Well 3060Ti is a lot more powerful. Perhaps everything else was in place and 3060Ti was indeed the missing link. Until you know the rest of his system , it's easy judge .
@@GameslordXY 3060ti ain't that much powerful, 30% more performance tops. I'd understand, if he got a 13600k and wanted to upgrade to a 3080ti/4070TiS, but 3060ti really is a bad decision
Crazy is that a 3060 is barely faster than a regular GTX 1080, and the GTX 1080 Ti is even faster than the 3060 lol You only benefit from DLSS with a 3060
@@me-df9re based on the bottleneck calculator site. for 1080p gaming 12400F and 4060 have a 26% bottle neck but in some test videos on youtube it doesn't look that much ~ 🤣🤣
I tested them about a month ago to see if going to a 7800xt from a 6600xt on a 5900x would be a serious bottle neck and the result just didn't add up so I just discounted them as rubbish but as it happens its academic as the games I play (Wow classic world of tanks, and ships) getting a 7800xt is massive over kill and just a waist of £500. so just going to save for zen 6 and upgrade then.
Fabio, I think you are right about the calculators being useless, but you have a misunderstanding in the percentages that they show. For example: A bottleneck percentage of 22% can be the same using different GPUs and the same CPU. Even if the GPU2 is faster than GPU2. It just shows the percentage of the FPS that is slower to their max potential performance. So a 3600x with 7800xt can have the same percentage of bottleneck like 3600x with 7900xtx even if the FPS is much faster using the 7900xtx.
You want to see nuts? Try the DCS:World bottleneck calculation on the pc-builds bottleneck calculator with a Ryzen 7 5800X3D and an RX7900XT. It says 43.3% CPU bottleneck. It claims that at 3440x1440 I would get on: low settings Maximums of 120 fps, Averages of 100 and lows of 80. medium settings Maximums of 95fps, Averages of 80 fps, and lows of 63 fps High settings Max fps of 75, average Fps of 63, and lows of 51 Ultra Settings Maximums of 47 fps, averages of 39 fps and lows of 31 fps. The reality, with my optimized settings (some very high, some high, some medium and some off, like a lot of post-processing.) and with me running into the 180 fps engine frame limit. On the Caucasus map (free map) in a combat scenario with the Su-25T (free aircraft) and with me dropping clusterbombs, aka the ultimate performance killer in DCS: P99 fps 304.6 fps, Median 179 fps, Average 177 fps, P5 (5%) 135.7, P1 (1% fps) 118.6 fps, P0.2 (0.2% lows) 106 fps and P0.1 (0.1% lows) 97.3. On the Marianas Map in a free flight (non-combat scenario) in the F-14, mostly low level. P99 fps 301 fps, median 152.2 fps, Average 151.8, P5 104 fps, P1 93.5, P0.2 79.7 and P0.1 of 66.2 fps. So in short in most cases with optimized settings my lows were beyond the highs of the medium settings and mostly above the highs of the low settings, as claimed by that bottleneck calculator. My highs were doubling and tripling the claims of the bottleneck calculator and that is on settings that ran the gamut from very high to medium. And even on the infamous Syria map on in the F-14 I was bouncing off the 180 fps limiter most of the time, and the F-14 is the biggest performance hog of all aircraft. That's not just off by a little, that is not off by a mile, that is off by the length of a maximum range Phoenix shot, aka 120 nautical miles.
Fabio chooses to mention different tech outlets that do better benchmarking than these "Bottleneck Calculators" and mentions his own channel last. Can't get more humble than that, that's why I love this channel and will always support its content! Lots of respect for Fabio.
He's the best. He knows what he's talking about and he isn't sponsored by Nvidia and AMD reading a script they give him like SOME reviewers. Edit: I will say he is sponsored by GVGMall, which he makes abundantly clear, and I used it because of him for Windows 11. That was around a year ago. It's good to go..... his code and the key.
Was going to hit you up on Twitter since you were missing for 3 days😂. Adding to the subject, you could have all settings set correctly and yet still have a bottleneck somewhere due to either a game or OS bug and you can't do much there until the bug is fixed.
