There is still one thing that I find very hard to understand here.. Why on earth does people find this topic SO VERY HARD to understand?! I mean it is extremely simple. A GPU is drawing a bunch of images, the CPU is then processing those same images. If the GPU is capable of drawing 100 images in 1 second, but the CPU is capable of processing only 50 of those images in 1 second, you will get an FPS output of 50 images in 1 second! If you raise the quality of the images that needs to be drawn high enough so that the GPU now is capable of drawing 30 images in 1 second, you have effectively moved you bottleneck / limiting factor away from the CPU and offloaded it on the GPU instead. NO MATTER THE DANM RESOLUTION USED! Come on people. Is it really necessary to have a million different UA-cam channels going over this very same topic over and over and over again, seriously?
Ultimately, I think it comes down to people not knowing what specific computer parts do. It's the same people that pop up in various hardware discussions and only mention their GPU when talking about the relative power of their pc. I saw quite a few recently due to the discourse surrounding the MH: Wilds pc requirements.
@ You might be right about that. Then I have another question: if someone is uncertain about what the individual parts actually do, then what compels them to ask a question that implies that a reviewer, who’s knowledge on the topic far exceeds their own, is doing his reviews or benchmarking wrong/that the way he does it does not reflect current setups that people have? 🤦🏼♀️ I am not saying that asking questions is a stupid thing in itself, my point is that if someone’s own knowledge is full of holes, the least that person can do is to ask their question properly, instead of making it seem like everyone else is in the wrong, especially the people that actually spends a hell of a lot of time on research in the first place.
That you have to explain this to people is mindboggling. I still remember the days when they were benchmarking Celeron's and Pentiums at 360p resolution in 1999.
Back in the day only nerds were into this stuff and you had to use your brain and read benchmark reviews, today PC gaming is very mainstream and people expect to be spoonfed information by techtubers, so can't or don't think anymore.
@@ooringe378 I think it's more that modern platforms like UA-cam make it easier for everyone to comment. In 1999 you basically had to write a letter to a newspaper if you wanted to complain about a cpu test not being in 480p. 🤣 I'm not saying it's bad, though. At least this way the people putting out the reviews can see what are some of the pain points of viewers.
@@ooringe378 ye. Now unfortunately explaining to some is useless because they really are brainless arrogant manchild. Instead of conceding they lack the tools and should educate themselves thw parrot the same points all over. It is just frustrating
in short: 1. find a CPU review that also consist of the game you want at lowest resolution possible. 2. check if a certain CPU able to push X FPS that you want 3. find a GPU review that also consist of the game you want at resolution you want to. 4. check if a certain GPU able to push X FPS that you want. 5. get them so you can play at X FPS. Doesn't require a bachelor degree to understand this.
@@hansyulian3671 Settings affect the load on the CPU. I haven't payed attention to it in a while but I remember they normally use the lowest resolution and the min settings. If that is the case then it isn't as straightforward as you mention.
I love that you have the patience to explain a whole video with another whole video, even if the questions are always the same every time someone tests a cpu.
@@AncientGameplays indeed mate. But what a pain. Hardware Unboxed's latest has Steve suggesting links in the description to previous videos explaining why. You might do the same with this one, might save you some time in the future. 🇦🇺
"Why are they testing the speed of that Ferrari against that Fiat on a straight road without a speed limit? No one is going to drive at 200kmh all the time in a straight road. They should test it on regular traffic with regular speed limits which is what most people will experience in reality."
People, don't look at this like resolution "X" or "Y". Look at it like this CPU gives "X" FPS in this specific game scenario :D The resolution barely affects the CPU (usually doesn't). So, a CPU that with those settings and in that game gives 100 FPS WITHOUT A GPU BOTTLENECK, will give the same 100FPS at 720P, 1440P, 4K, 8K or 8,000K. The Resolution doesn't matter at all.
I think that 1% and .1% lows are also important particularly in fast paced games. If you have a frame rate target of 144hz or 240hz, then the closer your lows are to your target the smoother the experience will be.
Hi Fabio! People just don't understand that within expected resolution they play games at chosen quality game settings. Meaning for example: i can play Starfield at 1080p with Medium quality settings OR play Starfield at 1440p at Low quality settings. In two examples it might be that i could have exact same FPS.
@@AncientGameplays I get it but some people need to see data to really understand. If you used a 7700xt or 7800xt GPU with 9800x3d, 7800x3d, 7700x, 7600x at 1080p they would see some games become GPU bottled necked and fps at 1080p would be similar proving whats happening.
Daniel Owen tried to explain this on his channel for several videos, too. I guess some people just seem to have a real hard time understanding the difference between "CPU test" and "CPU buying advice". Some people will watch a CPU review that's done in 1080p and go "that's bullshit I'm playing at 4K, how the hell does this review help me, why would I buy this?!" because they are looking for buying advice. They don't understand that watching a CPU review is just one component of understanding what you should or shouldn't buy. The process should be something like this: 1. Figure out which games you want to play at a) which FPS and b) at which resolution and graphics settings. 2. Watch CPU AND GPU reviews to determine a) which CPU is able to deliver that framerate in your chosen games overall, and b) which GPU can deliver that framerate in your chosen games at your chosen resolution. 3. Make note of what RAM, power supply, display etc. requirements have of course, too. Also consider what productivity requirements you may have for your PC, and how important those are to you, as the current fastest gaming CPUs are not the fastest in productivity. But it's key to understand that your - CPU determines if your PC can produce a given framerate AT ALL, with ANY resolution* - GPU determines if your PC can produce a given framerate at A PARTICULAR resolution or PARTICULAR settings, but only IF your CPU is fast enough to manage that framerate. This is why resolution doesn't matter FOR CPU TESTING. Or, more precisely, resolution should be LOW enough that the CPU can do everything it can without waiting for the GPU that's being used in testing. So IT IS BETTER FOR YOU, THE VIEWER, that CPU tests are being done at a lower resolution than you may be using. And so finally: A CPU test shouldn't be understood as complete buying advice. YOU, THE VIEWER need to make the decision whether a CPU is a good buy for you. And IT DEPENDS on your other components. Do YOU have a fast enough GPU to reach the framerate the tester is getting at the resolution you want to play at? If NO, then it won't help you to upgrade to this CPU (unless you also upgrade your GPU). If YES, then it will help you to upgrade to this CPU. So you NEED the other component of watching GPU reviews also - this is WHY GPU reviews have multiple resolutions tested, usually. And of course eventually you need to consider price and value, something only you yourself can decide upon. I don't know if I can explain it any better. I guess this is about 80% rant and 20% genuine advice but I still hope someone reads it and gets something useful out of it - or Fabio and Daniel's excellent videos. * Note: Graphics settings other than resolution can affect CPU load, too, so they can make a difference. See kognak6640's comment below
@@AncientGameplays Sure. I mostly meant that as an exception to the idea that ALL graphics settings yield the same result for the CPU. I edited the note.
@@AleksiJoensuu Any graphics setting decreasing polygons on screen will decrease also drawcalls per frame thus lowering workload for CPU(per frame). Draw distance, LOD settings, decreasing number of objects in scene etc are like this. It's the reason why I hate reviewers using low presets in CPU testing with assumption it only affects GPU side. Tells me they don't know how graphics rendering pipeline works. Lowered resolution with commonly used presets is the correct way test, otherwise you get skewed results. Resolution is not a CPU factor because it's internal for GPU. It comes in when true coordinate based 3D gets rasterized into 2D plane for the display at the end of the pipeline. Ray tracing stuff is different, I'm not sure how it works on CPU side but I'm guessing higher resolution requires higher precision and density of rays and therefore extra work for CPU too. So there could be workload scaling based on resolution on CPU side if RT in enabled.
Additions: Obviously you won't always know which games you want to play in the future, so you won't always have CPU and GPU tests available for your particular case. Then, you try to extrapolate by what tests you do have. If there's an upcoming game that's using, say, the CryEngine or Unreal Engine 5 or Frostbite engine or whatever, you can look up tests that feature other games that also use this engine, and maybe those will give you some approximations for how the future game might work. Or, you might look for a more general "this CPU and this GPU seem to match up pretty well in most cases". Or you might just go "I'll get the fastest I can from every category, who cares about the budget", or whatever. Anyway the point is, testers can't give exact buying advice for your particular case, because they don't know every variable in your life and goals. They don't know which GPU you're using, which motherboard socket you have, what games you want to play, which resolution do you want to play them at, and what FPS targets you have in mind. That's why it's done this way. That's why reviewers give you one part of the puzzle, and your job is to find the other parts and put them together for your goals and your budget. And with CPU reviews, the way to provide that puzzle piece for you, is to test in a way that lets the CPU run as unlimited by other factors as it possibly can. By using as monstrously fast a GPU, RAM and other components as the reviewer has access to, and then to test the games at a low resolution.
I appreciate you explaining this, some people don't seem to understand this concept when specifically testing CPU's it is very important to remove the gpu limited factor.
In a CPU review the focus isn't on the resolution or GPU used but to make sure that CPU is consistently the performance limiting factor. Its that simple. We need to know the total amount of FPS the CPU can generate without the GPU becoming the limiting factor.
I find it hard to believe that people seeing your review have not already heard of hw unboxed's channel. They did this same video nearly 2 years ago and they post it in every cpu review they do since then.
Guys, not the testing method is wrong, you just have another question in mind than "what's the fastest gaming cpu?". Your question is more "money limited": Is it worth it to spend 500 EUR on this cpu to upgrade your individual configuration. In the end its totally up to you to decide that. Without knowing your system, use case, targets and most important, budget, noone can even give you a hint.
For higher resolutions, I always look at 1% fps lows. The smaller the gap between 1% fps lows to the average fps the better, consistent frame-times and smoother experience. I for example have a 4k 120hz monitor, which means that I want a CPU that the 1% fps lows it can deliver as much as possible close to 120fps.
Boggles my mind that in 2024 we still have people who do not understand the way their computers work. The higher the resolution goes, the more the GPU becomes the bottleneck, not the other way around. In fact, the only times there should ever be a CPU bottleneck in games at 4K is open world games like Cyberpunk with a lot of pedestrian and vehicle traffic. Even then, the bottleneck will be the same regardless of resolution. If you drop to 20fps at 4K, you'll drop to 20fps at 1080p as well, because the CPU has become the bottleneck. It's why we know if there is a CPU bottleneck by dropping settings in game. If we see no improvement in FPS when dropping graphical settings, you have a CPU bottleneck.
@ChrisM541 If you are getting 20fps at 4K due to a CPU bottleneck then you absolutely will see 20fps at 1080p as well. How is this hard to understand? Lowering the resolution doesn't help the CPU. The pedestrians and vehicles and everything else it has to calculate are still there regardless of the resolution.
Yes yes, great video. My FPS target is always based off the 1% Lows in relation to my Refresh Rate. Then i set my FPS to a little above that and just play at the smoothest experience. I do the same on All resolutions.
Once I heard Tech from The channel Teach Deal explained things easily for people. It was something like this: your CPU is creating/generating/ rendering all the structures you see in games, all floors, walls, objects and characters, the CPU is giving all of them the structure and your GPU is just painting everything making it looks nice. So, if your GPU IS NOT UNDER 100% usage it means you are bottlenecked by something else in your system CPU/RAM/SSD/software, usually is your CPU that can't generate enough structures and your GPU is just waiting dor them to get "painted"...
And the opposite also happens. A lot of "CPU COMPARISON" using like a Ryzen 5500 vs 5800x3D in 4k RT ultra and a 3060. Claiming that there isn't any different between theses 2 CPUs. In both scenarios the poor 3060 is 100% and 30fps.
Hehe, at least in that scenario they are correct that the CPU isn't making any difference. But to then claim that there *IS* no difference between those CPUs together, well...
I appreciate your efforts in explaining every aspects regarding pc harware Fabio. The lower the resolution the higher the draw call from a gpu to cpu which becomes cpu/bottleneck and most people still dont understand that.
Thank you for the video, you actually made me realize something! While I always understood why people test at lower res for CPU (to avoid GPU bottlenecking), I always disregarded the results regarding practical implementation in my case: I still own a Ryzen 5 3600 to this day, coupled with a RX 6800, because I always gamed at 4K or at 1440p, therefore, seeing those CPU results always made me associate the FPS targeting scene as something that didn't made sense to me, as I don't play fast paced multiplayer games that require 200+ fps so you can gitgud. However, when I play AAA or AA demanding games in 4K, I ALWAYS upsample at least from quality (1440p) using either FSR, XESS or library hijacks for DLSS. So I actually do play in lower res., and in this scenario, maybe the 1080p and 1440p CPU benchmarks make sense to me. For instance, eTeknix published a 2024 revisit of my 3600 in those demanding games about 8 months ago, and the results baffled me. While the R5 7600 performed on par with the r5 3600 in 4K (making old me think upgrading was a bad idea), in 1440p the min. fps jumped from 40 to 70 in CP2077 and other games. So I might actually be leaving a lot of critical performance on the table given that my lows start to get into choppy low fps territory.
