It’s just another fucking thing like motion blur, chromatic aberration and lens flare. It makes the game look and feel worse at the expense of performance.
Upscaling was meant for older gpus. To lower the resolution in the game while making it fit the monitor resolution. I have no clue why flagship models need upscaling.
No, it was meant to offset the performance loss of raytracing in games. That is why it was introduced.
I always use DLSS Quality if available (and sometimes it is the only kind of render scaling available) and always have a ~20% performance boost with almost no image quality loss.
I’ve started downscaling
So, while the little RTX 5060 can churn out a 258 fps result in NBA 2K27 at 1080p Ultra settings with ray tracing and DLSS 5 enabled, the game underneath will actually be running at around 43 fps.
Lol. Nvidia with the toilet paper math. “You’re only getting 6 frames but it’s actually 11 billion frames.”
Like, I’m playing Bully. Render the locked 30, barely consume any power and shut up. “Run your computer at full tilt all the time so your games can run like shit!” I cannot stand any frame gen. Makes games look like static.
Like, I’m playing Bully. Render the locked 30, barely consume any power and shut up
I getting older and honestly only recently came to appreciate 60fps, it blows my mind to hear people say shit like “doesn’t do 120fps, unplayable.” …like wtf?
I think that “more fps” is important when there’s things that move quickly compared to their size. My go-to game here is Batman: Arkham City - in the steel mill sections, there’s a lot of “sparky” particle effects; in general, there’s a lot of dust flying about. That stuff does look noticeably better at 90 fps than it does at 60 fps. My eyes can’t really see the difference at 120 fps, but I can definitely hear the GPU fans, so locked to 90 is fine. Makes absolutely no difference to the gameplay at all, of course.
Games that are designed for 30 fps tend not to have the same kind of particle effects, mostly because they would look terrible. I do feel sorry for the graphics artists having their hands tied, rather than being able to fully express themselves, but all engineering is about weighing things up and finding the right compromise, of course.
Yassifying all of your games with a slop filter is the complete opposite of artistic expression, of course.
I always just play at locked 60 fps. It’s a simple life.
in some cases only 1% of pixels are actually rendered.
When will they realize that good graphics does not equal a good game?
Not anytime soon since gamers are still overwhelmingly indicating they will pay out the nose for the latest graphics.
Ehhhhh have you seen the numbers lately? People seem to be moving away from AAA titles more and more.
Yeah, life has infinite graphics and it’s a 3/10 at best.
That’s if you even consider AI generated frames to be “good graphics”.
I love beautiful graphics. I’m not picky about them and don’t insist on photorealism only (Valhala is a beautiful game and uses this strange mix of high and low fidelity graphics), but I do love me some good photorealistic rendering.
I’m not interested in this “AI tries to predict the pixels between actual rendered ones”. It’s just a push towards average, like everything else big data, when it works well. And maybe this one does work well but I still dgaf because I don’t want slop frames.
And I suspect that’s a reason you don’t see true frame to frame comparisons between AI upscaling and that same frame fully rendered at the higher resolution. It’s difficult for random reviewers to get the exact same frame rendered like that so I mostly see frames taken from the same viewpoint, but you can’t do apples to apples comparisons because the scene is dynamic and things other than the camera have changed. But nvidia could generate the same frame both ways for a direct comparison but only seem to want to show comparisons like DLSS off (at low res) vs DLSS on (at high res) or different DLSS versions against each other, which I think is telling.
So traumatised by AI ads on YouTube that I think this post is a tai chi ad.
Hey, wanna get rid of that stubborn belly fat? Try this chair tai chi.
I don’t care about or want this technology
If I had a video card that could run this I’d be so disappointed.
You know, I don’t particularly care about having the need to see every freaking pore on someone’s model. I hate graphic purists.