Great vid! That first bottleneck calculator was a “tool” I’d used in the past before finding out it was bogus, great to see proof it’s definitely terrible!
Good thinking. I chose my GPU and CPU based on other important factors like, what I actually use them for, as well as energy efficiency and ability to cool, both of which contribute to longevity, and things like VRAM size and memory bandwidth that help with future proofing (within reasonable boundaries). So bottlenecking is the least of my problems.
Great video as always. An idea you might be interested in since you're already on the "fake things in pc world" streak videos. I recently seen more and more people that are afraid of bottlenecks when ordering parts, a video showing no combo is actually perfect in some scenarios (like on certain games and on certain settings) and there's always a "bottleneck" (cpu or gpu depending on the case) could be a nice idea (basically you pairing a few cpus with a few gpus and showing "bottlenecks" in game scenarios even when the combo is a really decent one to show nothing is actually perfect).
I had many people try and tell me my 7600 CPU I paired with my 7900XTX was a waste and a big bottleneck. Yes, it is...if I play at 1080p. One doesn't buy a 7900XTX to play at 1080p though. lol
@@Gofr5 îs it tho? Checking twitch streamer theyve been playing competitive games and for competitive many players think 360hz is becoming a standard and a must have
@@Gofr5 i dunno how niche it îs. There îs plenty of cs fortnite dota lol pubg etc players. I would even Argue that the majority of gamers are actually playing competitive games online and there's a niche of single players.
Mate bottleneck calculators are useless. The fact of the matter is they never accurate or even close to what they report. So people are making buying decisions on misinformation just because a fake calculator says so. I messed around with a few min/max setups and everytime it was wrong. @@me-df9re
When recommending/building a pc to clients, I always get bombarded with "I don't think this CPU is enough for my GPU". Now I have the perfect weapon (your video) to just shut them up and complete their builds lol
Wow, this actually makes sense so much sense. I have for the longest time thought that my rig was bottlenecked when its not.... Thx for de-bunking and clearing it up what it actually means and how to know if you actually got it!!! More of these type of vids is what we need!!!
Glad to hear the ram speed mentioned in the context of a CPU bottleneck. I heard a short hand idea once that 200 in frequency is equivalent to two in cast latency. Everything is specific use case and it really depends what settings and it what resolution with what hardware and everything can be manipulated to make anything else the bottleneck.
The PC Builds one is insane! Told me that with 7800X3D / 7900XTX is 32% @1080p, 25% @1440p and 14.5% @4k!!! So, my high-end parts from the same generation are incompatible out of the box!!! NUTS
@AncientGameplays Yo Fabioooooo!!! Do you have a 7900GRE? I Think Overclocking the Shyte out of it and the VRAM would make great content. An AIB card would be preferable, since you would likely reach higher sustainable clocks. It just seems to be gimped out of box.
@@stephanhart9941 noo, whne u go to calculate the bottleneck you need to choose between general task and GPU or Cpu intense tasks U probably chose general tasks
You are absolutely correct. As simple as playing two different games can make the difference. One game can choke the CPU and another is throttled by the GPU.
@@AncientGameplays bought a 5700x3d for 199 euros, adds 5-10fps in normal games and like 15-20 in CPU intensive games. I've also OC and UV the GPU for a 15% gain. I'm really satisfied right now.
Yup. I'm old enough to remember when the bottleneck was the northbridge. Now the bottleneck is either a Lamborghini on a motocross track, or a rock-crawler on the Autobahn...like one wasn't enough. I'm getting too old for this shizzz.
I’ve done some ram tests for older Zen CPU’s (ie; Ryzen 5-2600) when I lowered the speed and tuned the ram with tighter timings. I was able to achieve higher throughput at lower speeds and lower latency overall by significant margins. I used TCreate Classic DDR4-3200 running at 2933 CL14. Primaries, secondaries, and tertiary timings tightened and 1.4v to the sticks. Results were very impressive. So there are things you can do to increase performance on low budget rigs.