This discussion will never end because both sides have a point. "1080p is the most used resolution" is irrelevant because that's overrepresenting laptops, gaming cafes, and people with really old hardware. AKA people *_NOT_* buying this stuff or watching _YOUR_ channel. These are hardware _enthusiast_ channels. If that's your argument then you should test everything with 60 class GPUs since they're the most prevalent as well, right? The argument for 1080p/4090 is for a testing environment _only,_ and even then only shows you how it compares to other CPUs. It's that last part that the reviewers are so laser focused on that they don't see the problem. A lot of viewers *_WANT_* to know where that bottleneck kicks in with higher resolutions so they don't overspend on their CPU, or even worse spend a bunch to replace a CPU, motherboard, and memory that was still completely fine at the resolutions and settings they play at. Knowing a CPU's theoretical max is useful, but in the real world it's just that. _Theoretical._ It doesn't tell you how it will perform with _your_ GPU and resolution. Reviewers are so focused on that theoretical max because they already have the hardware to know where the bottlenecks occur so they can avoid them. But if you never show _where_ bottlenecks occur the consumer has no idea until they buy the hardware to find out they made the wrong purchase. Seeing reviews that only test tires on Ferraris and Lamborghinis tells me squat on how useful those tires are on my Toyota. Telling me with a voice of superiority "Well, obviously the max is 200mph right there in the test. Stop asking stupid questions." is completely useless information when you don't take into account the speed limits of a car most people can afford, let alone the real roads they drive on. Car sites don't do this. Boat sites don't do this. RV sites don't do this. Only PC hardware sites do. NO other hobby I'm into does testing with this mentality. Let alone tell your own viewers they're stupid for asking for real world scenarios. No one's asking for 1080p testing to be thrown away. We aren't stupid. We KNOW it's needed for a baseline to directly compare CPU to CPU. All we ask is for a random 1440p test thrown in with a 6750xt or 4070 to realize the massive waste of money a 9800X3D is if you can't afford a $2000 GPU and your 5600x is still more than adequate. Especially when GPUs have gotten so expensive that CPUs get more and more irrelevant by the day. It's not about accuracy of the scientific method. It's about providing a service to your viewer.
Couldn’t agree more with this! This is exactly what a lot of people have been thinking. It’s frustrating when reviews focus so much on theoretical maxes, but don’t show how things perform in real-world scenarios where most of us are gaming or working. Like you said, we don’t need 1080p testing to be thrown out, but it’d be great to see something that gives a better idea of where the bottlenecks actually start to happen when you’re at higher resolutions and using mid-range GPUs.
EXACTLY THIS! ZOMG I couldn't have word it any better, that's what the "1080P" and "720P" supporters DO NOT UNDERSTAND. People are not asking for low resolution tests to be removed, we actually understand the purpose. :D Knowing the info at 1080P doesn't translate into how the game performs at higher resolution as it depends on the GPU vs resolution vs game engine vs settings. If ever I make such a content (actually I plan on making one) I think I want to use a screenshot of your comment here. TIA!
This is exactly what I've been telling hardware unboxed to do but they are arrogant fks and refuse! 1080p is the majority is complete bs! The data is probably accounting for 3rd world countries in Asia where people get $1 wages!
@@Grievous- but... you already have that info, the video will just be titled GPU review. They already usually show you 1080p, 1440p, 4k data in the GPU review of your chosen GPU. You already know where the bottleneck happens, you're just watching the wrong video and reiterating GPU testing results in a cpu review says nothing you don't already know.
I experienced it myself how upgrading from Ryzen 5 3600 to Ryzen 7 5800x3D change the entire gaming experience. The max fps is same, almost never increase (I play on 2k resolution). But the frame time is smooth af. No stuttering, less input lag, and no flickering. Even after I bought 4K monitor, as long as my low fps is 40-60, it's still enjoyable. But still, for me who gaming on 4K, every CPU video test is useless, because it will never increase my fps if I upgrade my CPU again.
That's definitely not true. Just because games are almost always GPU bottlenecked at 4k doesn't mean they can't be CPU bottlenecked, which will happen with your CPU eventually if you continue upgrading your GPU. So look at it like this, whatever your fps is when you are CPU bottlenecked is the fastest the game will ever be able to run. If you have a game that can't run at your desired framerate at 720p or 1080p because it is too CPU intensive for your CPU to handle, then getting a faster GPU and playing at 4k will not get you any more fps, but upgrading your CPU will increase your framerate at all resolutions. I'm hoping I'm just misinterpreted your comment or taking it too literally.
@The_Chad_ What I mean is, my PC is already peak for 4K (I tried overclock, try any driver, cleaning/repasting and it still never increase my fps except using upscaling) so upgrading CPU is useless so any CPU video also useless for me. I need to upgrade my GPU and then probably upgrade CPU again because I already experienced how powerful CPU can impact gaming experience (not looking for higher fps but smoother experience if my Ryzen 7 5800x3D having stutter like my Ryzen 5 3600).
This kind of nonsense happens both ways. AMD fans now are saying the 60% power increase for 8% performance is a non issue while with intel 30% power for a 11% uplift was unacceptable to them... to me and my use case that added power draw on the 9800X3d and the 25 watts more idle power draw over the expected 5 yearsd lifespan will cost me 240 euro more in power usage. i calculate that into cost of running and compared to the 7800X3D it makes the effective price of the 9800X3d 640 euro (thats 700 dollar). It's crazy to me i have no bias just demands of products. If something is positive state it, if something is negative state it. No product past 5 years has bin perfect. I now run intel cpu +amd gpu and soon i will run amd+amd and later in time i will likely run amd + nvidia as amd threw in the towel on high end gpu.
And still 12 threads zen 1 shows better low and 0.1 than 4 cores i5, cost same and have future to upgrade lol. Me personally came from 1200 to 2600x, and now i got 5700x3d for 135$ from ali, while having same msi b350m 50$ motherboard from 2017. God bless Liza Su
@@auturgicflosculator2183 but why would i spend 150 more money then a 7800X3D from a budget bin when i have to enable eco mode to make it effient for zero gains over that 7800X3D. it just makes no sense. effectifly its just AMD pulling a intel with massivly more power draw for marginal performance gains in certain games. cyberpunk was 60% power for 8% performance and thats just silly. im here hoping for undervolt and curve optimazation that fixes some of this nonsense.
Thankyou so much for the great content and you're right getting a new GPU is the way to go in the short term. I am waiting for the 8800 XT early next year but still I am looking to also upgrade my 5800 X3D as well next year. Great vid.
Lmao I've been watching Daniel Owen try and educate people about this recently, sucks people need to make these videos on such a (to me at least) basic concept, but I'm glad we have people like you and Daniel willing and able to do so. Hopefully you guys are putting it in ways they can understand so they can make better informed purchases, I always hate to see when people buy the wrong hardware combo because of bad research.
According to Steam Hardware Survey for October 2024: 1920x1080 56.98% meaning more than half users actually use 1080p resolution 2560x1440 20.00% meaning every fifth person use 1440p resolution 3840x2160 3.68% yep, 4K users, right... I think above stats speak for itself.
@@asdfqwerty-u9nbro . When Ur blind making things bigger doesn't help You cure your blindness.. change the lenses :))) Sorry i wear glasses too but this was funny for me.
Stats don't mean shit for this situation. All the people at 1080p have lower end hardware, not a 4090 and high end cpu. Basically nobody playing at 1080p is gonna care about looking up benchmarks of high end hardware.
There are people who genuinely think that if a game causes a CPU bottleneck, then raising the resolution will reduce the bottleneck. No, Timmy, that's not how it works. You'll have exactly the same cpu bottleneck in all resolutions...
Entirely and completely untrue. 4k uses far more of the GPU power than 1080p, raising the resolution makes the cpu speed and ipc matter less the higher you go up.
@@drunkhusband6257 matter less comparatively, which means the cpu is still going to do the same workload, but at 4k chances are the limiting factor is the GPU, so a cpu can be much faster than another one and will produce the same FPS on 4k native res ultra settings. But that's just a stupid way to think about things, it only makes sense to think like this if your only usecase for the PC is to play games that demand way more from the gpu than the cpu and you have no intention of keeping the same system for multiple years as is.
@@ChrisM541 you'd be been lied to if they only tested in 1440p Ultra and 4k, the benchmark's purpose is to show the improvement of one cpu over another, not to show real performance on a GPU bound game. Did you even watch the video? also if you want to see "real use" scenarios there will shortly be no shortage of random smaller channels uploading every combination of cpu and gpu in different games and settings after every new release Stop trying to find fault in something that is not faulty at all
It's 2024 and people now don't know a thing about CPU performance taxing, or about their pc in general. Idk, I don't remember it being this bad quite a few years back, was it? Awesome video as always.
Took me 8 years to upgrade from a Phenom x4 to a Ryzen 3600 because the gains weren't substantial until then. I've been using my Ryzen for the last 5 years and haven't felt the need to upgrade since I play below 60fps at 1440p.
I have one of the new 240hz 4K OLEDs, I currently use a 5800x3D and a 4090. I play all my games at 4K but if I want to take advantage of my monitors high refresh rate I need to play with DLSS which shifts the bottleneck over from my GPU to my CPU. He is 100% right about FPS targets, if you ONLY play 4K native without any upscaling then your target FPS will be lower and your required CPU will be lower because even a 4090 at native 4K will have a hard time reaching high FPS.
Many people nowadays going with 1440p after buying new PC, but there is still a difference on those processors even on cards like 4070 Super or 4080. It will be even more visible in next gen with something like 5070 Ti or so.
You should do a video about only competitive games in comp settings with this kind of cpu(9800x3d) because lets be real most the people that are buy this high end focus in games cpus are gonna play in 1080p 1440p competitive games and is when this cpus really shine
Thumbs up! I hope people get it! For me it is clear but there are still so many out there who does not understand why a CPU is testet Benchmarked in 720p 1080p. Of course we need also Benchmarks in 1440p and 4K and we have to Bench AMD CPU with AMD GPU and the same goes for Nvidia. Intel is... Not really here anymore 😅
Intel really need to come up with their own version of x3D vcache, they are gettimg murdered on gaming purposes, their fastest cpu is still 14900k which eats up way more power, not to mention killing itself before microcode patches, we need competition to prevent stagnation.
Very interesting analysis Fabio, as I play mostly in 4k whenever possible, I will keep my 7800x3d, also because I have seen in the reviews that it uses considerably less power than the new 9800x3d
in gaming the power used is around the same, in render it uses more. I have that on my review video as well. At 4K you'll be GPU limited in most scenarios, so there's no point in getting a better CPU even if you're using DLSS because after all, you have a 7800X3D
Its 1:30 AM here, I should be sleeping but screw it, Ill start my 2 PC builds now while listening to your video 😂 the parts are sitting for a week now 😂
Even at 1440p you wil see a difference in the 1 % lows. Thats actually whats most important not the avarage fps. So people are just being lazy not including the 1440p benchmarks. Even at 4k with RT you see a big difference between cpu performance. LIke Dragon Age Veilguard. So only testing 1080p is outdated.
There's plenty of benchmarks that show the % difference of 1% lows between 1080p and 1440p between cpu comparisons is the same. So you're asking for redundant testing when you can get the same % diff, also GPU bottlenecks so you most likely would see the 1% lows diff be tighter and less pronounced.
@Mathster_live Its not the same for every game. Thats why having the 1440p is a good thing that gives you a more complete picture. It doesnt detract from anything it just adds to it. Its similar to testing gpus at 1080p and we still do it.
again, resolution doesn't matter. People aren't being lazy, you're being Dumb. Because if you're testing a game engine that is completely GPU driven, even the 1% lows will be equal. Also, you can test at 1440P but with a 7800XT instead of the 7900XTX and in that case, you're 100% GPU bottlenecked, so no difference in the 1% lows as well. You guys need to try and learn for once
@@dante19890 It is the same for MAJORITY of games, the exceptions are not the rule. Also, RT is a special case because it really depends on the implementation, and a better CPU does help and sometimes it doesn't and from what I know of Dragon age veilguard benchmarks, it's apparently something with the driver version affecting the performance with RT with 9000 series cpus
@AncientGameplays its the same argument for testing gpus at 1080p. We still do it even with top level cards even though it's being bottlenecked at that res. The same argument can be had for testing top cpus at 1440p so it is just laziness not to include it. It's something that also is highly requested by people.
Fabio, it's 2024 we should be doing 16K resolution tests. Please, we skipped 8K and are already at 16K. /s I still do hope you do tests with some older CPU/GPU combos like the Ryzen 3000s vs newer stuff to see if some things are worth upgrading vs price which would include mobo/ram. And keep up those great undervolting/ocing vids.
More than "not understanding", I think that what many want is for you to give them everything already done, to reproduce in your tests exactly the same scenario that they have in their setups and thus not have to make any effort of interpretation...
Fabio, I would even argue that 1080p is TOO HIGH for a CPU test. A CPU gaming test should have, ideally, the ultimate in potato graphics settings. I'm talking 720p or 900p AT MOST. Then there would be no question about whether or not the GPU is interfering.
@@AncientGameplays i have follow up question for 720p, i know its unrealistic, is there still games that being bottlenecked by 4090 in 1080p? if that happens do you usually give extra test in 720p if that happens?
720p is in fact deprecated, at least 1440p quality(960p) , cuz u dont use 1080p dlss quality, looks like shit almost always, if u want 1080p dlss quality is cuz ur GPU is the problem. REAL CASES: 4k dlss performance , 4k dlss balance, 1440p native, 1080p native,1440p quality(960p). 1440p quality will be for me the best test, and that is 960p( looks the same has 1080p native usually)
Most reviewers were testing cpus in 720p 2-3 years ago. Now almost everyone switched to 1080p. Probably due to upscaling being common now? Or the FPS at 720p was too high and some game engines were having artificial lock so it was affecting the results and providing misleading
Before 1080p, in GTX900 era and earlier, 1080p CPU testing was 720p CPU testing, even though NO ONE EVER played at this res on a modern machine (laptops aside). You pick a cpuFPS and gpuFPS and get the lower one and voila you have your rough resulting FPS.
