Gamers Are Wrong About 1440p vs 1080p CPU Benchmarking

Added: 2026-05-05

[00:00:00]For years now, I have invested a good chunk of time and energy into trying to educate gamers about why CPU testing is done the way that it is. And I'm aware those of you who already understand are sick and tired of me talking about it.

[00:00:13]And believe me, I do get it. Uh we've already made countless videos explaining this topic, covering it from several different angles. But despite that, very little progress seems to been made.

[00:00:29]Now, for those of you thinking, I just need to get over this. It's just a vocal minority that doesn't understand the very basics of CPU testing. Most of you get it. So, Steve, just move on. Well, I'm here to disappoint you with a pretty shocking reality. You see, whenever I post CPU benchmarking related content, which has been quite a bit lately, the comment section ends up flooded with people demanding 1440p and even 4K testing. Some are just curious as to why those resolutions don't get included, while others claim the content's misleading or just dumb because it's not real world. But it's just the comment section, right? After all, less than a percent of all the people who watch our videos on YouTube leave a comment, making them by definition a vocal minority. And even then, we're talking about a vocal minority of a minority of viewers. Sadly though, as much as I'd like to believe that's the case, evidently it isn't. And my evidence for this is a poll that I recently ran on the Harbor and Box channel. The poll was simple. Do you want to see 1080p testing with medium and ultra presets as I'm currently doing, or would you rather see two resolutions tested, 1080p and 1440p, using a single preset? And shockingly, the majority of people, 56% voted for the inclusion of 1440p data, which I have to say is a bit disappointing and for a few reasons. Firstly, the 1440p data is useless. It tells you nothing that you can't learn from the 1080p data. There's really not much more to say about that. It's just a fact.

[00:02:06]There's nothing useful about 1440p CPU testing or just higher resolution testing for that matter. It's not real world. It doesn't tell you how your system might perform. It really provides you with no additional insight. It only serves to potentially mislead you. The poll results are also disappointing because the data we get from testing two presets is often very insightful.

[00:02:29]Testing games like Cyberpunk and Spider-Man 2 with and without ray tracing enabled has provided some really interesting results, especially as it pertains to CPU performance. The push for higher resolution CPU testing is also a bit baffling at a time when most gamers are now using some form of upscaling when gaming. Love it or hate it, that's just the reality. And to prove this, I again pled the same audience asking them if they use upscaling at 1440p. And 64% of you said that you do. And the reason for why this is relevant is because if you're gaming at 1440p with DLSS or FSR enabled and you're using the highest quality upscaling mode, which is quality, the base render resolution is actually below that of 1080p, you end up at 960p. And the base resolution falls to 835p with balanced and then just 720p using performance. Today's video is sponsored by the AOC Q27 G40 ZDF, a highly affordable, highquality QDLED gaming monitor. It's got a 27in panel, 1440p resolution, and 240 Hz refresh rate along with all the latest features like full G-Sync compatible support, lightning fast response times, and a picturein picture mode. With the beautiful QD OLED panel in the Q27 G40 ZDF, you're getting a brilliant true HDR experience, including deep blacks, per pixel control, and high peak brightness.

[00:03:56]AOC are offering all of this for a super low price of below $600 Aussie, making this a great option for your first OLED gaming monitor. To learn more, check the links in the description below. So, if you're gaming at 1440p and use any form of upscaling and also believe CPU benchmarking should be done at 1440p rather than 1080p because 1080p is not real world or whatever, you're actually gaming at a resolution that's below 1080p. And it's also worth noting that upscaling techniques work so well now that performance mode at 4K is acceptable to use and that has a base render resolution of 1080p. Anyway, I hear what many of you are probably saying. Why not just test 1440p with quality DLSS enabled. So, that's what we're going to do. Comparing that data to the native 1080p and then some native 1440p results as well. Of course, as usual for all of this testing, I'm using an RTX 5090, which is a separate issue that I'll tackle towards the end of the video. But for now, I want to look at this fixation gamers have on resolution when it comes to CPU testing. And for this, we're going to compare the Ryzen 7 9800 X3D, 5800 X3D, and 3800X. So, let's get into the data.

