Yesterday, we shared our thoughts regarding the 10GB of GDDR6X memory that the RTX3080 is equipped with. And, today, it’s time to benchmark this GPU in the 17 most demanding PC games.
As said, the games we’ve benchmarked are the following ones.
- Marvel’s Avengers
- Cyberpunk 2077
- Watch Dogs Legion
- Crysis Remastered
- Project CARS 3
- Horizon Zero Dawn
- Red Dead Redemption 2
- Quantum Break
- Borderlands 3
- CONTROL
- Kingdom Come Deliverance
- Ghost Recon Wildlands
- Anthem
- Metro Exodus
- Assassin’s Creed Valhalla
- Immortals Fenyx Rising
- Godfall
For our benchmarks, we used an Intel i9 9900K with 16GB of DDR4 at 3800Mhz. Naturally, we also used the latest version of Windows 10 64-bit, as well as the latest version of the GeForce drivers.
For both our 1440p and 4K benchmarks, we used the default Ultra settings. We also focused on pure rasterization (as we’ve already talked a lot about DLSS and Ray Tracing). We’ll be also comparing our RTX3080 with the RTX2080Ti. Do note, however, that we’ve slightly overclocked our RTX3080 (+90Mhz). Additionally, our RTX2080Ti has been overclocked (via the curve functionality of MSI Afterburner). Both of these GPUs were operating between 1.9Ghz and 2.0Ghz.
Back when NVIDIA released the RTX2080Ti, I described it as a great 1440p graphics card. And I can wholeheartedly say the same about the RTX3080. At 1440p, this GPU can run these extremely demanding PC games with more than 60fps.
At 4K/Ultra, things are not looking that great for the RTX3080. As we can clearly see, most of these games run below 60fps on the RTX3080 at 4K/Ultra. This could give you an idea of what you can expect from the next-gen only games that will release in 2021 or 2022. Also, and given these numbers, we can assume that the RTX3080 will require more power for gaming at 4K/Ultra in newer titles (and not more VRAM).
Now as we can see below, the performance improvement over the RTX2080Ti was not that great. What also surprised us was the very low improvements in minimum framerates (in some games). For instance, we only saw a 9-10% performance improvement in minimum framerates in Cyberpunk 2077. Similarly, Immortals Fenyx Rising saw a 10-15% performance improvement in minimum framerates. On the other hand, Godfall saw a 29-33% performance boost in minimum framerates.
The games that saw the biggest improvements in average framerates were Borderlands 3, CONTROL, Red Dead Redemption 2 and Horizon Zero Dawn.
Here are also all the percentage differences between the RTX3080 and the RTX2080Ti.
In conclusion, the NVIDIA GeForce RTX3080 appears to be an amazing 1440p graphics card. For 4K/Ultra gaming, though, you’ll either need DLSS or highly-optimized PC games. Owners of OC’ed RTX2080Ti GPUs should also not worry about the RTX3080, as the performance difference between them is not that big.
John is the founder and Editor in Chief at DSOGaming. He is a PC gaming fan and highly supports the modding and indie communities. Before creating DSOGaming, John worked on numerous gaming websites. While he is a die-hard PC gamer, his gaming roots can be found on consoles. John loved – and still does – the 16-bit consoles, and considers SNES to be one of the best consoles. Still, the PC platform won him over consoles. That was mainly due to 3DFX and its iconic dedicated 3D accelerator graphics card, Voodoo 2. John has also written a higher degree thesis on the “The Evolution of PC graphics cards.”
Contact: Email
Project Cars 3 is not demanding, its just horribly optimized.
Yup but John also says Dirt5 is badly optimized…i dunno but i can play this game in 4k all maxed out till launch. Seems that their gaming rig is not that well optimized. Grid 2019 as well. And this games are looking way way better that project cars 3.
Sadly almost any other racing game looks better than Project Cars 3, a shame I really enjoy the first one and the second one was ok.
The 2 last Codemaster games was indeed badly optimised at launch, that’s a fact.
Wow has Horizon gotten patched THAT well? That game used to run so bad that it shouldve been one of the worst performers but now its at the top of the charts? Interesting. Maybe I can reinstall it and play it finally
Yeap, it runs pretty good now after all those patches.
I have a 3080 gaming OC with a 8700K, 32GB 2666 mhz, Games on SSDs and a NVME and i cant confirm what you say. The 3080 is a crystal clear 4K card and you can achieve a minimum of 60Fps in every game you mentioned. You only need to slightly tweak the graphic options to achieve that on some. Why you feature games like Project Cars 3 and leave DIRT5 for example is unbelievable. Normally i really like this page but this article is nonsense.
Apparently, you did not pay attention to the article. We specifically talk about 4K/Ultra and not about what GPU can handle 4K/60fps gaming with lower graphics settings. If that was the case, some could also say that the RTX2070 is a 4K GPU as it can run games at 4K/60fps with Medium settings. Numbers are numbers and they don’t lie.