@Ancient Gameplays: Maybe look at it in a different way. There are two reasons to watch 9800x3d reviews: - Theoretical/Coolness factor: People who want to see performance of new cool 9800x3d - Practical: People who want to see if it makes sense to buy 9800x3d for their setup, or maybe something cheaper will work as well Reviews at 1080p satisfy a lot of people. Your channel can attract more viewers with honey than vinegar. Instead of saying "It is pointless to review at 1440p" why not do 1440p review? => You will get views other channels don't, due to unique content. => You get more viewers happy, so more subscribers. => You can save people money by showing which cheaper CPU work as well as 9800x3d at 1440p. People love saving money. I just don't see a downside to doing special separate 1440p review of 9800x3d. There are videos that claim even with 4070 TI on 1440p there are benefits to using 9800x3d versus cheaper 9700x.
Your Witcher 3 Wild Hunt FPS is correct, I was playing it yesterday and I had Hairworks cranked all the way up. Frame gen adds another hundred to it. My hardware config of course.
correct me if I'm wrong. are you saying if you got a GPU and the best CPU and tested the GPU FPS on 1080p & 1440p the FPS will be the same so there's no point in doing the 1440p test.?
Yes, this is precisely what he mentioned; now I have to purchase a 1080p monitor to view more fps and then return to my 1440p monitor to observe the exact performance. 🤭 However, we are the foolish ones who fail to realize that if the CPU performs well at 1080p, the rest is insignificant if you gain only 5fps more at 1440p with the new CPU. 🤣 If the internet claims that 1080p is used by many, Steam represents everyone globally who uses a PC, and we must rely on that data, so can you understand what we’re facing here? If you are gaming at 1440p and move from a 7800X3D to a 9800X3D, you'll see a performance boost of 5-10%. Does spending 500$ make sense for that increase? No negativity! I simply can't resist commenting when I observe individuals deceiving others with a significant falsehood, claiming this is the greatest CPU of all time; I despise fakers and those who mislead others. How can I upgrade my CPU if I use 1440p? Should I check a video on 1080p tests?
@@DEM1GOD7 he used 1080p in testing to remove the GPU bottleneck and instead let the CPU do all the work to see how much fps it can push. If it can push 500 fps in one game at 1080p, it can do 500 fps on any other resolution. Now you need a powerful enough GPU to get to that 500 fps at higher resolutions. That's what he's explaining here. He was testing the CPU, not the GPU.
Exactly, if you have no GPU bottleneck, you could test the CPU at 480P, 720P, 1080P, 4K or even 8K. No GPu bottleneck? The CPU would deliver the same FPS technically
im with you on all this ancient lots of people skip parts of vids jsut to get to what they want to hear but not me im here for all of what i can learn and possibly in irl tell soemoen else who is need to know about all of this or i can always redirect them to this video as well. and resoultions do not matter if your cpu is bottlenecked and can not pus hup to 100 fps your not able to run higher res but if it can like you said then your perfectly fine jsut as long as the good ol trusty gpu can keep up on what the cpu is doing plus there is so much more to consider as well liek you have also said.
Great video Fabio. Higher resolutions have close to no impact on CPU performance. ''Close to no impact'' because as the resolution increases, the framebuffer size increases, and that requires more memory bandwidth for transferring data between the CPU and VRAM. That increased data transfer can ''technically'' lead to slightly higher CPU usage. However, that increase is generally not significant compared to the overall workload of the GPU, which handles the bulk of the rendering process. I feel like people who don't understand bottlenecks don't know that data is sent from the CPU to the GPU.
Seems this kind of clarification needs to be made at least once a year ... I've just bookmarked this and Daniels video now to send to others in the future since i got tired trying to explain this to people that rather get their "knowledge" from tictoc or reddit.
absolutely this is a well spread misinformation i agree with you when testing cpu res shouldn't matter the only thing matter is the fps limit the cpu can reach but since 1080p is the most used res its better to test at 1080p bravo and thank you for clearing this misinformation about cpu testing
@@shannonhubbard4026 Oh i do all the time haha. I'm tempted to pour a foundation for a home patio all the time. I just don't have a propperty big enough to make a patio as large as i would need.
Fabio, greetings from Latvia. How long do you think my R9 5950X will last, in games of course. Let me adjust the question, what is the most powerful video card that this processor can handle? Thank you.
@@AncientGameplays I bought an Asus VG32UQA1A (4k/160Hz) a month ago. Older games can reach 160Hz, but modern games are a complete failure. I have a Pure RX 7700XT in my computer right now. Will it make sense to replace it with a conditional RX 8800/8900 XT next spring? The current card has a performance of 35 teraflops, will it make sense to put a card with a performance of 50 teraflops to this CPU? Or Zen3 is already the past...?
Thanks aot for the work you do. It is sad that we hve a cult aslready of people who cry about using lower resolutions for testing CPU. But you really gave a nice explanation. IF my 12600k+6800XT cannot deliver the FPS i need i will change res and lower settings and i will find out the component that is bottleneking. I wish one day when you have time to test World of Warcraft. Alot of people play it and my 12600k+DDR4 4000 cannot deliver 60 FPS in raids or in the main city. No matther what resolution i use i cannot go over 60 if there are many people around. Still some say you do not need a better CPU at 4k. Same thing for ACC. I have a 2k 14 Hz monitor and i cannot see more then 120-130 FPS. thanks again.
Fabio amazing thumbnail lol. October 2024 steam survey, 1080p is at 57%, still the majority of the player base.1440 is only at 20%, and 4k is at 3.7% so completely irrelevant. The higher the resolution the more pixels need to be rendered and that is far easier for the GPU to do then CPU. The proper and definitive way to test 1080p is i believe at low settings as well so it offsets everything to the CPU and not the gpu. That's the way it's always been, people just dont wan't to understand that at 1440p+ the difference between CPUs won't be anywhere near as high as 1080p. This is also why the new 24h2 update for W11 makes a much bigger difference at 1080p, it's all about the frame calls.
@@monotoneone Correct, but not a single person in their sane mind uses ray tracing because it's pointless and a waste of time lol. I love that some games will tell you in the settings if it affects CPU/GPU or both. Every single game should be like that, it's just text doesn't affect the game performance at all :).
Nobody playing at 1080p owns high end hardware, and most aren't looking up benchmarks on you tube to begin with. 1080p for benchmarks should have died in 2015
It's not that people don't know what the aim of 1080p benchmarks is. It's that people who are thinking it's all about 1080p benchmarks (or benchmarks at non actual use-case resolutions) are missing the point that the CPU performance at real-world use-case resolutions is the actionable information that is relevant to people's interests when learning how a CPU performs. It's what purchasing decisions are, or should be, based on. Showing spreads at resolutions that are beneath what people will actually be using, when there will be no spread at the resolutions they actually use, is misleading in terms of messaging of what people should expect from the CPU. When the results of multiple CPUs show the same performance at 1440p because even the most powerful consumer GPU can't deliver higher FPS... that's good, actionable review information that lets people know they aren't going to get any more performance if they shell-out for a more expensive CPU. And very-few people even have an RTX 4090. Even in another few years, few people will have RTX 4090-level GPU performance in their system - which is something I go into further detail of in my first reply below. And people saying "then you're testing the GPU and not the CPU" aren't correct, as the GPU is only part of the equation. It's testing the CPU to the extent it's able to stretch its legs with (typically the maximum) GPU performance that's available, and letting people know that they aren't going to get more performance out of the CPU than what benchmarks at real-use resolutions show. Even in another several years, the large majority of 7800X3D and 9800X3D owners still won't have RTX 4090 GPU performance to go with their CPU. So, showing 1080p benchmarks without all this additional context to interpret the 1080p results is misleading and makes people think they're paying for performance they probably won't ever actually see. Reviewers need to think about the reason people are watching reviews. Showing FPS spreads at 1080p for a top-end gaming CPU is just academic information. It's not a practical information for consumers. By all means, show the 1080p spread. But don't do it without showing real-world use-case gaming results at 1440p and 4k. BTW, Linus addressed this topic in LTT's 9800X3D review, and noted that 1080p benchmarks aren't relevant for most people, and that looking at 1440p benchmarks shows there's not really any reason to upgrade to a 9800X3D if already owning a 7800X3D and only using the CPU for gaming at 1440p or higher. That fact is hidden in 1080p benchmarks, and so only showing 1080p benchmarks misleads people into paying more for something they won't benefit from.
still....why? if the 9800x3d can give me 200fps avg on 1080p and i know that a 4090 maxes out at 100fps avg in 4k, i now know, that the 9800x3d can support GPU´s twice as fast as a 4090 in 4k. if i only test this combination at 4k, i will NEVER know, if this cpu can handle a faster future GPU for example... now if you want to pair your gpu with a perfect cpu combination, just look at a GPU benchmark with your resolution, pick the GPU you want, look at cpu benchmark and ceck for the cpu that can deliver you that amount of fps and done
@@georgwarhead2801 But why would you only test in 4k? Likewise, why would you only test in 1080p? In typical use resolutions, you're unlikely to ever see the spread you do at 1080p in your lifetime use of a top-end gaming CPU. To get to see that spread at 4k, GPUs will first have to become so powerful that they're able to push the FPS at 4k that an RTX 4090 pushes today at 1080p. How many years away do you think it will be before there's even a GPU that can do that: maybe 5+? And then how many more years do you think it will be before even 25% of gamers have that level of GPU performance in their system, to make today's 1080p benchmarks relatable for a notable segment of the consumer base for that CPU? So, do people want hardware review information that informs them of what value a CPU gives them today, or of what value it could hypothetically give them in 5+ years, if they buy a top-end GPU at that time? When I buy new hardware, I want to know information that can firstly be applied today. And I'm not planning for 5+ years in the future, because I'll have bought a new CPU by that time, anyway. Reviews containing only academic FPS stats at below-typical-use resolutions is at least as bad, if not worse than if they only contained benchmarks of typical-use resolutions. Because the purpose of a review is to inform for a purpose (in this case consumer purchasing awareness), and the typical-use resolution data would probably be much more helpful to a lot more people. But reviews should contain normal-use FPS data, to go along with the academic 1080p data.
@@thischannel1071 its the same reason, why you dont test a 4090 paired with a12100f and than start argue that at 4k all gpus are the same, YOU WANT TO REMOVE OTHER FACTORS that can limit the performance. if you want to pick a cpu for a gpu, you look up benchmarks for your resolution, pick the gpu that fits your needs, look up cpu tests and pick one that fits in the fps range of your gpu ( BECAUSE THE RESOLUTION DOESNT MATTER) and BAM done, otherwise you have to look 100´s of possible configurations beeing tested and you come up with the enlightement, that this tests dont tell you how much faster this is then your current build.
@@georgwarhead2801 Saying you'd only test in 1080p to remove all other factors, to get academic results, doesn't answer a question of "why would you only test in 1080p" - it ignores the point that non-practical, academic data isn't only what a product review should have, or even what it should focus on, and the point that practical use-case is a fundamental part of a product review. Like I said in my previous comment, reviews should contain normal-use FPS data, to go along with the academic 1080p data. The resolution absolutely matters, though, because people use their PCs in certain resolutions. And there aren't that many resolutions that are used as standard: 1080p, 1440p, and 4k are the main ones. But if benchmarks are done in even just two resolutions, such as either 1440p and 4k, or 1080p and 1440p, then that creates a scaling trajectory, and people can generally mentally apply that scaling to figure-out where their CPU and GPU would sit along that trajectory.
I couldnt have said it better. I wish every reviewer saw this comment. For some reason people only think in black & white when the solution is SO simple: test the cpu in all 3 resolutions. Problem solved. People who want to see the raw improvement of the cpu over the competition and how "future proof" it is, can look at the 1080p results, and the people who want to see if it will be an upgrade for them specifically, right now, can look at 1440p/4K. Thats all there is to it. How is this so hard to grasp?
People think that 1080p is irrelevant when it comes to gaming and benchmark and they boast 1440p, 4k gaming is the new generation yet they still complain people will never be happy or satisfied no matter what, heck my old hardware R7 5700X is still doing fine at 1080p and I'm happy for it.
1080p still relevant. People want also to know for 1440p because a vast majority of player are on 1080p and 1440p. But guys. He's right most people still on 1080p and usually upscaling goes from 1080p to 1440p if you want the best quality. And can go even lower if you let the base settings.
@@zf8496 Resolution do not become obsolete. It's just the more resolution you have the greater the screen you can have without having more pixelated images. That's why I'm sticking to 1440p 27 inches
Hey ancient, you think you can do a vid on educating people about fake benchmarks also? It would be nice to know who to trust when it comes to scoping out an upgrade. UA-cam has a lot of channels with interesting technology combinations but many people say they’re fake. Would be nice to know who to trust when it comes to benchmarks.
haha, Intel fanboys fuming hard that their brand is underperforming hardware canucks had same thing, in short, 8 years ago Intel fanboys were all for 720p testing since it gave better numbers for intel, now it's other way around, so it's userbenchmark level of cope, don't mind those Fabio
AMD fanboys crying because they cant get hold of an 9800x3d 😂 Its not about being a fan, its about having competition for competative prices, and right now the consumer is pretty much screwed. We need intel to compete. Its not rocket science
@@prussell890we need Intel to compete yes. But consumers won't buy the shittier product just out of love for a corporation. Or will they ? See this îs why so many Intel fans are brainless. They have a better cpu for better price and better cooling and power usage but they are still choosing Intel because of confort and loyalty and hatreed in team Red. The story of producticity comes here . Really now . Producticity...majority of Intel i7 i9 users are just gaming. They don't produce shit. U dont need i9 to scroll webpages and use multitask. Almost no one will actually feel the difference in day to day use between i9 and 7800x3d. We need Intel to compete. But we need to punish Intel when they don't compete. So they will compete. My last processors were i5 7400-i7-7700k-i5-11400-now 5800x3d. AMD was punished everytime by the consumers on every single mistake they did with their producta. Just imagine If AMD pulled this 14900k 13900ks overvolt instability problems in one of their products. Like it happoened with bulldozers. AMD almost went bankrupt because of the anticonsumerism wave that came from Intel fanboys. And now they(Intel board) are knocking on lisa's door proposing deals and partnerships. Life is life. Intel fucked up a few years in a row.