[00:05:12]Battlefield 6 using the overkill preset is a good place to start because at first glance, this proves the point those requesting high resolution testing think they're making. If we look at the 1080p data, which I normally provide, the 9800 X3D is still GPU limited. It's just 8% faster than the 5800X 3D, but a massive 106% faster than the 3800X.

[00:05:36]The 5800X 3D is also 90% faster than the 3800X. Again, a massive increase. But these margins alone only tell part of the story, and it's not even the story you should be interested in, especially if you think resolution has any kind of relevance here. As important as the margins are, the real sticking point is the frame rate. Battlefield 6 is a competitive online shooter. I'd personally consider it to be more of a casual shooter. Trigger alert, I guess, but the goal is to eliminate enemy players and win the match. So, by its very nature, it is competitive. The more casual players probably do play Battlefield 6 for the immersion and therefore might prioritize visuals over frame rate, meaning the 102 FPS we see from the 3800 X at 1440p would be deemed acceptable. But not all gamers are the same, and I'm certainly not saying there's a right or a wrong way to play a game like Battlefield 6. And there will be plenty of Battlefield 6 players who are playing to win the game. And a lot of those gamers will want every competitive advantage on their side. For this, those players will turn down the quality settings, which does two things, both of which are highly desirable for this type of player. Firstly, it increases the frame rate, reducing latency, which gives you a better shot of getting your shot off first. And secondly, it makes it easier to spot enemy players as it's now easier to see in or through foliage, explosions, and darker areas. So, circling back, the margins are just part of the story because ultimately there's a subset of gamers who desire a certain frame rate.

[00:07:12]Personally, I'm not playing a game like Battlefield 6 at less than 140 FPS, and ideally, I would want around 200 FPS in any competitive shooter. This isn't some kind of elitist take either. You don't need an extreme high-end GPU for this level of performance. You just need to turn down the quality settings, but you do need a CPU capable of throughputting those frame rates. And this ultimately is where the bottleneck lies for a lot of competitive shooters. The Ryzen 7 3800X is going to struggle to push much over 100 FPS regardless of the quality settings you use. And while we can reduce the CPU load a little bit, you're not going to maintain 140 FPS or better with this CPU. But looking at the 1440p overkill results in a vacuum, you'd probably be misled into believing that the 3800 X is just as fast as the 5800X or even the 9800X 3D. So, the point is the 1080p CPU limited results tell everybody, whether it be casual gamers, competitive gamers, doesn't really matter, but it tells all gamers exactly what they need to know. If you only require 60 fps, then yeah, sure, the 3800 X is fine. And even at 100 FPS, the 3800 X still gets the job done. But what about 120 FPS? At that point, the 3800 X is no longer sufficient, and you'll have to start looking at faster processors like the 5800 X 3D in this example. Now, for those of you gaming at 1440p with any form of upscaling enabled, you quickly jump back up towards 1080p levels of performance. For example, the 5800 X 3D is now 53% faster than the 3800X with DLSS quality enabled. Still, we are seeing a GPU bottleneck of 158 FPS. If this is the level of performance you require, then the 9800 X3D isn't going to offer much or really anything over the 5800 X3D. But then it didn't even at 1080p with the overkill preset.

[00:09:09]So nothing's really changed there. And that means the addition of 1440p testing with upscaling didn't really tell us anything new. And at native 1440p, we just learned that under this test configuration, the RTX 5090 will limit you to 103 FPS. And therefore, it's a safe bet that all other GPUs will be much slower than this. The 1440p data is completely useless because if you already know you're happy gaming at 60 fps, then the fact that the 9800 X3D can push over 200 FPS, it's not of interest.

[00:09:40]But if you want more than 120 fps, you have no idea which CPU can deliver this and which one does so while offering the most value. Then if we switch to the medium preset, everything changes without changing. By that I mean resolution scaling when compared to what we just saw with the overkill preset has changed, but the results, the margins and learnings from the previous 1080p testing hasn't changed at all.

[00:10:05]Regardless of the resolution, whether that be 1080p, 1440p, or 1440p upscaled, the 9800 X3D is still about 32% faster than the 5800 X 3D, while the 5800X 3D is about 47% faster than the 3800X.