Why would you just crank everything blindly at ultra when those settings are where any optimization goes out the window ? Thats the setting where they just dump everything on there, nearly no visual difference from the next step. That doesnt mean the 3080 is not a 4k card. It runs literally every game on the market at 4k if you just slightly adjust the settings where needed. A lot of games dont even need adjusting.
The difference between the FE card compared to the 2080Ti is around 35%, going to 40 or a bit more if ocd or aib model. Not really sure how you got the conclussion that theres not much difference between them
And again, youre bencharking minimum framerates instead of 1%, jesus christ. The minimum framerate isnt indicative of anything. Nobody cares what the minimum is. You can have 80 frames 99% of the time and in a single second drop to 20. It doesnt matter, you care what the framerate is in 99% of the time while you play
Thank you, thats what i was trying to say?…the 3080 is a monster in 4k. Its the first card where you can enjoy 4k graphics maxed or nearly maxed out in all its glory.
Yes, agreed. This is a strange article, I must say.
We are seeing way, WAY lower performance improvements than 30-40%. It’s more around 20% (and in some games, it’s 15%). Our EVGA RTX2080Ti XC Gaming is also not one of the best AIB models (it’s nowhere close to a FTW3 model for instance). That doesn’t mean that the RTX3080 is not great. Especially for 1440p, it’s an amazing GPU. Here are all the percentage differences (again, both GPUs are running at 1.9-2.0Ghz with overclocks made via the MSI Afterburner OC scanner. They are on air and they are rock solid stable).
https://uploads.disquscdn.com/images/c7f57a37fe2d4b12bab4c3a5a584956f46ccd9c842c36d7d94635fb0b2092b2d.jpg
you managed to get the lowest differential between a 2080ti and a 3080 than the entirety of the gaming press. Any other review. Im not joking. The results you have for the 2080ti are abnormaly high. You got the 3080 at just 21% faster at 4k. Hardware unboxed for example got it at 32% over 14 games.
Many games that you bencharked dont have an ingame benchmark so you must’ve did manual benchamrking and more than likely situations couldnt be reproduced exactly between runs.
As the title says, these are the most demanding PC games right now. That’s how these GPUs perform in them. It’s not a review, and we had a similar article for the RTX2080Ti when we bought it (and we’ll have similar articles whenever we get newer GPUs). We are not selectively picking games that favour AMD or NVIDIA in order to affect (one way or the other) the results. Obviously some of them are unoptimized and that’s why they are so demanding. Quantum Break is a perfect example. CONTROL looks and runs better on a newer version of the Northlight Engine. HOWEVER, it’s interesting – at least for us – witnessing how newer GPUs perform in unoptimized games that could not run with 60fps/Ultra/1440p, even on previous gen high-end GPUs. Again, the whole purpose of the article is to see how the GPU performs in the most demanding PC games right now (obviously when it comes to GPU requirements. That’s why we haven’t included Microsoft Flight Simulator as it’s bottlenecked by the CPU).
regardless of the how big or small the performance is… it’s still a way cheaper card
Also should be comparing a 2080 to 3080, not a 2080ti. Wait for the 3080ti then compare again.
Do any of you really even use Ultra settings for demanding games?
I’ve only ever used a mix of low to high settings so I could keep an avg 144fps (I have 165hz/1440p monitor). If Nvidia, AMD and Reshade effects don’t reduce the fps too much then I’ll turn that in too.
Every single game I play I flip every single setting to Ultra and then scale down if I have too bit by bit
What avg fps do you play at and does routine change for you depending on resolution?
50fps is my lowest i’ll go. I play everything at 1440p. I enjoyed cyberpunk at 50ish fps but I had the entire game set to ultra. Worth it for me since i am a graphic wh0re
Never use low for anything. Ultra and then turn a couple settings down only if there is no noticeable reduction in quality. The image is paramount. As long as It’s not competitive fps the framerate isn’t as big a deal as it used to be since I got a 240 hz monitor. The input lag difference is massive since there are more instances for the game to refresh. 60 goes into 240 4 times. That’s twice as many chances as a 120hz panel which is f*king huge. Playing control with 60 hz doesn’t feel like sh*t like it did with my last 144 hz panel. Even cyberpunk is playable unless it drops below 45 hz.
A little late to the party. RTX 3080 benchmarks and reviews are old news at this point.
Still, nobody’s able to get a damn card so..
I got one in September, but the situation is still not good.
Got my 30 series card quite exactly 1 month after launch
3080? 3070? Or what?
Could you kindly tell me how much you paid for it and where did you buy it?
Europe? USA? Asia? or where?