@pR0ManiacS yes i totally agree with you. I dont think the latest intels are bad, they are more efficient and cooler, they just cannot compete with X3D. However in the uk, the ultra 7 is £350 and the 9800x3d is £450. Thats 25% more expensive. The 7800x3d has been discontinued, the 9800x3d is out of stock and the ultra 7's are available to buy. Paying that price really depends on how patient your are waiting, your type of gaming and gpu. Personally, I would obviously prefer the 9800x3d, but I would not be disappointed with the intel ultra range.
@@pR0ManiacS Unfortunately that is current trend of things to come. Especially with Intel and Nvidia fanbois 😂 with the latter being the worst offender.
I know it’s a bit off topic, but could you make a video with the best set up for amd adrenaline settings for video cards?(please 7900xtx included). With all the new features and functions, it’s a mess to optimize all for the best results. An updated overclocking setup would also be appreciated. Best regards, keep up the good work!
Most people who only use PC to game aren't understand how to do benchmarking and are also too ignorant to try to learn how it works. They just want a quick answer of how it perform with their hardware and their resolution target. I saw other youtubers explaining the same thing over and over about why people benchmark CPU on 1080p. I doubt they even read your replies.
A key point is the same game can be both GPU & CPU bottlenecked at different times. So there's weaknesses in even the target framerates as in some games other things matter too. Personally buying the 5800x3D was for reducing late game lag to avoidable situations like watching massive battles close up, where the GPU was hit hard in games like Stellaris. I was pleasantly surprised at the quality improvement in other games feeling smoother, despite the previous 5600x having faster ST benchmark scores.
This just speaks volumes in terms of peoples arrogance and ignorance, this doesn't just apply to benchmarking and pcs, this is happening in almost every subject.
Incredible that people don't know this and you've had to make a video to explain it. Good explanation video just like all your videos, keep up the good work!
Tho I agree that 1080p is the best resolution to test the CPU power in gaming, it wouldn't hurt to include 1440p and 4k benchmarks. I just hope techtubers do it like that. It's hard to find reliable benchmarks for 1440p and 4k that compares new and old Cpus
It would hurt very much, people have no notion of the amount of work I have to test just one CPU, especially since I already test at stock, Tweak RAM and OC
@ makes sense. I only watch the output and don't really know what's going on in the production part. I'm sure it's a lot of work. Keep it up, I'll be watchin
if they show tests in 2k/4k resolution, they will be forced to agree that the performance of current processors is more than enough for top-end videocards to work at their maximum, which means there is no point in upgrading the CPU. This will slow down sales. It's not profitable for them! Reviewers will give various reasons for this, but keep in mind that this is the main reason. Money first!) People look at the numbers and don’t fully understand how it works. Unfortunately, most casual players, looking at 1080p tests, think that they will get an increase in 4K resolution as well, when buying a new CPU. And to leave them in blissful ignorance, 2k/4k tests are not done) This gives a boost in sales at the start. Nothing personal, just business😉
Love 1080p and always will. For me personally its the perfect compromise between efficiency, visual pixel clarity and smoothness. Most of the time im running AFMF2 at 90FPS (180fps with AFMF2) at 1080p. My system is super quiet at 30% system usage and i can game the entire day without worrying about my power bill.
But, imagine rendering at 1080p and then upscaling to 1440p, with FSR it beats anti aliasing in some games, and with dlss likely all of them. For it's also really good for lightweight competitive games where someone's head can be a few pixels where as on 1440p they're way easier to see. 1080p is awesome but don't sleep on 1440p because it's "more premium"
It's funny how many UA-camrs have to explain this over and over. But people still make those uneducated comments. To me, this is common sense. And simply wanting something to be true doesn't make it so.
Next thing this new Cult will argue it is "Why reviewers aretesting these games in very CPU intensiv areas??!!!". and they will be crybabies because they only play 1-2 minutes in that area, and is not relevant, etc. Th problem wll not stop with them, no matter how much explanation you will have. But yea, congratulation to you, having this much patience.
In theory it also depends on how efficient the drivers use the particular CPU. Lets say the driver is very inefficient with the CPU, then the FPS drops. That does however not mean the CPU is weak, it just gets misused. To be precise we should also monitor how much cpu time the driver software consumes, compared to some benchmark software. Pretty hard to do as the OS also plays a role in this.
Most people who watch these videos completely understand the need for testing at 1080p, but i think what most people also want is the higher resolutions to actually show how it changes from CPU bottleneck to GPU bottleneck. I personally would like to see how my 7800X3D holds up against the 9800X3D at 1440p to see only a 1-2 fps difference. It would be nice to see the proof that there is zero reason to upgrade. And it would be even better to see the comparisons using DLSS and FSR because the theory would say uspscaling from 1080p to 1440p should show the same (or similar) results as a native 1080p comparison... but practically speaking.... will it really show those big difference as it does in native 1080p?. these are the things we would like to see. So please, do... 1080p native 1440p native 4K native 1440p quality/balanced upscaled 4K balanced/performance upscaled
Exactly... I can't believe that literally none of these guys are doing that. It's literally the exact same content on every single channel. 🙄 Yet, comment sections are FULL of people wanting to see 1440p tests and so on.
You can check 1440p/2160p tests like that on Techpowerup. Yes, once you become GPU bottlenecked the differences disappear and it is no longer a CPU test. If you have a 7800X3D you really shouldn't upgrade unless you play competitive games and you really want to max out your 240-480FPS screen... at 1080p... I just went from an i7 2600k to a R7-7700 not because I can't afford a 7800/9800x3d but because I'm not going to pay 2.5x the money to get +30FPS at 1080p and +10FPS in 1440p once I get my 5070/5080... I'll get a 10800/11800x3d once they have 16 core variants on a single CCD
Soon UA-cam will be full of those kinds of benchmarks so no worries man, then you will see that you are correct sir, so if you play at 1440p/2160p stick with your 7800X3D and max your GPU if you want more frames, not your CPU. If you have money to spend then wait for the 5080/5090 and don't waste here. Even if you play at 1080p who cares? from 500 frames you will fall to 400-450? damn that's a shame, the game will be unplayable for sure for those who had a 144-175Hz monitor. Other point is and I'm sure you know it, almost all of the test you watch on YT, doesn't matter if the channel is Pro AMD or Pro Intel, those will always show a better result than the one you are going after, so keep that in mind.
Tienes razón toda la razón y como dices ese target de fps dependerá de varios factores. La resolución y calidad gráfica deseada es importante como así el juego ya que depende de que tan optimizados estén. Por ejemplo The last of Us parte 1 es muy dependiente de la cpu aún a 1440p por lo que un procesador más rápido hace la diferencia en este caso. Pero nuevamente comparto totalmente lo que dices por eso al jugar en 1440p prefiero una mejor gpu y un cpu que pueda exprimir cerca del 99% a mi gpu. Además si activamos el trazado de rayos también estresan a la cpu por lo que como dices depende de muchos factores.
No, watch the video and understand. The use doesn't matter. Because otherwise we would have to test every single scenario in the world to fulfill all needs.
@@FO0TMinecraftPVP That's a real world scenario for maybe 1% of gamers. You would need tons of more CPU/GPU/resolution combinations to answer the question "Do I benefit from this CPU?" for a meaningful amount of gamers. And it would be needless, because everyone can answer this question for himself much easier. Just start your favorite games and look at the GPU load. Is it above 95% most of the time? You don't need a new CPU. Is it below that? A new CPU can help you get more FPS if you want. The rest is just a matter of looking at you budget and a CPU review to get the best CPU for your money and use case.
@@FO0TMinecraftPVP videos are almost like books: if u didnt understand something, watch again and again and again.. so just..watch again, he explained it :)
It's really simple to understand CPU tests done at low resolutions. It shows the max FPS the CPU can provide as a baseline. So even if you play at 4K you'd know that if you have GPU that can push the FPS really high. Then you know what the limit it would be hitting. Resolution doesn't matter to a CPU because it is not rendering the image at 4K fps. That is what the GPU renders and outputs. The CPU only puts together the physics and information feeding the GPU that data and at how fast it can do so. For example hitting 178hz/fps max in a given game/engine or 150 in another and 400 in another. The reason the resolution is low. Is to remove the GPU as the bottleneck because it matters when it comes to GPU. I have a 5800X3D and a 4090. Playing Hogwarts at 4K, I know my CPU cannot push past a certain limit in feeding the 4090 enough FPS. The 4090 at times sit at 50-58% Utilization. Now if I had a 9800X3D, I could potentially get the GPU to go full 99% Utilization and actually provide the full capability. So in a CPU test seeing the 5800X3D only reaching 93fps in Hogwarts (example I don't recall the real number) and my 4090 at only 60% Utilization. It's pretty clear to see it's being CPU bottle-necked and if I had a 9800X3D and it could reach 135fps (again made up example number). Then I could actually get the 4090 at 4K reach perhaps 124fps. Reaching my monitors 120Hz limit and be fully utilized.
I can literally make benchmarks and explain why those claims are dumb haha, but hardware unboxed already did that too and people still ask dumb questions
There is still one thing that I find very hard to understand here..
Why on earth does people find this topic SO VERY HARD to understand?!
I mean it is extremely simple.
A GPU is drawing a bunch of images, the CPU is then processing those same images.
If the GPU is capable of drawing 100 images in 1 second, but the CPU is capable of processing only 50 of those images in 1 second, you will get an FPS output of 50 images in 1 second!
If you raise the quality of the images that needs to be drawn high enough so that the GPU now is capable of drawing 30 images in 1 second, you have effectively moved you bottleneck / limiting factor away from the CPU and offloaded it on the GPU instead. NO MATTER THE DANM RESOLUTION USED!
Come on people. Is it really necessary to have a million different UA-cam channels going over this very same topic over and over and over again, seriously?
Ultimately, I think it comes down to people not knowing what specific computer parts do. It's the same people that pop up in various hardware discussions and only mention their GPU when talking about the relative power of their pc. I saw quite a few recently due to the discourse surrounding the MH: Wilds pc requirements.
@
You might be right about that.
Then I have another question: if someone is uncertain about what the individual parts actually do, then what compels them to ask a question that implies that a reviewer, who’s knowledge on the topic far exceeds their own, is doing his reviews or benchmarking wrong/that the way he does it does not reflect current setups that people have? 🤦🏼♀️
I am not saying that asking questions is a stupid thing in itself, my point is that if someone’s own knowledge is full of holes, the least that person can do is to ask their question properly, instead of making it seem like everyone else is in the wrong, especially the people that actually spends a hell of a lot of time on research in the first place.
The CPU is making the draw calls for the GPU, not the other way around.
@@hugoclery5789
I never said anything about draw calls mate.
All I talked about was the capabilities relative to each other.
says it's extremely simple then proceeds to explain it in a convoluted way🤣
That you have to explain this to people is mindboggling. I still remember the days when they were benchmarking Celeron's and Pentiums at 360p resolution in 1999.
Yup...good old times
Back in the day only nerds were into this stuff and you had to use your brain and read benchmark reviews, today PC gaming is very mainstream and people expect to be spoonfed information by techtubers, so can't or don't think anymore.
@@ooringe378 I think it's more that modern platforms like UA-cam make it easier for everyone to comment. In 1999 you basically had to write a letter to a newspaper if you wanted to complain about a cpu test not being in 480p. 🤣
I'm not saying it's bad, though. At least this way the people putting out the reviews can see what are some of the pain points of viewers.
@@ooringe378 ye. Now unfortunately explaining to some is useless because they really are brainless arrogant manchild. Instead of conceding they lack the tools and should educate themselves thw parrot the same points all over. It is just frustrating
Oh, I'm new on this, so it's good people explains why they choose those resolutions!
in short:
1. find a CPU review that also consist of the game you want at lowest resolution possible.
2. check if a certain CPU able to push X FPS that you want
3. find a GPU review that also consist of the game you want at resolution you want to.
4. check if a certain GPU able to push X FPS that you want.
5. get them so you can play at X FPS.
Doesn't require a bachelor degree to understand this.
steps unclear, now my cat thinks she is a bird and jumped out the window
@@georgwarhead2801 Situation unclear, what FPS was the cat pushing when it jumped out of the window?
@@AleksiJoensuu unclear, it had a ping of 200+ ms
@@georgwarhead2801 well, considering it's a cat. i'm glad and not shocked
@@hansyulian3671 Settings affect the load on the CPU. I haven't payed attention to it in a while but I remember they normally use the lowest resolution and the min settings. If that is the case then it isn't as straightforward as you mention.
I love that you have the patience to explain a whole video with another whole video, even if the questions are always the same every time someone tests a cpu.
I needed to make a video otherwise people keep talking crap
Kinda naive If You think You fix it with a video. ;))@@AncientGameplays
@@AncientGameplays indeed mate. But what a pain. Hardware Unboxed's latest has Steve suggesting links in the description to previous videos explaining why. You might do the same with this one, might save you some time in the future. 🇦🇺
"Why are they testing the speed of that Ferrari against that Fiat on a straight road without a speed limit? No one is going to drive at 200kmh all the time in a straight road. They should test it on regular traffic with regular speed limits which is what most people will experience in reality."