[00:10:22]So, in this example, 1440p is no more realworld than 1080p. Not only that, but again, it's about the frame rate. If we limit performance to 110 fps via an introduced GPU bottleneck, that doesn't mean the 5800 X 3D and 9800 X3D are pointless and don't offer any real world gains over the 3800X. It just means you've decided that 110 FPS is an appropriate level of performance to introduce a hard performance cap, which would be pretty silly to do. But let's move on.

[00:10:54]Up next, we have Marvel Rivals. And here we see when looking at the original 1080p data with the ultra preset that the 9800 X3D is 23% faster than the 5800X 3D and 90% faster than the 3800X.

[00:11:07]And this configuration does introduce some GPU bottlenecking for the 9800 X3D.

[00:11:13]Now, if we switch to 1440p, but enable upscaling, this actually increases the margin out to 29%. It's important to note that DLSS isn't a completely free performance booster. There is a small amount of overhead, and that's what we're seeing here. Because the 5800X 3D and 3800X are already maxed out, it's possible that enabling DLSS can slightly reduce performance, though this won't always happen, allowing performance to remain much the same. But if you have loads of available CPU power, like in the case of the 9800 X3D, if the GPU can render more frames with DLSS enabled, it will. And we're seeing that here with a small 2 1/2% performance boost over native 1080p.

[00:11:53]Now, if we compare native 1080p to native 1440p, we see that the 9800X 3D runs into a strong GPU bottleneck at the 1440p resolution, whereas this wasn't the case for the 5800X 3D or 3800X.

[00:12:08]These 1440p results saw the 9800 X3 lead the 5800 X3D by a mere 6% margin. And it reduced the margin over the 3800X from 90% at 1080p down to 58% at 1440p.

[00:12:23]But now we're capped at 133 FPS. And this is as fast as the RTX 5090 can go at 1440p with the ultra quality settings regardless of what CPU you use. So, if you want more than 133 FPS on average, you'll have to decrease the resolution and/or quality settings. And as you can see here, enabling DLSS quality upscaling will get you up to 164 FPS on average. Of course, though, if you're happy with Sage's 60fps, then the CPU in the context of this testing, it doesn't matter. And even the 3800X could be deemed as overkill. But if you're willing to turn down the visual quality settings in order to boost frame rates, like what many Marvel Rivals players end up doing, we see some very big changes in our testing. Firstly, the RTX 5090 jumps up from an FPS cap of 133 FPS at native 1440p to 210 FPS, an almost 60% increase by downgrading from ultra to medium. And even at native 1440p, the 9800 X 3D is still 27% faster than the 5800 X and 88% faster than the 3800 X.

[00:13:32]But again, these margins are only part of the story. If you desire more than 140 FPS in Marvel Rivals, it doesn't really matter that the 9800 X 3D is 88% faster than the 3800 X. At that point, it might as well be 400% faster because the 3800 X simply isn't achieving your minimum desired target given it is limiting the average frame rate to just 112 FPS. As for the 9800 X3D, if you wish to talk about real world differences, at 1080p, it was 40% faster than the 5800 X3D, but at 1440p with quality upscaling enabled, that margin actually increases to 52%. And this is because the 9800 X3D does a better job of dealing with the DLSS overhead.

[00:14:17]Whether or not you need over 250 FPS on average in Marvel Rivals is really up to you to determine, but at least you can now make that determination while knowing the true performance of all of the CPUs tested.

[00:14:30]Next up, we have Bouldersgate 3. And this absolutely isn't a game that needs to be played at high frame rates.

[00:14:35]Anything that keeps the 1% lows over 60 fps is highly enjoyable in my opinion.

[00:14:40]Though of course, as always, opinions will vary. But at least in my opinion, this data is less about the performance in the here and now and more about using Boulders Gate 3 as a benchmark tool for gauging not just the raw gaming performance of each CPU, but how they might compare in future games. Using the ultra preset, we see pretty consistent scaling regardless of the resolution used. The 3800 X does become an extreme bottleneck, limiting the average frame rate to 79 FPS in our test with 1% lows of 55 fps. Though, as I said, this should be sufficient performance for most of you playing Boulders Gate 3.