Missing by your sponsor GVGmal the voice from S.A. „Just do it“ - as always tnks for your effort and information you did here. Greetings from Austria 🇦🇹
Ill admid when i build my first pc 2 months ago i had no idea that these sotes arent good representations of real world since i always heard people talk about bottlenecks. It made sense to me so i just googled and found the calculators, thankfully it didnt make me do any stupid decision (7600X & 7700XT) i hope
There are a lot of people on foruns advising others about bottlenecks without even understanding what a bottleneck actually is, they see those rubbish calculators once and Dunning-Kruger does the rest
Very notorious especially on local facebook gaming groups, the moment people spit out but this calculator says I need a better gpu than your recomended CPU/GPU I just call them out and say that they are dumb and cannot be helped. It doesn't help that they are also almost always fanboys of nvidia/intel/amd zzz.
Or that you are AMD fanboy. Not saying you are, but isn't that a possibility? I noticed that AMD crowd doesn't care to take some things into consideration. Like Raytracing, regardless of how powerful their GPUs. 1 of the reasons might be I guess the AMD not having the alternative to DLSS3.5 Ray Reconstruction. They should have called it just Ray Reconstruction or DLSS Ray Reconstruction. Some think that it Works only on RTX4000 , which is fauls as it Works on 2000 and 3000 as well. Natively.
@@GameslordXY They don''t care to take about raytracing because they don't care about raytracing. With my 7900 xtx or 4090 i'd always go for 165+fps @ 1440p max settings / 4k100-120 fps over raytracing at 60 . Every time. Have you ever played something at max settings at 170-200 fps ? It's glorious.
@@GameslordXY why should a amd crowd take into consideration a feature made by nvidia into their gpu-s? And its not like rt is a life changer settings
Looks like it's more of a "Panic Calculator". You panic - you buy new stuff. I know i got a bottleneck (Ryzen 5 1600x on a X370 Board + RX 6600XT + 16 gb Trident Z 14CL B-Die) and need to upgrade my CPU, flash the BIOS, install Win 11 and run for 3 more years without concerns. But as @13loodLoost said: my real bottleneck is my wallet. PS: just checked the Board too. It has PCIe 3.0, so my bottleneck gets worse. RIP to my wallet.
put 3600 and u are GOOD! My 6700xt was paired with 1600af for a while, and it's not too bad combo as ppl would think...so 6600xt with 3600 is golden! Pci 4.0? u should not care for that also... Wallet should be fine!
As someone who just upgraded from a Ryzen 5 5600X to a Ryzen 7 5800X, and is also planning to do an upgrade from a 3060 ti to a 7800 XT (Already ordered) , I was a bit concerned when I saw the bottleneck test saying my 5800X would be too weak and I would get like 20-25% bottleneck. It even says my 5800x is too weak for 3060 ti as well. But thanks to this video I guess I can say that the calculator is wrong, so I appreciate this video.
@@AncientGameplays Ah ok I didn't know that and that sucks because if I did I would have then chosen the 5700x3d. However as I did see a decent improvement with the 5800X over the 5600X I am not that disappointed, as it is still a upgrade over my current setup. I am also sure that the 5800X will be able to handle the 7800 XT, atleast better than what the calculator said.
You're absolutely right. I'm using a 2680v4 with a 6800 and there is absolutely no bottleneck and I'm using all the games that have the absolute best graphics. Furthermore I've seen a bottleneck on a 5700x with a 3060 which shouldn't exist according to the calculator.
Cooling matters too. If you have a 13900k but with a 30$ amazon cooler the bottleneck due to throttling will be higher than if you had a 7900x with a 360 AIO
I was once gaming off an external hard drive and it worked before, so I didn't question it. I was gaming in my parent's pc and it was unplayable on games that seem like they should be easy to run, like overcooked. I just assumed it was lack of ram, or the 5300g I gave them... out of desperation, I tried plugging my hard drive into the motherboard usb3 ports (because the case's front ports were usb2) and it turns out, my bottleneck was the fact I was gaming off of 20 year old technology.