A good analogy indeed
Don't know if all people will get it though
People, don't look at this like resolution "X" or "Y". Look at it like this CPU gives "X" FPS in this specific game scenario :D
The resolution barely affects the CPU (usually doesn't). So, a CPU that with those settings and in that game gives 100 FPS WITHOUT A GPU BOTTLENECK, will give the same 100FPS at 720P, 1440P, 4K, 8K or 8,000K. The Resolution doesn't matter at all.
I think that 1% and .1% lows are also important particularly in fast paced games. If you have a frame rate target of 144hz or 240hz, then the closer your lows are to your target the smoother the experience will be.
Hi Fabio! People just don't understand that within expected resolution they play games at chosen quality game settings. Meaning for example: i can play Starfield at 1080p with Medium quality settings OR play Starfield at 1440p at Low quality settings. In two examples it might be that i could have exact same FPS.
@@flavoredmold6432 Yes, but I stated that aswell, 1% lows will improve a bit usually, but not as much as people think
@@flavoredmold6432 thats very important for esports games, especially for FPS title
@@AncientGameplays I get it but some people need to see data to really understand. If you used a 7700xt or 7800xt GPU with 9800x3d, 7800x3d, 7700x, 7600x at 1080p they would see some games become GPU bottled necked and fps at 1080p would be similar proving whats happening.
Daniel Owen tried to explain this on his channel for several videos, too. I guess some people just seem to have a real hard time understanding the difference between "CPU test" and "CPU buying advice".
Some people will watch a CPU review that's done in 1080p and go "that's bullshit I'm playing at 4K, how the hell does this review help me, why would I buy this?!" because they are looking for buying advice. They don't understand that watching a CPU review is just one component of understanding what you should or shouldn't buy. The process should be something like this:
1. Figure out which games you want to play at a) which FPS and b) at which resolution and graphics settings.
2. Watch CPU AND GPU reviews to determine a) which CPU is able to deliver that framerate in your chosen games overall, and b) which GPU can deliver that framerate in your chosen games at your chosen resolution.
3. Make note of what RAM, power supply, display etc. requirements have of course, too. Also consider what productivity requirements you may have for your PC, and how important those are to you, as the current fastest gaming CPUs are not the fastest in productivity.
But it's key to understand that your
- CPU determines if your PC can produce a given framerate AT ALL, with ANY resolution*
- GPU determines if your PC can produce a given framerate at A PARTICULAR resolution or PARTICULAR settings, but only IF your CPU is fast enough to manage that framerate.
This is why resolution doesn't matter FOR CPU TESTING. Or, more precisely, resolution should be LOW enough that the CPU can do everything it can without waiting for the GPU that's being used in testing. So IT IS BETTER FOR YOU, THE VIEWER, that CPU tests are being done at a lower resolution than you may be using.
And so finally: A CPU test shouldn't be understood as complete buying advice. YOU, THE VIEWER need to make the decision whether a CPU is a good buy for you. And IT DEPENDS on your other components. Do YOU have a fast enough GPU to reach the framerate the tester is getting at the resolution you want to play at? If NO, then it won't help you to upgrade to this CPU (unless you also upgrade your GPU). If YES, then it will help you to upgrade to this CPU. So you NEED the other component of watching GPU reviews also - this is WHY GPU reviews have multiple resolutions tested, usually. And of course eventually you need to consider price and value, something only you yourself can decide upon.
I don't know if I can explain it any better. I guess this is about 80% rant and 20% genuine advice but I still hope someone reads it and gets something useful out of it - or Fabio and Daniel's excellent videos.
* Note: Graphics settings other than resolution can affect CPU load, too, so they can make a difference. See kognak6640's comment below
Not all ray tracing are relevant to CPU or memory load, depends actually, but I agree with you in all the rest!
@@AncientGameplays Sure. I mostly meant that as an exception to the idea that ALL graphics settings yield the same result for the CPU. I edited the note.
@@AleksiJoensuu Any graphics setting decreasing polygons on screen will decrease also drawcalls per frame thus lowering workload for CPU(per frame). Draw distance, LOD settings, decreasing number of objects in scene etc are like this. It's the reason why I hate reviewers using low presets in CPU testing with assumption it only affects GPU side. Tells me they don't know how graphics rendering pipeline works. Lowered resolution with commonly used presets is the correct way test, otherwise you get skewed results. Resolution is not a CPU factor because it's internal for GPU. It comes in when true coordinate based 3D gets rasterized into 2D plane for the display at the end of the pipeline.
Ray tracing stuff is different, I'm not sure how it works on CPU side but I'm guessing higher resolution requires higher precision and density of rays and therefore extra work for CPU too. So there could be workload scaling based on resolution on CPU side if RT in enabled.
Additions:
Obviously you won't always know which games you want to play in the future, so you won't always have CPU and GPU tests available for your particular case. Then, you try to extrapolate by what tests you do have. If there's an upcoming game that's using, say, the CryEngine or Unreal Engine 5 or Frostbite engine or whatever, you can look up tests that feature other games that also use this engine, and maybe those will give you some approximations for how the future game might work. Or, you might look for a more general "this CPU and this GPU seem to match up pretty well in most cases". Or you might just go "I'll get the fastest I can from every category, who cares about the budget", or whatever.
Anyway the point is, testers can't give exact buying advice for your particular case, because they don't know every variable in your life and goals. They don't know which GPU you're using, which motherboard socket you have, what games you want to play, which resolution do you want to play them at, and what FPS targets you have in mind. That's why it's done this way. That's why reviewers give you one part of the puzzle, and your job is to find the other parts and put them together for your goals and your budget.
And with CPU reviews, the way to provide that puzzle piece for you, is to test in a way that lets the CPU run as unlimited by other factors as it possibly can. By using as monstrously fast a GPU, RAM and other components as the reviewer has access to, and then to test the games at a low resolution.
@@kognak6640 Good info. I knew this but omitted it, foolishly. I've edited the post accordingly. Thanks!
I appreciate you explaining this, some people don't seem to understand this concept when specifically testing CPU's it is very important to remove the gpu limited factor.
That is one of the best videos out for the 9800x3d because you highlighted the difference with the 5800x3d as well.
Yes that is why I loved it.
Ok, then we need 720p, 1440p and 2160p version of this same video.
😂💪
720P?
1080i? P or p?
In a CPU review the focus isn't on the resolution or GPU used but to make sure that CPU is consistently the performance limiting factor. Its that simple. We need to know the total amount of FPS the CPU can generate without the GPU becoming the limiting factor.
I find it hard to believe that people seeing your review have not already heard of hw unboxed's channel. They did this same video nearly 2 years ago and they post it in every cpu review they do since then.
And yet, here we are
@@AncientGameplays lol
Guys, not the testing method is wrong, you just have another question in mind than "what's the fastest gaming cpu?". Your question is more "money limited": Is it worth it to spend 500 EUR on this cpu to upgrade your individual configuration. In the end its totally up to you to decide that. Without knowing your system, use case, targets and most important, budget, noone can even give you a hint.
For higher resolutions, I always look at 1% fps lows. The smaller the gap between 1% fps lows to the average fps the better, consistent frame-times and smoother experience. I for example have a 4k 120hz monitor, which means that I want a CPU that the 1% fps lows it can deliver as much as possible close to 120fps.
That face on the thumbnail had me rolling 😂😂😂
Great job explaining how CPU benchmarks work. Keep up the great work, my friend.
Boggles my mind that in 2024 we still have people who do not understand the way their computers work. The higher the resolution goes, the more the GPU becomes the bottleneck, not the other way around. In fact, the only times there should ever be a CPU bottleneck in games at 4K is open world games like Cyberpunk with a lot of pedestrian and vehicle traffic. Even then, the bottleneck will be the same regardless of resolution. If you drop to 20fps at 4K, you'll drop to 20fps at 1080p as well, because the CPU has become the bottleneck.
It's why we know if there is a CPU bottleneck by dropping settings in game. If we see no improvement in FPS when dropping graphical settings, you have a CPU bottleneck.
@ChrisM541 If you are getting 20fps at 4K due to a CPU bottleneck then you absolutely will see 20fps at 1080p as well.
How is this hard to understand? Lowering the resolution doesn't help the CPU. The pedestrians and vehicles and everything else it has to calculate are still there regardless of the resolution.
Yes yes, great video. My FPS target is always based off the 1% Lows in relation to my Refresh Rate. Then i set my FPS to a little above that and just play at the smoothest experience. I do the same on All resolutions.
One of the very best video comparisons out there.
Thanks!
Once I heard Tech from The channel Teach Deal explained things easily for people. It was something like this: your CPU is creating/generating/ rendering all the structures you see in games, all floors, walls, objects and characters, the CPU is giving all of them the structure and your GPU is just painting everything making it looks nice. So, if your GPU IS NOT UNDER 100% usage it means you are bottlenecked by something else in your system CPU/RAM/SSD/software, usually is your CPU that can't generate enough structures and your GPU is just waiting dor them to get "painted"...
And the opposite also happens. A lot of "CPU COMPARISON" using like a Ryzen 5500 vs 5800x3D in 4k RT ultra and a 3060. Claiming that there isn't any different between theses 2 CPUs. In both scenarios the poor 3060 is 100% and 30fps.
exactly...sad
@@diegolce lol. Yes, sometimes i see comparisons, things looks weird, than you see they bottleneck one piece or the other making it completely useless
Hehe, at least in that scenario they are correct that the CPU isn't making any difference. But to then claim that there *IS* no difference between those CPUs together, well...
😂
I appreciate your efforts in explaining every aspects regarding pc harware Fabio.
The lower the resolution the higher the draw call from a gpu to cpu which becomes cpu/bottleneck and most people still dont understand that.
Thanks! I feel smarter today, because of this update 💙🏅
Thank you, feel glad someone learns and gets better!
Thank you for the video, you actually made me realize something!
While I always understood why people test at lower res for CPU (to avoid GPU bottlenecking), I always disregarded the results regarding practical implementation in my case: I still own a Ryzen 5 3600 to this day, coupled with a RX 6800, because I always gamed at 4K or at 1440p, therefore, seeing those CPU results always made me associate the FPS targeting scene as something that didn't made sense to me, as I don't play fast paced multiplayer games that require 200+ fps so you can gitgud.
However, when I play AAA or AA demanding games in 4K, I ALWAYS upsample at least from quality (1440p) using either FSR, XESS or library hijacks for DLSS. So I actually do play in lower res., and in this scenario, maybe the 1080p and 1440p CPU benchmarks make sense to me. For instance, eTeknix published a 2024 revisit of my 3600 in those demanding games about 8 months ago, and the results baffled me. While the R5 7600 performed on par with the r5 3600 in 4K (making old me think upgrading was a bad idea), in 1440p the min. fps jumped from 40 to 70 in CP2077 and other games. So I might actually be leaving a lot of critical performance on the table given that my lows start to get into choppy low fps territory.
Improtant to remember that not all areas of the game are equally heavy so it really depends, but to eleviate most drops you'd want better cpu
Shattered the world of "know it all" Intel folks. lmaoo I thought we all gamed at 8k already lmaoo much love and thanks for the hard work mate
Thanks for your very helpful and insightful video.
Thank you for this explanation. You did a great job 👍. Now everything is clear.
thank you
This discussion will never end because both sides have a point.
"1080p is the most used resolution" is irrelevant because that's overrepresenting laptops, gaming cafes, and people with really old hardware. AKA people *_NOT_* buying this stuff or watching _YOUR_ channel. These are hardware _enthusiast_ channels. If that's your argument then you should test everything with 60 class GPUs since they're the most prevalent as well, right? The argument for 1080p/4090 is for a testing environment _only,_ and even then only shows you how it compares to other CPUs.
It's that last part that the reviewers are so laser focused on that they don't see the problem. A lot of viewers *_WANT_* to know where that bottleneck kicks in with higher resolutions so they don't overspend on their CPU, or even worse spend a bunch to replace a CPU, motherboard, and memory that was still completely fine at the resolutions and settings they play at. Knowing a CPU's theoretical max is useful, but in the real world it's just that. _Theoretical._ It doesn't tell you how it will perform with _your_ GPU and resolution. Reviewers are so focused on that theoretical max because they already have the hardware to know where the bottlenecks occur so they can avoid them. But if you never show _where_ bottlenecks occur the consumer has no idea until they buy the hardware to find out they made the wrong purchase.
Seeing reviews that only test tires on Ferraris and Lamborghinis tells me squat on how useful those tires are on my Toyota. Telling me with a voice of superiority "Well, obviously the max is 200mph right there in the test. Stop asking stupid questions." is completely useless information when you don't take into account the speed limits of a car most people can afford, let alone the real roads they drive on. Car sites don't do this. Boat sites don't do this. RV sites don't do this. Only PC hardware sites do. NO other hobby I'm into does testing with this mentality. Let alone tell your own viewers they're stupid for asking for real world scenarios.
No one's asking for 1080p testing to be thrown away. We aren't stupid. We KNOW it's needed for a baseline to directly compare CPU to CPU. All we ask is for a random 1440p test thrown in with a 6750xt or 4070 to realize the massive waste of money a 9800X3D is if you can't afford a $2000 GPU and your 5600x is still more than adequate. Especially when GPUs have gotten so expensive that CPUs get more and more irrelevant by the day. It's not about accuracy of the scientific method. It's about providing a service to your viewer.
Couldn’t agree more with this! This is exactly what a lot of people have been thinking. It’s frustrating when reviews focus so much on theoretical maxes, but don’t show how things perform in real-world scenarios where most of us are gaming or working. Like you said, we don’t need 1080p testing to be thrown out, but it’d be great to see something that gives a better idea of where the bottlenecks actually start to happen when you’re at higher resolutions and using mid-range GPUs.