[00:15:18]Even so, at native 1440p, the 5800 X3D was 71% faster and the 9800 X3D an insane 162% faster. These figures are only slightly increased at 1080p. And now the 5800 X3D is 85% faster and the 9800 X3D 186% faster. Though all three testing configurations maintained the same margins between the 5800 X3D and 9800 X3D. So in reality it wouldn't matter if we tested at 1080p or 1440p in this particular example. The margins would end up being about the same. And we see much the same story with the medium settings. The 9800 X3D was 50% faster than the 5800 X3D when testing at 1080p and then 43% faster at native 1440p or 48% faster at 1440p with upscaling. The gap to the 3800 X did grow even larger.

[00:16:11]And now at 1440p, the 9800 X 3D is 165% faster. Though again, you probably don't need 228 frames pers in a game like Boulders Gate 3. Next up, we have Cyberpunk 2077 Phantom Liberty. And these results are quite interesting.

[00:16:28]Firstly, the 9800 X 3D saw higher frame rates when using ray tracing at 1440p with upscaling when compared to 1080p.

[00:16:36]And this is due to the fact that these results are extremely GPU limited. And 1440p DLSS quality runs at a lower base resolution than 1080p.

[00:16:46]However, at native 1440p, we drop down to just 94 fps. And here the 9800X 3D is a mere 3% faster than the 5800 X3D, though it's still 24% faster than the 3800X.

[00:17:00]That said, with upscaling enabled, the margin increased to 88%. And here, the 9800 X3D was also 28% faster than the 5800 X3D. More notably though, that took the average frame rate from just 112 FPS up to a very nice high refresh rate experience with 143 FPS. It's also worth noting that while native 1440p does heavily restrict the performance of the 9800X 3D at sub 100fps levels of performance, if you enable DLSS, which most gamers apparently do, the margin ends up larger than the one shown in our original 1080p testing, growing from just 13% up to 28%.

[00:17:39]Now, if you don't care for ray tracing or would rather enjoy as many frames as possible, the medium quality preset still looks really good and performance goes through the roof. We also find a scenario where the 1080p and 1440p resolutions deliver the same results with the RTX 5090 as all three configurations are CPU limited, which is what you want to see when testing CPU gaming performance. So, the data here goes unchanged from our original testing. So, let's move on. Moving on to Space Marine 2. Here we have a very CPU demanding game. So, I wasn't shocked to find that including 1440p results provides very similar numbers to what we saw at 1080p, even with the ultra quality preset. And this is a great example of a game that really is primarily limited by the CPU, even at very high resolutions using maxed out quality settings. And that being the case, switching over to the less demanding medium preset shows more of the same. Again, whether we test at 1080p or 1440p does not matter here. But even so, why would you run the risk of unnecessarily introducing a GPU bottleneck into CPU testing by increasing the resolution? It just doesn't make sense to do so. Now, although ACC is a racing simulator, it also represents the kind of performance you'll see in unrelated games like Apex Legends and Fortnite, for example, being that it is a lightly threaded game that plays well at high frame rates. Even with the epic quality settings, we're comfortably looking at over 200 FPS with the 9800 X3D, making it at least 50% faster than the 5800 X3D. So, given that the Epic presets already revealed identical performance metrics at both 1440p and 1080p, it's entirely unsurprising that the subsequent testing using the medium quality preset yields the same universally CPU limited data.

[00:19:30]This clearly indicates that across a range of graphical settings, the bottleneck preventing higher frame rates is not the GPU, but rather the CPU. The system simply hitting a wall where the CPU cannot process and feed the necessary data to the GPU fast enough.

[00:19:45]Moving on to Spider-Man 2 using the ultimate ray tracing preset. Previously, we found when testing at 1080p that the 9800 X3D was an average 44% faster than the 5800 X3D, which is obviously a significant performance uplift. We also saw that the 5800 X3D was 29% faster than 3800 X, which is another significant performance uplift. Now, when we introduced the 1440p results, either at the native resolution or with upscaling enabled, the margin between the 5800X 3D and 3800X goes completely unchanged. as the data was already heavily CPU limited to begin with. And we know the RTX 5090 can render at least 104 FPS on average because we can see that with the 9800X3D.