You know what AI reminds me of? That Xavier the renegade angel episode where the monks built that machine to do the monk chants for them and all but two monks remained afterwards and the two that were left in the monastery were playing video game and had put on weight. And the monks said soemthint liek surely engaging in western luxuries will not make us lazy
Good thing that this shit won’t run on AMD cards anyways lmao. AMD saving me once again.
Well its a good thing no one plays games anymore, its just benchmarking and dick measuring right? Right?
If chopping people’s arms off to make them run faster is a strategy then this is equivalent to chopping of someone’s legs to make them prettier.
Can’t wait for this vendor lock even more now.
So developers will create crap models, crap lighting, and then let DLSS 5 take care of the rest and now not only will AMD and Intel users get a subpar gaming experience, so will the bulk of Nvidia users who can’t afford the flagship card.
The future of gaming looks bright, like a train in a tunnel while your on the tracks.
Live service gaming is about to get so much worse when companies start obsessing about adding this.
From my understanding DLSS 5 can be targeted. Meaning the developer can choose to make a chair look more realistic while reducing the effects for others objects (like NPCs). I think the marketing group dropped the ball on the utility of DLSS 5 and went with the most dramatic NPC altering features. I’ve seen some DLSS5 lighting comparisons in Cyberpunk and I have to admit… It’s pretty damn good.
Not gonna lie, I’d try this in the same way I’ve tried AI for 5 minutes like a morbid curiosity thing. I’m also gonna admit that I can see it working with a very specific category of games, if it didn’t have a performance impact. Those games being the kind of games I have absolutely no interest in; sports games. I can see it for porn games too but I’ve not played more than a few and didn’t really see the appeal, but I can imagine it’d be a cool thing for some people.
None of that has any effect on this being a purely bad thing though, I’d rather stop using technology completely than being forced to use AI in any way.
Measure it when it’s out. Stop clucking and cackling about unlaid eggs.
The launch most likely will not be flawless. You will get your opportunity to some feathers when the damn chicken finally comes to roost.
I’m actually kinda interested in how this will change what emulated games look like. Can it improve the visuals in Aria of Sorrows? Or like old racing games?
I’ve been a fan of the most cutting edge videogame graphics since that Half Life 2 E3 demonstration in 2002. Didn’t hear anyone complaining about better graphics for less performance at that time. The ragdoll physics, the water reflections, the smart new NPC AI, got people hyped to buy the latest 256mb or 512mb GPUs that were expensive at the time. Know what I didn’t hear? “Ragdoll physics is destroying gaming. The real way would be to pay a village of animators to do keyframe animations of every possible way a character can fall over, instead they use this tool as an excuse to cut back on animators. This is an affront to devs artistic vision and a slap in the face to animators who are now out of potential work.” I didn’t see these complaints when Nvidia release DX9, DX10, or DX11, that all required the latest GPUs of the time and slowed down framerates on the “more affordable” GPUs. Except now that this has an association to AI, people are being prejudiced.
I think the only reason people are pissed at this fancy new graphics technology, which once upon a time would have us excited, is because the economy is so shit and nobody can afford the GPUs anymore (thanks Trumps economy.) That and the association of AI not being used on our general QoL but at our expense so that a few people can make a buck. You’d have to be a very special moron to say “AI is the reason the economy is bad” and say all this is the fault of AI. It seems that everybody is still looking past the dominant religion of “capitalism” to blame our current problems on. AI implies fancy new tech that everyone could have access to. Capitalism limits that access, and then weaponizes it against us. And yet the current bandwagon is the “AI Hate Train” meanwhile turning a blind eye to Trumps economy and capitalism. Makes sense tho considering most gamers are pro-Trump, right-leaning, economically illiterate incels.