I use those old Mickey's beer wide mouth bottles. Hardly any neck at all. I was rendering an animation at 1080p...while simultaneously playing WarThunder at 1440p maxed out and hitting 200+ fps. Sometimes it got slightly jerky and dropped to around 150fps. (when it started a new animation frame)Temps in mid 60s. What more do you want? (I might try a 5950X) 3080 10g - 3700X - 4x8 Ballistix - B550M/ac - RM1000x - 8 Sickleflows @ 65cfm each - CM240L push/pull
Idk, I just recommend people by just asking them what CPU theyre buying new If R3/i3 - anything GPU performance number of 50 or 60 (eg. RTX3050 or RX7600 If R5/i5 - anything GPU performance number of 60 or 70 (eg. RTX4060 or RX7700) If R7/i7 - anything GPU performance number of 70 or 80 (eg. RTX4070 or RX7800) If R9/i9 - anything GPU performance number of 80 or 90 (eg. RTX4080 or RX7900) ...and the PSU is just simply asking them using 650W / 750W / 850W and 1000W
That's why I always recommend buying a very good Mobo+CPU+Ram and pairing it with a weaker GPU, because you can upgrade the GPU in the next generation quite cheaply if you sell your old one and gain a lot of FPS, while upgrading a CPU won't happen so often and is more expensive.
So, I can upgrade my 4790k to a 7800x3d and still keep my 1080ti would give me better performance. I would like to have less lag/stuttering in Fallout 4 and Skyrim SE maybe this combo would be better.
@@Savethepandabears Why? It was a suggestion by someone on Discord to do that had already done that last year. So, I can the PC Parts now and upgrade the GPU later.
@@Savethepandabears He said the fps had increased a constant 100fps and 70fps 1% low and that there was hardly any stuttering/lag in his game. He went on to say that he got a rtx 4080 in November last year and with a new 4k 120hz monitor now he can play Fallout 4 at 4k with 140+ fps and a 120 fps 1% low.
@AncientGameplays I hear you, but like i said, just an idea of what would work together is that most people do more research, but a lot of people will just follow the bottleneck calculator to heart. So it's good you made this video, but at the end of the day, it all depends on the games or software you are trying to run. For example, if you play cs having a much stronger cpu, it will help even when a less strong cpu won't bottleneck your gpu.
@@no5ense40195% they are useless they do give you an idea but you can still get a better idea by watching yt videos with certain build cpu gpu combs and see if theres bottleneck and depending on resolution,its just takes more time
No, but seriously. I hate all the magical online calculators that don't disclose the method and just spit out a random percentage. Sadly, you also aren't blameless here. Way too many percentage comparisons are thrown around, and luckily I'm too lazy to do all the maths, so just consider the example below. There are 3 graphics cards: Card A can deliver 30fps, Card B - 60fps, Card C - 100fps. Card C is our baseline. The simplified graph would look something like this: C |=|=|=| 100% B |=|=| | 60% A |=| | | 30% Would you say card B is only 30% faster than card A? No, it's twice as fast, so it's 100% faster (or 200% as fast). Is card A 30% slower than card B? No, it's half as fast, so it's 50% slower. Similarly, card C is ~66% faster than card B. Thus, for example, the comparison between 7900XTX and 7900XT at ~ 18:30 is slightly off. There is no 15% difference between those cards. There is a 22fps difference, which means (according to the graph) that 7900XTX is about 11% faster and 7900XT is about 13% slower.
Question: I admit I'm not experienced with the topic of bottlenecks. And I understand this might not be enough detailed information to give an accurate answer. But I use an i7-10700 CPU (8 cores-16 threads @ approximately 4.8 GHz), 32GB of RAM (2933 MHz), and an RTX 4070. Do you think there is any significant bottleneck going on with my PC? The i7-10700 is a really good CPU, but I've assumed that it's not exactly *optimal* for a GPU as powerful as the 4070. Just from my own experience, I honestly don't notice any performance issues in games. All games run very smoothly on my PC. In fact, it seems my PC often runs games smoother than other people who also have powerful PCs. So I guess you could say "Well, there's your answer. If your PC runs everything smoothly, then there is no significant bottleneck at all." And of course, it would be fair to tell me: "Stop being lazy. Install MSI Afterburner, look at the GPU utilization while playing games, and see for yourself." Anyways, sorry for the stupid question. I was just curious. Thanks for any feedback you can give.