3090 (asus strix oc), about 40-50$ over mrsp for that card. Scandinavia. Ordered at the release day. Figured that card will sell out and be out of stock for quite sometime so placed the order and waited.
Could you kindly tell me how much you paid for it and where did you buy it?
Europe? USA? Asia? or where?
Man that quantum break number sucks big time.
The game was pretty but simply requires too much rasterization. It’s a beast.
1440p is what I’ve rolled with since 2017, and it’s honestly becoming the new sweet spot for some, like 1080p was.
Also 1080p shows more from the CPU side vs GPU usage when you’re going 2-4k res.
I’ll continue seeing 4k as useless until we get cards that can handle well above 60fps at high settings, but at the same time, for most games to actually sport 4k res textures on practically anything, not just character/guns and the UI.
1440p really is the sweet spot. It only takes a small amount of fps going from 1080p and looks amazing. I will probably stick with 1440p for years
Another stupid set of benchmarks using “Ultra” settings, which is unrealistic, useless and kills performance for litte to no gains. Seriously, are you guys really PC gamers?
some advice maybe.. Rather than questioning credential that didnt prove anything (is there even any certification to be the real and only de facto PC gamers ?)
moron
Huang doesn’t give a flying about the 3080 VRAM. It was all part of the plan. He is anything but dumb and stupid. He is a genius at making money.
Want more? Buy a 3080Ti! Comments? No? Thanks for buying our new card! Or wait for our higher priced more VRAM, Super variants! – Huang
More Vram is useless, like you
I have been using RTX 3080 for 3 months now and I can say that it is an excellent graphics card indeed.
Lol…john sounds kind of salty because it took him so long to get a 3080 honestly. I have to disagree with him saying this a 1440p…nope. just no. While this is not the 4k/60fps on everything like some expected, it definitely reaches that mark far more than 2080ti, and with a few tweaks one can acheive the 60fps. Lol…he makes it seem like it’s only a tiny jump in performance improvement over the 2080ti…lol
Also, and given these numbers, we can assume that the RTX3080 will require more power for gaming at 4K/Ultra in newer titles (and not more VRAM).
*laughs in maxed Doom Eternal*
Going by the 1440p benches, you can clearly spot the games and respective companies who screwed the PC ports (CDPR included).
Surprised to see HZD gaining more after those plethora of patches though.
Remedy games will continue to suck on PC hw wise as per usual. Glad I skipped Control and QB.
Gotta agree, remedy sucks at optimizion.
Disagree with the cdpr assessment – The PC version is indeed heavy but considering the gfx fidelity it provides now its actually well done. As a bonus that means we can enjoy it with even better fidelity down the road when we get improved hardware.
Don’t mistake heavy for unoptimized… especially in a open world game! If the gfx is good now and runs quite ok… all above that is an icing on the cake as it will future proof the game… and once all bugs are squashed many will no doubt do a bunch of play throughs.
As for remedy – Agree with that assessment
Quantum Break is quite good. You should give it a shot. It has reconstruction that looks good so there’s no need for 4K anything. I’m going to play Control later today maxed out with RT (and DLSS of course).
Please add ultrawide performance (3440 x 1440) with ray tracing, DLSS, HDR and 10–bit color ?
They are testing a scene that is not THAT GPU-heavy (I know that scene, I’ve finished the game). The test chamber scene (at the beginning of the game) is more taxing and that’s the one we’ve tested.
Reasonably adjusting worthless settings that make no visible differences, but pull down performance in some ultra settings just need to be reduced when gaming at 4k.
At this point, if the RTX 3080 is a card for the 1440p … is my 2080 to be thrown away?
I have your same LG monitor …. 32 inch-G-Sync and 144hz (165 Overclocked)
Through the reviews I’ve seen so far, the 3080 is “almost” twice as good as my poor 2080 …… so what should I think? That my GPU is only good for 720p, for upcoming next-gen games?
NO, i don’t think so!
I’m used to playing above 100 fps in 90% of the games I have …. and my CPU is a “meager” 4790k @4800 Mhz …. paired with 32 Gigs of 2400mhz DDR3 Ram ….
You have simply chosen the heaviest and/or worst optimized games on the market!
And above all …. maxing out these heavy games doesn’t make any sense …. often lowering some details from Ultra to High will double the fps ….
Ok, I understand this is another matter, but you have really painted a bad picture for the 3080!
C’mon guys! It’s a monstrous card! it should not be labeled as a 1440p card! ….. it is suitable for 4k and beyond!
It’s a 4K card – what a dumb article
If I can’t run 4k at 90+ fps on high/ultra then it isn’t a 4k card.
So no it’s not a 4k card, It’s a 1440p card like the 2080Ti before it.
Sure you can piss fart around with settings on both 2080ti and 3080 to run 4k, But that’s not what I spend a few grand on.
What about 1080p?