EXACTLY THIS! ZOMG I couldn't have word it any better, that's what the "1080P" and "720P" supporters DO NOT UNDERSTAND.
People are not asking for low resolution tests to be removed, we actually understand the purpose. :D Knowing the info at 1080P doesn't translate into how the game performs at higher resolution as it depends on the GPU vs resolution vs game engine vs settings.
If ever I make such a content (actually I plan on making one) I think I want to use a screenshot of your comment here.
TIA!
@@goldfries Feel free, but that's a lot of text for a SS. 😂
This is exactly what I've been telling hardware unboxed to do but they are arrogant fks and refuse!
1080p is the majority is complete bs! The data is probably accounting for 3rd world countries in Asia where people get $1 wages!
@@Grievous- but... you already have that info, the video will just be titled GPU review. They already usually show you 1080p, 1440p, 4k data in the GPU review of your chosen GPU. You already know where the bottleneck happens, you're just watching the wrong video and reiterating GPU testing results in a cpu review says nothing you don't already know.
Well done Fabio, great vid
I experienced it myself how upgrading from Ryzen 5 3600 to Ryzen 7 5800x3D change the entire gaming experience.
The max fps is same, almost never increase (I play on 2k resolution).
But the frame time is smooth af. No stuttering, less input lag, and no flickering.
Even after I bought 4K monitor, as long as my low fps is 40-60, it's still enjoyable.
But still, for me who gaming on 4K, every CPU video test is useless, because it will never increase my fps if I upgrade my CPU again.
That's definitely not true. Just because games are almost always GPU bottlenecked at 4k doesn't mean they can't be CPU bottlenecked, which will happen with your CPU eventually if you continue upgrading your GPU. So look at it like this, whatever your fps is when you are CPU bottlenecked is the fastest the game will ever be able to run. If you have a game that can't run at your desired framerate at 720p or 1080p because it is too CPU intensive for your CPU to handle, then getting a faster GPU and playing at 4k will not get you any more fps, but upgrading your CPU will increase your framerate at all resolutions.
I'm hoping I'm just misinterpreted your comment or taking it too literally.
@The_Chad_ What I mean is, my PC is already peak for 4K (I tried overclock, try any driver, cleaning/repasting and it still never increase my fps except using upscaling)
so upgrading CPU is useless so any CPU video also useless for me.
I need to upgrade my GPU and then probably upgrade CPU again because I already experienced how powerful CPU can impact gaming experience (not looking for higher fps but smoother experience if my Ryzen 7 5800x3D having stutter like my Ryzen 5 3600).
@@wisnudJat ohh ok lol. That makes more sense. I took it as keeping the same CPU for decades while you continue to upgrade the PC around it😂
am still running on 1080p haha. If the monitor is too big, i would get headaches. Good video thanks, your room's looks sick man.
remember intel fanboy was only using 720p benchmark and bullying Zen1-2 users now it don't matter anymore........
That's framechasers ?:)))
This kind of nonsense happens both ways. AMD fans now are saying the 60% power increase for 8% performance is a non issue while with intel 30% power for a 11% uplift was unacceptable to them... to me and my use case that added power draw on the 9800X3d and the 25 watts more idle power draw over the expected 5 yearsd lifespan will cost me 240 euro more in power usage. i calculate that into cost of running and compared to the 7800X3D it makes the effective price of the 9800X3d 640 euro (thats 700 dollar).
It's crazy to me i have no bias just demands of products. If something is positive state it, if something is negative state it. No product past 5 years has bin perfect.
I now run intel cpu +amd gpu and soon i will run amd+amd and later in time i will likely run amd + nvidia as amd threw in the towel on high end gpu.
And still 12 threads zen 1 shows better low and 0.1 than 4 cores i5, cost same and have future to upgrade lol. Me personally came from 1200 to 2600x, and now i got 5700x3d for 135$ from ali, while having same msi b350m 50$ motherboard from 2017. God bless Liza Su
@@auturgicflosculator2183 but why would i spend 150 more money then a 7800X3D from a budget bin when i have to enable eco mode to make it effient for zero gains over that 7800X3D. it just makes no sense. effectifly its just AMD pulling a intel with massivly more power draw for marginal performance gains in certain games. cyberpunk was 60% power for 8% performance and thats just silly. im here hoping for undervolt and curve optimazation that fixes some of this nonsense.
Studies show that 3 out of 4 don’t think bullying is bad. 🥲
Thankyou so much for the great content and you're right getting a new GPU is the way to go in the short term. I am waiting for the 8800 XT early next year but still I am looking to also upgrade my 5800 X3D as well next year. Great vid.
Lmao I've been watching Daniel Owen try and educate people about this recently, sucks people need to make these videos on such a (to me at least) basic concept, but I'm glad we have people like you and Daniel willing and able to do so.
Hopefully you guys are putting it in ways they can understand so they can make better informed purchases, I always hate to see when people buy the wrong hardware combo because of bad research.
Guess I need to a livestream with Daniel Owen haha
I have a 1080p 240hz monitor, and my specs are a 7800x3d with a 4070tiSuper. I use 1660p with DLDSR and its amazing.
According to Steam Hardware Survey for October 2024:
1920x1080 56.98% meaning more than half users actually use 1080p resolution
2560x1440 20.00% meaning every fifth person use 1440p resolution
3840x2160 3.68% yep, 4K users, right...
I think above stats speak for itself.
Does this include upscaling? Because i think most gamers are actually gaming on 1080p lol with dlss balanced on 1440p😂😂😂
everyone that has glasses plays at a lower ress since my 24 monitor are like 17 inch for me - I need a 32 inch in 1080p or a tv
@@spookyskellyskeleton609 you can also ask if does this include Frame Generation. Which might be also a good question to yours! 🙂
@@asdfqwerty-u9nbro . When Ur blind making things bigger doesn't help You cure your blindness.. change the lenses :)))
Sorry i wear glasses too but this was funny for me.
Stats don't mean shit for this situation. All the people at 1080p have lower end hardware, not a 4090 and high end cpu. Basically nobody playing at 1080p is gonna care about looking up benchmarks of high end hardware.
Hi ! I continue to play in 1080p60 ! My 5600X stay great ! ^^ Good video !
There are people who genuinely think that if a game causes a CPU bottleneck, then raising the resolution will reduce the bottleneck. No, Timmy, that's not how it works. You'll have exactly the same cpu bottleneck in all resolutions...
exactly! You're just shifting from a CPU to a GPU bottlleneck. If you have a CPU bottleneck, you better just lock fps with Radeon Chill or RTSS
Entirely and completely untrue. 4k uses far more of the GPU power than 1080p, raising the resolution makes the cpu speed and ipc matter less the higher you go up.
@@drunkhusband6257 matter less comparatively, which means the cpu is still going to do the same workload, but at 4k chances are the limiting factor is the GPU, so a cpu can be much faster than another one and will produce the same FPS on 4k native res ultra settings.
But that's just a stupid way to think about things, it only makes sense to think like this if your only usecase for the PC is to play games that demand way more from the gpu than the cpu and you have no intention of keeping the same system for multiple years as is.
@@ChrisM541 you'd be been lied to if they only tested in 1440p Ultra and 4k, the benchmark's purpose is to show the improvement of one cpu over another, not to show real performance on a GPU bound game. Did you even watch the video?
also if you want to see "real use" scenarios there will shortly be no shortage of random smaller channels uploading every combination of cpu and gpu in different games and settings after every new release
Stop trying to find fault in something that is not faulty at all
To be fair if you go game at 4k with 8x multisampling you will no longer have a cpu bottleneck @ 9 fps :P
It's 2024 and people now don't know a thing about CPU performance taxing, or about their pc in general. Idk, I don't remember it being this bad quite a few years back, was it?
Awesome video as always.
Took me 8 years to upgrade from a Phenom x4 to a Ryzen 3600 because the gains weren't substantial until then. I've been using my Ryzen for the last 5 years and haven't felt the need to upgrade since I play below 60fps at 1440p.
exactly as I stated in this video. Playing at 60FPS? almost any CPU nowadays does it in almost any game
I have one of the new 240hz 4K OLEDs, I currently use a 5800x3D and a 4090. I play all my games at 4K but if I want to take advantage of my monitors high refresh rate I need to play with DLSS which shifts the bottleneck over from my GPU to my CPU. He is 100% right about FPS targets, if you ONLY play 4K native without any upscaling then your target FPS will be lower and your required CPU will be lower because even a 4090 at native 4K will have a hard time reaching high FPS.
Many people nowadays going with 1440p after buying new PC, but there is still a difference on those processors even on cards like 4070 Super or 4080. It will be even more visible in next gen with something like 5070 Ti or so.
Unless they're playing light games
It's crazy that you have to make this video, I thought HUB already straightened these folks out.
You should do a video about only competitive games in comp settings with this kind of cpu(9800x3d) because lets be real most the people that are buy this high end focus in games cpus are gonna play in 1080p 1440p competitive games and is when this cpus really shine
Thumbs up! I hope people get it! For me it is clear but there are still so many out there who does not understand why a CPU is testet Benchmarked in 720p 1080p. Of course we need also Benchmarks in 1440p and 4K and we have to Bench AMD CPU with AMD GPU and the same goes for Nvidia. Intel is... Not really here anymore 😅
Intel really need to come up with their own version of x3D vcache, they are gettimg murdered on gaming purposes, their fastest cpu is still 14900k which eats up way more power, not to mention killing itself before microcode patches, we need competition to prevent stagnation.
agreed
Again, only murdered at 1080p, nobody with an x3d cpu or 14900k is using 1080p, you people are brain damaged.
Very interesting analysis Fabio, as I play mostly in 4k whenever possible, I will keep my 7800x3d, also because I have seen in the reviews that it uses considerably less power than the new 9800x3d
in gaming the power used is around the same, in render it uses more. I have that on my review video as well. At 4K you'll be GPU limited in most scenarios, so there's no point in getting a better CPU even if you're using DLSS because after all, you have a 7800X3D
@ Thanks for the clarification Fabio
Its 1:30 AM here, I should be sleeping but screw it, Ill start my 2 PC builds now while listening to your video 😂 the parts are sitting for a week now 😂
Hahaha
Thank you, I'll look for other CPU benchmarks👍
You should
i remember daniel owens also have to explain this 😂, thank god you have the patience to explain this carefully and clearly.
i still don't think people get it ...........even though he's explained it..........this will be asked many more times.......
@Anon1370 sadly
Even at 1440p you wil see a difference in the 1 % lows. Thats actually whats most important not the avarage fps.
So people are just being lazy not including the 1440p benchmarks. Even at 4k with RT you see a big difference between cpu performance. LIke Dragon Age Veilguard.
So only testing 1080p is outdated.
There's plenty of benchmarks that show the % difference of 1% lows between 1080p and 1440p between cpu comparisons is the same. So you're asking for redundant testing when you can get the same % diff, also GPU bottlenecks so you most likely would see the 1% lows diff be tighter and less pronounced.
@Mathster_live Its not the same for every game. Thats why having the 1440p is a good thing that gives you a more complete picture. It doesnt detract from anything it just adds to it.
Its similar to testing gpus at 1080p and we still do it.
again, resolution doesn't matter. People aren't being lazy, you're being Dumb. Because if you're testing a game engine that is completely GPU driven, even the 1% lows will be equal. Also, you can test at 1440P but with a 7800XT instead of the 7900XTX and in that case, you're 100% GPU bottlenecked, so no difference in the 1% lows as well. You guys need to try and learn for once
@@dante19890 It is the same for MAJORITY of games, the exceptions are not the rule. Also, RT is a special case because it really depends on the implementation, and a better CPU does help and sometimes it doesn't and from what I know of Dragon age veilguard benchmarks, it's apparently something with the driver version affecting the performance with RT with 9000 series cpus
@AncientGameplays its the same argument for testing gpus at 1080p. We still do it even with top level cards even though it's being bottlenecked at that res.
The same argument can be had for testing top cpus at 1440p so it is just laziness not to include it. It's something that also is highly requested by people.
Fabio, it's 2024 we should be doing 16K resolution tests. Please, we skipped 8K and are already at 16K. /s
I still do hope you do tests with some older CPU/GPU combos like the Ryzen 3000s vs newer stuff to see if some things are worth upgrading vs price which would include mobo/ram. And keep up those great undervolting/ocing vids.
😂😂💪
More than "not understanding", I think that what many want is for you to give them everything already done, to reproduce in your tests exactly the same scenario that they have in their setups and thus not have to make any effort of interpretation...
Yes, people are lazy nowadays
Fabio, I would even argue that 1080p is TOO HIGH for a CPU test. A CPU gaming test should have, ideally, the ultimate in potato graphics settings. I'm talking 720p or 900p AT MOST. Then there would be no question about whether or not the GPU is interfering.
Reason I always loved anandtech was their 720p and lower cpu tests.
720P is not realistic nowadays I believe as noone uses it, but 1080P? Sure, still the most used one
@@AncientGameplays i have follow up question for 720p, i know its unrealistic, is there still games that being bottlenecked by 4090 in 1080p? if that happens do you usually give extra test in 720p if that happens?
720p is in fact deprecated, at least 1440p quality(960p) , cuz u dont use 1080p dlss quality, looks like shit almost always, if u want 1080p dlss quality is cuz ur GPU is the problem. REAL CASES: 4k dlss performance , 4k dlss balance, 1440p native, 1080p native,1440p quality(960p). 1440p quality will be for me the best test, and that is 960p( looks the same has 1080p native usually)
Most reviewers were testing cpus in 720p 2-3 years ago. Now almost everyone switched to 1080p. Probably due to upscaling being common now? Or the FPS at 720p was too high and some game engines were having artificial lock so it was affecting the results and providing misleading
@@hensem4891 if the GPU is the bottleneck with 1080p, reduce the quality settings never the resolution, 720p will destroy your game, looks really bad.