@@AncientGameplays Thanks for the feedback, Fabio. I appreciate it. In any case, I don't notice any sort of performance issues when playing games, like I was saying. So that's the good thing. Thanks again!
@tGameplays PS: I just read an article written by a person that performed several benchmarks with specifically the i7-10700 paired with a 4070. And according to this article, it's a bottleneck of approximately 15.1 on average. And I'm losing about 30-45 fps on average in games. Damn, that's actually pretty significant. But the good thing is that I use a 60Hz monitor, so my max framerate is always 60fps anyways. And the 4070 is so powerful that even with that reduction of frames due to the bottleneck, the game never drops below 60fps. So I'm not being impacted by the bottleneck. So that's the 'bright side' of this situation, at least. Anyways, sorry for the wall of text. Thanks again for the feedback man. Have a good one!
Bottleneck calculators are good for a Basic understanding. You need to do your own research afterwards ofc. You cannot trust the Result of a single Website. You should always do your own research afterwards.
i dont care if my 5800x will "bottlenecking" the future gpu i will grab but i will move to 4k@144Hz instead of 1080p@144Hz upgrading just my monitor and my GPU (currently on 6750xt). Since i dont have money to spend on a new PC i prefer to put my money on a future rdna5 or 6, or maybe even rtx60 or intel arc 3rd gen. it has to be an 80-class card with at least 4090 raster performance. All i know is that i play my fav games with video settings that ALWAYS makes the gpu "a bottleneck" so 99% utilization like it's designed to be: in 20 years of gaming i noticed that the bottleneck it's all placebo effect _(edit: if you choosed the right tier parts to balance ur build)._ Maybe in this or that specific tasks i need more cpu or gpu or ram or vram but they are quite rare. Enjoy ur PC guys stop looking at these things that melts ur brain :)
I just bought a "broken" rx 580 8gb for 15$ , the person who I bought said he can't start his pc, when I tested it, it worked but it was full of artifacts, I switched the bios from the gpu and boom everything works flawlessly 👍👍
A proper bottleneck calculator would be extremely complex and AT LEAST needs to consider some basic RAM information and ingame settings or at least an FPS target.
1:23 I would like to add that some games only bottleneck in specific areas, making the calculation even more difficult. EDIT: lol never mind you point this out at 5:47
There was a channel called 'Testing Games' on TH-cam that was banned a week ago. Now that they're banned a bunch of copycat channels have popped up using their name, a similar name, and logo. Maybe you should do a video about this.
I haven't built a computer since 2017 and haven't kept up with many released. So my knowledge was severely out of date a couple of months ago when I started planning another build. I was checking a lot of these bottleneck calculators and was just baffled by what CPUs and GPUs I was supposed to be pairing together. But several of the tools were giving me the same bottlenecks, so I just assumed they were right and my intuition was simply wrong or just out of date. Thanks for making a video like this and showing why the bottleneck calculators are often wrong.
They wouldn't be _completely_ useless, if they were just more accurate. If they were it could at least give a rough estimate for someone that knows absolutely NOTHING about PCs. Like you don't want to see anyone pair a 7800X3D or 13900k with a 3060. But the problem with that is anyone that knows nothing about PCs isn't going to be doing that kind of research in the first place and they'll just buy a prebuilt. Unfortunately prebuilts are often really lopsided too because they don't want to give you a competent GPU without making a fat margin on it. If someone could make one more accurate it might be a decent source to point people to that are completely alien to PC hardware. The REAL problem is these sites have absolutely _nothing_ to do with actually building a balanced system and everything to do with coercing you to spend more while using their affiliate links. It's not a calculator, it's an AD. And I wouldn't be even slightly surprised to find out they're probably owned by a seller. I doubt they've done even a single benchmark, probably just punch random stats into a deeply flawed calculator.