Nah, keep giving Nvidia money
oh no way
all I have to do is tank my gpu performance, and I get a non-deterministic result, plus it looks substantially worse?
dang yo, where do I sign up
thingsiplay@lemmy.ml 2 weeks ago
I wouldn’t use it, even if it was adding performance. Nvidia really found a new technology that is even worse than Frame Generation.
jaycifer@lemmy.world 2 weeks ago
DLSS5 looks bad, but after using for a little while frame generation does have some use cases. It requires a high refresh monitor to be worth anything first and foremost to ensure there’s space for the interpolated frames between a high base frame rate. I use a 144hz monitor, so with 2x frame rate I still get a perfectly playable 72fps based latency.
I found it to be good while playing Kingdom Come Deliverance 2 because it let me turn up the other settings while consistently looking smooth. It’s a slow enough game that the increased latency isn’t noticeable.
It’s especially helpful for CPU bound games like Helldivers 2 and Victoria 3 because the interpolation happens after the main rendering. I could not get HD2 above ~90fps with my CPU, but with frame gen on the CPU only had to render 72fps. It also meant I could turn DLSS scaling off, which meant the generated frames were based on native renders and everything looked better. Victoria 3 went from a painful 20-30fps to a much more tolerable 4-60.
All that said it does have serious issues. The increased latency means I’ll never use it for a reaction time based shooter like Enlisted or Deadlock. A few games just don’t look good with it. Star Wars Jedi Survivor had poor performance anyway, but turning frame gen on both makes it feel like it plays worse while also adding terrible ghosting effects and artifacts.
I’ve been interpolating movies and tv shows since 2015, so I think I was better able to set my expectations for the downsides of interpolation.
jukibom@lemmy.world 2 weeks ago
While I don’t disagree that frame gen at high refresh rates does look pretty decent, I feel the need to correct the input latency part… In your 72fps example you actually have the latency of effectively 36fps because the GPU has to hold back a frame in order to have two frames to interpolate between. It doesn’t (currently?) do extrapolation so this holding back a frame is a hard limit and applies pretty much equally from 2x frame gen all the way up to 6x. This isn’t exactly catastrophic for single player or PVE multiplayer but I would steer clear in PVP
Nvidia reflex 2 should sorta kinda tackle this by taking the updated input => camera position and reprojecting the image before rendering but 1) this will require the game to support it (e.g. declaring UI, crosshairs etc to exclude) and 2) only serves to tackle the game feel portion of the problem - in pve you’ll still be working with old pixels and your response times will still suffer accordingly.
Again, great tech but worth knowing the limits!
sakuraba@lemmy.ml 2 weeks ago
why???
burido@lemmy.world 2 weeks ago
AI frame generation makes sense on lower hardware but to turn your games into slop? Why the hell? Why not enhance the 3d rendering or ray tracing with this ** crap.
TheHobbyist@lemmy.zip 2 weeks ago
I would actually argue it does not. Frame generation requires a minimum of 45-60 FPS so with such a requirement it does not take a game from unplayable to playable. Additionally, those frames don’t lead to the same experience as if the GPU had natively rendered the game because it does not account for the user input for those frames. It literally delays the next frame to insert a fake frame in between, leading to an increase in latency, so the only experience improvement is a smoothing of the rendering but it makes the visual quality worse and the latency worse. To me, this does not make sense and has been very wrongly advertised by nvidia with claims such as an 5070 = 4090, so so wrong.
takeda@lemmy.dbzer0.com 2 weeks ago
Wasn’t the original premise of DLSS to add performance and now they even falling on that?
I’m not as familiar with it, but from what I understand DLSS was switching rendering to a lower resolution then used algorithms to upscale the video to make it look as if it is high res.
thingsiplay@lemmy.ml 2 weeks ago
It is. According to Nvidia, DLSS is a family and brand now, consisting of different technologies. So the name DLSS started as upscaler only, but now it contains the image styling and other techs in the same name. Which is great, because it confuses everyone.
ms_lane@lemmy.world 2 weeks ago
The original, original premise of DLSS became DLAA.
It’s in the actual name - Deep Learning Super Sampling - it was supposed take your native res rednering, run it through than and get fake supersampled output.
Then they realised they could use it to run sub-native resolutions.