[00:20:31]And this is really important to note because viewers often claim that testing lowerend or older CPUs with an RTX 5090 is unreasonable, unrealistic, or something along those lines. But this is a feelings-based argument that's devoid of logic. It doesn't matter if you use an RTX 5090 or a 5070 Ti. If the GPU can render over 70 FPS in this particular test configuration, you are going to end up 100% CPU limited every time, just as we see here. For the Ryzen 7 3800X configuration to be GPU limited, you would need to drop down to an RTX 5060 Ti, which is still a very good GPU, don't get me wrong, but going down this sort of path is missing the point anyway, because we're now talking about sub60fps PC gaming, which means I personally would always be looking to downgrade the quality settings for a performance boost because 60fps gaming kind of sucks and I would avoid it if I can. But even if we significantly reduce the visual quality settings in an effort to try and boost frame rates, the 3800 X struggles to push over 90 FPS for the 1% lows with an average frame rate of around 120 FPS, which is still very good performance overall. But again, if you want more performance, a part like the 5800X 3D is going to be over 50% faster, and that's while using settings that most would deem realworld.

[00:21:57]The second last game we're going to look at is Horizon Zero Dawn Remastered. And this title is heavy on both the GPU and CPU. As a result, it doesn't matter if we test it at 1080p or 1440p. The results end up being the same. And that's seen when testing with the very high quality preset, which is the highest quality preset in the game. So, given what we've seen with the very high quality preset, it comes as no surprise to find the 1080p and 1440p results are virtually identical with the medium quality preset.

[00:22:27]Finally, we have Crimson Desert, and this is yet another example where 1440p testing doesn't differ from 1080p testing. We learned that the 3800 X can't throughput more than 85 FPS on average using the cinematic preset. And this limitation obviously doesn't go away if you increase the resolution.

[00:22:45]Even when using the medium quality preset, the 3800 X doesn't give you much in the way of additional performance, up to a 5% increase. Though to be fair, we saw similar gains for the other two CPUs, strongly suggesting that all of this data is 100% CPU limited.

[00:23:01]Okay, so if we were to test at 1440p rather than 1080p, how would this have changed our findings and recommendations?

[00:23:09]Looking at these CPUs, not much would have changed. At 1080p, the 9800 X3D is 33% faster than the 5800 X3D. And at native 1440p, it's 27% faster. So, that really doesn't change anything. But had we tested at 1440p with quality upscaling using the settings that the majority of 1440p gamers apparently do, the 9800 X3D ends up 36% faster. The margin over the 3800 X does change a bit, but even so, the 5800 X 3D is still 44% faster at 1440p opposed to 62% faster at 1080p, which is a big difference. But as I've alluded to many times now, if you want more than 80 fps, it doesn't really matter if the 5800X 3D is 44 or 62% faster. It's much faster and it's possibly delivering the kind of performance that you're interested in.

[00:23:58]However, that was the maxed out quality preset data. If we switch to the medium preset in search of higher frame rates, the CPU becomes more of a key player.

[00:24:07]And now it doesn't matter at all if we test it at 1080p or 1440p. The data is still largely CPU limited. and therefore the margins don't change.

[00:24:17]So there you have it. Despite so many viewers requesting or even demanding 1440p CPU testing, it adds very little to the conversation while risking poisoning the well. I think the problem with a lot of these comments is that they're trying to slot all gamers into the exact same category. And that's just not how things are. Not all gamers play the same games, and not all gamers use the same settings or even hardware to play those games. And look, some will prioritize visuals over frame rate or vice versa. This is why showing the true gaming performance of any CPU makes the most sense as that data is not only relevant for those looking to maximize performance, but also those who aren't.

[00:24:58]It's what you'd call the full picture.

[00:25:00]At the risk of sounding like a broken record, testing CPUs under GPU limited scenarios doesn't actually tell you anything useful. And I get the feeling that some people want these GPU limited testing scenarios so they can maybe justify holding off on an upgrade. They can say something like, "Oh, look, the 9800 X 3D is no faster than a 5800 X at 4K when limited to 50 FPS by the GPU."

[00:25:25]But that's not the point of this testing. It's not meant to make you feel insecure about your current hardware configuration or even pressure you into upgrading. That is simply not the point of any of this content. If, for example, the 5800X can throughput over 300 FPS in a game like Rainbow Six Siege, and all you require is a frame rate north of 200 FPS, then it simply doesn't matter that the 9800X 3D can throughput over 500 FPS. You don't require that and your current hardware configuration is sufficient. So, kneecapping both processes with a GPU limited test configuration, it doesn't help anyone.

[00:26:01]showing the actual uncapped performance, at least to the best of our ability.