Before 1080p, in GTX900 era and earlier, 1080p CPU testing was 720p CPU testing, even though NO ONE EVER played at this res on a modern machine (laptops aside).
You pick a cpuFPS and gpuFPS and get the lower one and voila you have your rough resulting FPS.
@Ancient Gameplays:
Maybe look at it in a different way.
There are two reasons to watch 9800x3d reviews:
- Theoretical/Coolness factor: People who want to see performance of new cool 9800x3d
- Practical: People who want to see if it makes sense to buy 9800x3d for their setup, or maybe something cheaper will work as well
Reviews at 1080p satisfy a lot of people.
Your channel can attract more viewers with honey than vinegar.
Instead of saying "It is pointless to review at 1440p" why not do 1440p review?
=> You will get views other channels don't, due to unique content.
=> You get more viewers happy, so more subscribers.
=> You can save people money by showing which cheaper CPU work as well as 9800x3d at 1440p.
People love saving money.
I just don't see a downside to doing special separate 1440p review of 9800x3d.
There are videos that claim even with 4070 TI on 1440p there are benefits to using 9800x3d versus cheaper 9700x.
great video as ever
Tech Jesus is back
🤣🤣🤣🤣🤣🤣
Your Witcher 3 Wild Hunt FPS is correct, I was playing it yesterday and I had Hairworks cranked all the way up. Frame gen adds another hundred to it. My hardware config of course.
correct me if I'm wrong. are you saying if you got a GPU and the best CPU and tested the GPU FPS on 1080p & 1440p the FPS will be the same so there's no point in doing the 1440p test.?
Yes, this is precisely what he mentioned; now I have to purchase a 1080p monitor to view more fps and then return to my 1440p monitor to observe the exact performance. 🤭
However, we are the foolish ones who fail to realize that if the CPU performs well at 1080p, the rest is insignificant if you gain only 5fps more at 1440p with the new CPU. 🤣
If the internet claims that 1080p is used by many, Steam represents everyone globally who uses a PC, and we must rely on that data, so can you understand what we’re facing here?
If you are gaming at 1440p and move from a 7800X3D to a 9800X3D, you'll see a performance boost of 5-10%. Does spending 500$ make sense for that increase?
No negativity! I simply can't resist commenting when I observe individuals deceiving others with a significant falsehood, claiming this is the greatest CPU of all time; I despise fakers and those who mislead others.
How can I upgrade my CPU if I use 1440p? Should I check a video on 1080p tests?
@@Silent-1983 i did say correct me if I'm wrong. i cant tell if you are correcting me or being sarcastic towards Ancient Gameplays. ?
@@DEM1GOD7 he used 1080p in testing to remove the GPU bottleneck and instead let the CPU do all the work to see how much fps it can push. If it can push 500 fps in one game at 1080p, it can do 500 fps on any other resolution. Now you need a powerful enough GPU to get to that 500 fps at higher resolutions. That's what he's explaining here. He was testing the CPU, not the GPU.
@@ShaiiShaiiShaii Ah well i wised he said it a bit better they way he explained it was very confusing
Exactly, if you have no GPU bottleneck, you could test the CPU at 480P, 720P, 1080P, 4K or even 8K. No GPu bottleneck? The CPU would deliver the same FPS technically
im with you on all this ancient lots of people skip parts of vids jsut to get to what they want to hear but not me im here for all of what i can learn and possibly in irl tell soemoen else who is need to know about all of this or i can always redirect them to this video as well. and resoultions do not matter if your cpu is bottlenecked and can not pus hup to 100 fps your not able to run higher res but if it can like you said then your perfectly fine jsut as long as the good ol trusty gpu can keep up on what the cpu is doing plus there is so much more to consider as well liek you have also said.
Great video Fabio.
Higher resolutions have close to no impact on CPU performance. ''Close to no impact'' because as the resolution increases, the framebuffer size increases, and that requires more memory bandwidth for transferring data between the CPU and VRAM. That increased data transfer can ''technically'' lead to slightly higher CPU usage. However, that increase is generally not significant compared to the overall workload of the GPU, which handles the bulk of the rendering process.
I feel like people who don't understand bottlenecks don't know that data is sent from the CPU to the GPU.
Seems this kind of clarification needs to be made at least once a year ... I've just bookmarked this and Daniels video now to send to others in the future since i got tired trying to explain this to people that rather get their "knowledge" from tictoc or reddit.
All hail GVG MOOOOOOLL
hahahaha
absolutely this is a well spread misinformation i agree with you when testing cpu res shouldn't matter the only thing matter is the fps limit the cpu can reach but since 1080p is the most used res its better to test at 1080p bravo and thank you for clearing this misinformation about cpu testing
"You can't fix stupid"
- some people
+++
you can but its not legal.
@@pino_de_vogel I can still fantasize about trying.
@@shannonhubbard4026 Oh i do all the time haha. I'm tempted to pour a foundation for a home patio all the time. I just don't have a propperty big enough to make a patio as large as i would need.
Fabio, greetings from Latvia. How long do you think my R9 5950X will last, in games of course. Let me adjust the question, what is the most powerful video card that this processor can handle? Thank you.
Depends on the fps you're aiming at, basically that's it
@@AncientGameplays I bought an Asus VG32UQA1A (4k/160Hz) a month ago. Older games can reach 160Hz, but modern games are a complete failure. I have a Pure RX 7700XT in my computer right now. Will it make sense to replace it with a conditional RX 8800/8900 XT next spring? The current card has a performance of 35 teraflops, will it make sense to put a card with a performance of 50 teraflops to this CPU? Or Zen3 is already the past...?
Maybe they wanted to see the frame drops for CPU. X3D CPUs provide better 1% lows.
not entirely true all the times. Also, in some cases the difference would be null even in the 1% lows
@@AncientGameplays True, I think it happens when the GPU is under extremely heavy usage.
Thanks aot for the work you do. It is sad that we hve a cult aslready of people who cry about using lower resolutions for testing CPU. But you really gave a nice explanation. IF my 12600k+6800XT cannot deliver the FPS i need i will change res and lower settings and i will find out the component that is bottleneking. I wish one day when you have time to test World of Warcraft. Alot of people play it and my 12600k+DDR4 4000 cannot deliver 60 FPS in raids or in the main city. No matther what resolution i use i cannot go over 60 if there are many people around. Still some say you do not need a better CPU at 4k. Same thing for ACC. I have a 2k 14 Hz monitor and i cannot see more then 120-130 FPS. thanks again.
Fabio amazing thumbnail lol. October 2024 steam survey, 1080p is at 57%, still the majority of the player base.1440 is only at 20%, and 4k is at 3.7% so completely irrelevant. The higher the resolution the more pixels need to be rendered and that is far easier for the GPU to do then CPU.
The proper and definitive way to test 1080p is i believe at low settings as well so it offsets everything to the CPU and not the gpu. That's the way it's always been, people just dont wan't to understand that at 1440p+ the difference between CPUs won't be anywhere near as high as 1080p.
This is also why the new 24h2 update for W11 makes a much bigger difference at 1080p, it's all about the frame calls.
Unlike in the past, there are many graphics settings that affect the CPU usage today. Ray tracing is one of the bigger ones.
@@monotoneone Correct, but not a single person in their sane mind uses ray tracing because it's pointless and a waste of time lol. I love that some games will tell you in the settings if it affects CPU/GPU or both. Every single game should be like that, it's just text doesn't affect the game performance at all :).
Nobody playing at 1080p owns high end hardware, and most aren't looking up benchmarks on you tube to begin with. 1080p for benchmarks should have died in 2015
@@drunkhusband6257 another one who doesn't understand how a CPU benchmark works lol.
Thank you very much for your information I learned many things today❤
Thank you aswell!
It's not that people don't know what the aim of 1080p benchmarks is. It's that people who are thinking it's all about 1080p benchmarks (or benchmarks at non actual use-case resolutions) are missing the point that the CPU performance at real-world use-case resolutions is the actionable information that is relevant to people's interests when learning how a CPU performs. It's what purchasing decisions are, or should be, based on. Showing spreads at resolutions that are beneath what people will actually be using, when there will be no spread at the resolutions they actually use, is misleading in terms of messaging of what people should expect from the CPU.
When the results of multiple CPUs show the same performance at 1440p because even the most powerful consumer GPU can't deliver higher FPS... that's good, actionable review information that lets people know they aren't going to get any more performance if they shell-out for a more expensive CPU. And very-few people even have an RTX 4090. Even in another few years, few people will have RTX 4090-level GPU performance in their system - which is something I go into further detail of in my first reply below.
And people saying "then you're testing the GPU and not the CPU" aren't correct, as the GPU is only part of the equation. It's testing the CPU to the extent it's able to stretch its legs with (typically the maximum) GPU performance that's available, and letting people know that they aren't going to get more performance out of the CPU than what benchmarks at real-use resolutions show.
Even in another several years, the large majority of 7800X3D and 9800X3D owners still won't have RTX 4090 GPU performance to go with their CPU. So, showing 1080p benchmarks without all this additional context to interpret the 1080p results is misleading and makes people think they're paying for performance they probably won't ever actually see. Reviewers need to think about the reason people are watching reviews. Showing FPS spreads at 1080p for a top-end gaming CPU is just academic information. It's not a practical information for consumers. By all means, show the 1080p spread. But don't do it without showing real-world use-case gaming results at 1440p and 4k.
BTW, Linus addressed this topic in LTT's 9800X3D review, and noted that 1080p benchmarks aren't relevant for most people, and that looking at 1440p benchmarks shows there's not really any reason to upgrade to a 9800X3D if already owning a 7800X3D and only using the CPU for gaming at 1440p or higher. That fact is hidden in 1080p benchmarks, and so only showing 1080p benchmarks misleads people into paying more for something they won't benefit from.
still....why? if the 9800x3d can give me 200fps avg on 1080p and i know that a 4090 maxes out at 100fps avg in 4k, i now know, that the 9800x3d can support GPU´s twice as fast as a 4090 in 4k. if i only test this combination at 4k, i will NEVER know, if this cpu can handle a faster future GPU for example... now if you want to pair your gpu with a perfect cpu combination, just look at a GPU benchmark with your resolution, pick the GPU you want, look at cpu benchmark and ceck for the cpu that can deliver you that amount of fps and done
@@georgwarhead2801 But why would you only test in 4k? Likewise, why would you only test in 1080p? In typical use resolutions, you're unlikely to ever see the spread you do at 1080p in your lifetime use of a top-end gaming CPU. To get to see that spread at 4k, GPUs will first have to become so powerful that they're able to push the FPS at 4k that an RTX 4090 pushes today at 1080p. How many years away do you think it will be before there's even a GPU that can do that: maybe 5+? And then how many more years do you think it will be before even 25% of gamers have that level of GPU performance in their system, to make today's 1080p benchmarks relatable for a notable segment of the consumer base for that CPU?
So, do people want hardware review information that informs them of what value a CPU gives them today, or of what value it could hypothetically give them in 5+ years, if they buy a top-end GPU at that time? When I buy new hardware, I want to know information that can firstly be applied today. And I'm not planning for 5+ years in the future, because I'll have bought a new CPU by that time, anyway.
Reviews containing only academic FPS stats at below-typical-use resolutions is at least as bad, if not worse than if they only contained benchmarks of typical-use resolutions. Because the purpose of a review is to inform for a purpose (in this case consumer purchasing awareness), and the typical-use resolution data would probably be much more helpful to a lot more people. But reviews should contain normal-use FPS data, to go along with the academic 1080p data.
@@thischannel1071 its the same reason, why you dont test a 4090 paired with a12100f and than start argue that at 4k all gpus are the same, YOU WANT TO REMOVE OTHER FACTORS that can limit the performance. if you want to pick a cpu for a gpu, you look up benchmarks for your resolution, pick the gpu that fits your needs, look up cpu tests and pick one that fits in the fps range of your gpu ( BECAUSE THE RESOLUTION DOESNT MATTER) and BAM done, otherwise you have to look 100´s of possible configurations beeing tested and you come up with the enlightement, that this tests dont tell you how much faster this is then your current build.
@@georgwarhead2801 Saying you'd only test in 1080p to remove all other factors, to get academic results, doesn't answer a question of "why would you only test in 1080p" - it ignores the point that non-practical, academic data isn't only what a product review should have, or even what it should focus on, and the point that practical use-case is a fundamental part of a product review. Like I said in my previous comment, reviews should contain normal-use FPS data, to go along with the academic 1080p data.
The resolution absolutely matters, though, because people use their PCs in certain resolutions. And there aren't that many resolutions that are used as standard: 1080p, 1440p, and 4k are the main ones. But if benchmarks are done in even just two resolutions, such as either 1440p and 4k, or 1080p and 1440p, then that creates a scaling trajectory, and people can generally mentally apply that scaling to figure-out where their CPU and GPU would sit along that trajectory.
I couldnt have said it better. I wish every reviewer saw this comment. For some reason people only think in black & white when the solution is SO simple: test the cpu in all 3 resolutions. Problem solved. People who want to see the raw improvement of the cpu over the competition and how "future proof" it is, can look at the 1080p results, and the people who want to see if it will be an upgrade for them specifically, right now, can look at 1440p/4K. Thats all there is to it. How is this so hard to grasp?
People think that 1080p is irrelevant when it comes to gaming and benchmark and they boast 1440p, 4k gaming is the new generation yet they still complain people will never be happy or satisfied no matter what, heck my old hardware R7 5700X is still doing fine at 1080p and I'm happy for it.