[00:26:04]It's not always possible to fully remove GPU limits, and some games just inherently are more GPU limited than CPU limited, but having that goal doesn't make the content misleading. It's not somehow elitist, and it's not meant to make you feel negative emotions towards your hardware. Not what it's about at all. Look, I've been a PC gamer for well over 20 years now. And I do have a tendency, a bias if you want to call it, towards maximizing FPS performance because I grew up playing games like Unreal Tournament, loved Quake 3, Rocket Arena, played all of those games, and I did so on quite low-end budgety hardware. Admittedly, it was possibly more affordable back then, but then I was working a part-time job at the age of like 15 or something, so I wasn't making a whole lot of money. And yeah, I enjoyed those games on a 16 megabyte Voodoo Banshee. Again, tuning the quality settings for maximum FPS and back then overclocking was quite effective, so did a fair bit of that.

[00:27:00]But yeah, whether it's a a Voodoo Banshee or an RTX 5090, the goal here has always been the same. And even if you are a single player gamer aiming for around say 60 fps, again, not looking down on anyone who plays games at 60 fps. I think the PC experience should be better than that. But if you're perfectly fine with it, that's fine.

[00:27:20]It's still important to know how products within a price range that you're looking at really compare because that allows you to make an informed purchase and buy the best hardware you can for your given price range. I mean, to me, that just seems like common sense and should just make sense to everyone. Now, I'm sure we'll see some comments, some people saying, "Look, Steve, if the results are the same at 1440p for most games, why not just test at 1440p and sort of in an effort to keep everybody happy?" And the answer to that one's simple. It's a flawed test method, and it can sometimes, and it will introduce GPU limitations that will mask the CPU's true performance, resulting in misleading data. The only reason we no longer tested a resolution like 720p is because we got to a point in time where games that we were testing with the hardware configurations that we were using, they didn't really see any change in the margins between 720p and 1080p under any test conditions or, you know, any of the games we were using. In fact, we saw some instances where the frame rates at 720p regressed. And I believe this was because some of these really high-end modern GPUs, they just didn't have enough pixel data to keep all their cores fed. And that caused inefficiencies within the GPU, which yeah, led to a bit of a performance regression. We're not talking anything huge here, but we'd see frame rates go back by a few frames on average. Look, it could just be a simple thing like a driver bug. Not exactly sure, but we put it down to just, yeah, the increasing complexity of the GPUs. But whatever the case, it was something that we were starting to see more and more of, which prompted us just to focus on 1080p. At the end of the day, one of the big issues with all of this is the fixation on resolution, which is misguided to begin with. The resolution really doesn't matter for CPU testing, as in the CPU doesn't care about the game resolution. The resolution only matters to the degree that it isn't hammering the GPU so hard to the point where it becomes the primary performance limiting component. claiming that you game at 1440p or 4K and therefore 1080p testing is useless or whatever, it simply demonstrates a complete lack of understanding of what it is you're actually researching. And we've done our best to educate gamers on this topic over the years. And with 56% of the audience still requesting higher resolution CPU testing, does seem like we've got a bit of work ahead of us. But hopefully this video did help because ultimately the goal of this video was to help. Um whether or not we succeed in that goal, yeah, based on um our history of making this content, I'm not hopeful. But if it helped a few of you better understand why it is you're seeing reviews from us and others commonly focusing on low resolution testing, then I'll be happy with that. If we just get a few of you each time, then, you know, time well spent, I guess. Anyway, that is really going to do it for this video.

[00:30:09]Um, there's a thousand different arguments you can make here and I could address all 1,000 of them and yeah, maybe in the future we will. Uh, maybe I'll do like a reply in the comments.

[00:30:18]So, feel free to throw your best arguments for why the 1440p or 4K testing for CPUs are somehow valid and I'll pick them apart in an upcoming video. Maybe that'll be a bit of fun for everyone. Anyway, that is going to do it for this one. Um, if you like the content we make at Harbox and you want more Harbox goodness, we have the join button or Patreon that will give you access to our exclusive Discord server, monthly live streams, behind the scenes content and Q&A stuff. So yeah, check that out if you're interested. If not, that's perfectly fine. And of course, I would like to thank you for watching this video. I'm your host, Steve, and I'll see you again next time.