1080p still relevant. People want also to know for 1440p because a vast majority of player are on 1080p and 1440p. But guys. He's right most people still on 1080p and usually upscaling goes from 1080p to 1440p if you want the best quality. And can go even lower if you let the base settings.
no, upscaling goes from 960P to 1440P xD
@AncientGameplays Depends on setting but yeah.
i wont go beyond 1080P until it gets obsolete!
@@zf8496 Resolution do not become obsolete. It's just the more resolution you have the greater the screen you can have without having more pixelated images. That's why I'm sticking to 1440p 27 inches
@@FireCestina1200 yep correct I get it but still don't want to spend so much for better quality that's my motto lol
Hey ancient, you think you can do a vid on educating people about fake benchmarks also? It would be nice to know who to trust when it comes to scoping out an upgrade. UA-cam has a lot of channels with interesting technology combinations but many people say they’re fake. Would be nice to know who to trust when it comes to benchmarks.
haha, Intel fanboys fuming hard that their brand is underperforming
hardware canucks had same thing, in short, 8 years ago Intel fanboys were all for 720p testing since it gave better numbers for intel, now it's other way around, so it's userbenchmark level of cope, don't mind those Fabio
AMD fanboys crying because they cant get hold of an 9800x3d 😂 Its not about being a fan, its about having competition for competative prices, and right now the consumer is pretty much screwed. We need intel to compete. Its not rocket science
@prussell890 you missed the whole point, but ok
@@prussell890we need Intel to compete yes. But consumers won't buy the shittier product just out of love for a corporation. Or will they ?
See this îs why so many Intel fans are brainless. They have a better cpu for better price and better cooling and power usage but they are still choosing Intel because of confort and loyalty and hatreed in team Red. The story of producticity comes here . Really now . Producticity...majority of Intel i7 i9 users are just gaming. They don't produce shit. U dont need i9 to scroll webpages and use multitask. Almost no one will actually feel the difference in day to day use between i9 and 7800x3d.
We need Intel to compete. But we need to punish Intel when they don't compete. So they will compete. My last processors were i5 7400-i7-7700k-i5-11400-now 5800x3d.
AMD was punished everytime by the consumers on every single mistake they did with their producta. Just imagine If AMD pulled this 14900k 13900ks overvolt instability problems in one of their products. Like it happoened with bulldozers. AMD almost went bankrupt because of the anticonsumerism wave that came from Intel fanboys. And now they(Intel board) are knocking on lisa's door proposing deals and partnerships. Life is life. Intel fucked up a few years in a row.
@pR0ManiacS yes i totally agree with you. I dont think the latest intels are bad, they are more efficient and cooler, they just cannot compete with X3D. However in the uk, the ultra 7 is £350 and the 9800x3d is £450. Thats 25% more expensive. The 7800x3d has been discontinued, the 9800x3d is out of stock and the ultra 7's are available to buy. Paying that price really depends on how patient your are waiting, your type of gaming and gpu. Personally, I would obviously prefer the 9800x3d, but I would not be disappointed with the intel ultra range.
@@pR0ManiacS Unfortunately that is current trend of things to come. Especially with Intel and Nvidia fanbois 😂 with the latter being the worst offender.
I know it’s a bit off topic, but could you make a video with the best set up for amd adrenaline settings for video cards?(please 7900xtx included).
With all the new features and functions, it’s a mess to optimize all for the best results. An updated overclocking setup would also be appreciated. Best regards, keep up the good work!
Most people who only use PC to game aren't understand how to do benchmarking and are also too ignorant to try to learn how it works. They just want a quick answer of how it perform with their hardware and their resolution target. I saw other youtubers explaining the same thing over and over about why people benchmark CPU on 1080p. I doubt they even read your replies.
And that's why I made a video.
@AncientGameplays tell me you still love us ignorant gamers daddy 😢
A key point is the same game can be both GPU & CPU bottlenecked at different times.
So there's weaknesses in even the target framerates as in some games other things matter too.
Personally buying the 5800x3D was for reducing late game lag to avoidable situations like watching massive battles close up, where the GPU was hit hard in games like Stellaris.
I was pleasantly surprised at the quality improvement in other games feeling smoother, despite the previous 5600x having faster ST benchmark scores.
i think u underestimate average joe's stupidity so the quotes "why testing cpu/ram in 1080 its dead" will be there like...forever (sigh)
I guess I do, hope to help things change every day though!
This just speaks volumes in terms of peoples arrogance and ignorance, this doesn't just apply to benchmarking and pcs, this is happening in almost every subject.
Doing a CPU test while being GPU limited is like doing a drag race on public roads and staying under the speed limit. It makes no sense.
Incredible that people don't know this and you've had to make a video to explain it. Good explanation video just like all your videos, keep up the good work!
Tho I agree that 1080p is the best resolution to test the CPU power in gaming, it wouldn't hurt to include 1440p and 4k benchmarks.
I just hope techtubers do it like that. It's hard to find reliable benchmarks for 1440p and 4k that compares new and old Cpus
It would hurt very much, people have no notion of the amount of work I have to test just one CPU, especially since I already test at stock, Tweak RAM and OC
@ makes sense. I only watch the output and don't really know what's going on in the production part. I'm sure it's a lot of work. Keep it up, I'll be watchin
if they show tests in 2k/4k resolution, they will be forced to agree that the performance of current processors is more than enough for top-end videocards to work at their maximum, which means there is no point in upgrading the CPU. This will slow down sales. It's not profitable for them!
Reviewers will give various reasons for this, but keep in mind that this is the main reason. Money first!) People look at the numbers and don’t fully understand how it works. Unfortunately, most casual players, looking at 1080p tests, think that they will get an increase in 4K resolution as well, when buying a new CPU. And to leave them in blissful ignorance, 2k/4k tests are not done) This gives a boost in sales at the start. Nothing personal, just business😉
Hope you read this. So would you say that FSR/DLSS is actually hurting performance vs native resolution in MMOs, which tend to be CPU taxing?
It only hurts performance if you're cpu bottlenecked
@@AncientGameplays Well... some of the MMOs I play even make the 9800X3D sweat...
Finally someone with brains. People don't even understand about cpu and gpu test
Thanks for the explanation, I knew that lower resolutions are used to test CPU’s, I just didn’t know all that nerd stuff 😂
Love 1080p and always will. For me personally its the perfect compromise between efficiency, visual pixel clarity and smoothness. Most of the time im running AFMF2 at 90FPS (180fps with AFMF2) at 1080p. My system is super quiet at 30% system usage and i can game the entire day without worrying about my power bill.
But, imagine rendering at 1080p and then upscaling to 1440p, with FSR it beats anti aliasing in some games, and with dlss likely all of them. For it's also really good for lightweight competitive games where someone's head can be a few pixels where as on 1440p they're way easier to see. 1080p is awesome but don't sleep on 1440p because it's "more premium"
again, not about resolution haha
@@AncientGameplays good point 🤣
It's funny how many UA-camrs have to explain this over and over. But people still make those uneducated comments. To me, this is common sense. And simply wanting something to be true doesn't make it so.
WOW people still didnt figure out how to look at benchmarks?
Seems so
@@AncientGameplays im sorry to hear that!❤
Next thing this new Cult will argue it is "Why reviewers aretesting these games in very CPU intensiv areas??!!!". and they will be crybabies because they only play 1-2 minutes in that area, and is not relevant, etc. Th problem wll not stop with them, no matter how much explanation you will have. But yea, congratulation to you, having this much patience.
True haha
See the Hot hardware interview with Robert Halleck? He make comments on this subject.
I did not see, but guess I need to
@@AncientGameplays Worth a watch, gives a good indication on the time scale of incoming fixes for the disaster Arrow Lake launch.
In theory it also depends on how efficient the drivers use the particular CPU.
Lets say the driver is very inefficient with the CPU, then the FPS drops.
That does however not mean the CPU is weak, it just gets misused.
To be precise we should also monitor how much cpu time the driver software consumes, compared to some benchmark software.
Pretty hard to do as the OS also plays a role in this.
Yes, drivers overhead also matter, but I was talking regarding same assets of course
Most people who watch these videos completely understand the need for testing at 1080p, but i think what most people also want is the higher resolutions to actually show how it changes from CPU bottleneck to GPU bottleneck. I personally would like to see how my 7800X3D holds up against the 9800X3D at 1440p to see only a 1-2 fps difference. It would be nice to see the proof that there is zero reason to upgrade. And it would be even better to see the comparisons using DLSS and FSR because the theory would say uspscaling from 1080p to 1440p should show the same (or similar) results as a native 1080p comparison... but practically speaking.... will it really show those big difference as it does in native 1080p?. these are the things we would like to see.
So please, do...
1080p native
1440p native
4K native
1440p quality/balanced upscaled
4K balanced/performance upscaled
I am here explaining though, the same way I did on my CPU comparison. Not my fault people skip parts and don't absorve useful data
Exactly... I can't believe that literally none of these guys are doing that. It's literally the exact same content on every single channel. 🙄 Yet, comment sections are FULL of people wanting to see 1440p tests and so on.
You can check 1440p/2160p tests like that on Techpowerup. Yes, once you become GPU bottlenecked the differences disappear and it is no longer a CPU test. If you have a 7800X3D you really shouldn't upgrade unless you play competitive games and you really want to max out your 240-480FPS screen... at 1080p... I just went from an i7 2600k to a R7-7700 not because I can't afford a 7800/9800x3d but because I'm not going to pay 2.5x the money to get +30FPS at 1080p and +10FPS in 1440p once I get my 5070/5080... I'll get a 10800/11800x3d once they have 16 core variants on a single CCD
Soon UA-cam will be full of those kinds of benchmarks so no worries man, then you will see that you are correct sir, so if you play at 1440p/2160p stick with your 7800X3D and max your GPU if you want more frames, not your CPU. If you have money to spend then wait for the 5080/5090 and don't waste here. Even if you play at 1080p who cares? from 500 frames you will fall to 400-450? damn that's a shame, the game will be unplayable for sure for those who had a 144-175Hz monitor.
Other point is and I'm sure you know it, almost all of the test you watch on YT, doesn't matter if the channel is Pro AMD or Pro Intel, those will always show a better result than the one you are going after, so keep that in mind.
Tienes razón toda la razón y como dices ese target de fps dependerá de varios factores. La resolución y calidad gráfica deseada es importante como así el juego ya que depende de que tan optimizados estén.
Por ejemplo The last of Us parte 1 es muy dependiente de la cpu aún a 1440p por lo que un procesador más rápido hace la diferencia en este caso.
Pero nuevamente comparto totalmente lo que dices por eso al jugar en 1440p prefiero una mejor gpu y un cpu que pueda exprimir cerca del 99% a mi gpu.
Además si activamos el trazado de rayos también estresan a la cpu por lo que como dices depende de muchos factores.
The problem is that this is not how most people who buy this CPU would use it. 1440p should also be included
No, watch the video and understand. The use doesn't matter. Because otherwise we would have to test every single scenario in the world to fulfill all needs.
@AncientGameplays why would you not want to include real world scenario... meaning 1440p + Best CPU + Best GPU...
@@FO0TMinecraftPVP That's a real world scenario for maybe 1% of gamers. You would need tons of more CPU/GPU/resolution combinations to answer the question "Do I benefit from this CPU?" for a meaningful amount of gamers. And it would be needless, because everyone can answer this question for himself much easier. Just start your favorite games and look at the GPU load. Is it above 95% most of the time? You don't need a new CPU. Is it below that? A new CPU can help you get more FPS if you want. The rest is just a matter of looking at you budget and a CPU review to get the best CPU for your money and use case.
@@FO0TMinecraftPVP videos are almost like books: if u didnt understand something, watch again and again and again.. so just..watch again, he explained it :)
@FO0TMinecraftPVP he said watch the video and understand 😂
It's really simple to understand CPU tests done at low resolutions. It shows the max FPS the CPU can provide as a baseline.
So even if you play at 4K you'd know that if you have GPU that can push the FPS really high. Then you know what the limit it would be hitting.
Resolution doesn't matter to a CPU because it is not rendering the image at 4K fps. That is what the GPU renders and outputs. The CPU only puts together the physics and information feeding the GPU that data and at how fast it can do so. For example hitting 178hz/fps max in a given game/engine or 150 in another and 400 in another.
The reason the resolution is low. Is to remove the GPU as the bottleneck because it matters when it comes to GPU.
I have a 5800X3D and a 4090. Playing Hogwarts at 4K, I know my CPU cannot push past a certain limit in feeding the 4090 enough FPS. The 4090 at times sit at 50-58% Utilization.
Now if I had a 9800X3D, I could potentially get the GPU to go full 99% Utilization and actually provide the full capability.
So in a CPU test seeing the 5800X3D only reaching 93fps in Hogwarts (example I don't recall the real number) and my 4090 at only 60% Utilization. It's pretty clear to see it's being CPU bottle-necked and if I had a 9800X3D and it could reach 135fps (again made up example number). Then I could actually get the 4090 at 4K reach perhaps 124fps. Reaching my monitors 120Hz limit and be fully utilized.
Hmm.. a 1440p/2140p benchmark video could prove all claims! Fabio, stop refusing, we're still demanding 2K/4k benchmarks!:D
Sarcasm? 1440p/2140p will be for GPU benchmarks
I can literally make benchmarks and explain why those claims are dumb haha, but hardware unboxed already did that too and people still ask dumb questions
@@AncientGameplays We love to ask dumb questions!
2K is 2048x1080.
4K is 4096x2160.
@@AncientGameplays And you're both stubborn bastards.
Some other reviewers seem to manage to test with all common resolutions just fine.
Ah the thumbnail face says it all 😂😂😂
Testing your Patience