3060 ? 12GB model? Why? It can serve as Nice placeholder for those with relatively limited budget until a gamer can afford 7900GRE(finaly availible outside of China) or RTX4070Ti Super or something. Getting something more balanced can mean MUCH sooner upgrade of both CPU and GPU. With DLSS 3060 12GB can cover most players needs especially when serving as a placeholder for something much better.
@@GameslordXY If you have to cherry pick a placeholder you're missing the point entirely because you _still_ have to spend more down the road. Ironic you claim a balanced build will require an upgrade sooner, when using a placeholder means you still need an upgrade RIGHT NOW. Besides, if you do that staggered buildout you don't use a $300 GPU as a placeholder, you use a $100 CPU. Unless you only play @1080p low an entry level GPU will _always_ choke a CPU, but with real world settings a low end CPU won't be nearly as restrictive to a decent GPU. Besides, you're choosing to buy 2 of the most expensive component of the system making your overall costs much higher. If you spend all your time waiting for the "bigger better thing" you'll always be waiting. The Gimped Rabbit Edition is a terrible example as it isn't even a price to performance upgrade from the 7800xt you could have gotten long ago. May as well wait until RDNA4 later this year that WILL give you better value than buying that GPU AMD dug out of the trash.
It is very interesting how the "average PC builder" feelings on bottlenecking kind of went full circle. At the Core 2 era, most decently new PCs had strong enough CPUs to match or exceed 60fps on console games and most players were satisfied with that experience, so it didn't matter that much. The general knowledge was that most CPUs were fine for gaming, all you needed was a strong GPU. Hell I even remember some channels from my region testing 980Tis with Pentium G4560s, "proving" that CPU "didn't even matter" at 4K. I guess close to the 2015 era and after, high refresh gaming became more and more popular and people became more and more aware of the need for more CPU performance to reach high frame-rates. Yet, the general knowledge was still that a mid-range CPU from a couple years ago was enough to run most games at high resolutions at decent FPS. After all, how could your CPU bottleneck your system if it isn't at 100% usage? nonsense. I don't know exactly when it changed, maybe it was when current-gen consoles became a thing, but it felt like overnight a lot of people kept throwing "CPU bottleneck" at anyone that had any performance issue in games. Either you had the highest-end CPU possible or you had a CPU bottleneck. I feel like after one or two years of constant "CPU bottleneck" discussions, the calculators came around. And they never ever had any sort of basis in reality. I feel like people had mixed feelings that every system they put in the calculator returned a CPU bottleneck. It both proved them right that all systems were apparently CPU bottlenecked and also pissed them off because their system was also bottlenecking. That's when the discussion changed again and it feels like people are consciously rejecting the thought of CPU bottlenecking. Like they accepted it as an inevitability and are aggressively fine with it and very hostile to anyone that dares to point out a bottlenecking scenario. Specially if performance is already high or if CPU usage is very low. It is a complex topic. Most people simply lack the knowledge about game rendering to understand how each part of the system can affect the whole. People want a "blanket" term: If X then Y, to fix A improve B, and hardware is almost never that simple. Every piece of software utilizes hardware differently and every single system will have a bottleneck in some part of the process, if not, we would have infinite FPS.
Before watching past the first minute mark, I googled "bottleneck calculator", entered my hardware, selected "graphics card intense tasks" (i.e., gaming) and got: "This configuration has 0.0% of bottleneck . Everything less than 5% should not be concerned major bottleneck." I say these calculators work perfectly fine. For me. :)
When you pointed out the "Game calculator" existed, I tried that one with Cyberpunk 2077 - 1.9%. Still well below that 5% threshold, and I'm still happy.
When your Cyberpunk test starts, your CPU has 30% less usage than your GPU, so it seems the calculator got the right number, but has the metric backwards or something. They're interpreting their data incorrectly.
I mean in the introduction he is right to point out being cautions with these calculators But it's funny the first calculation when he opens the recording (assuming it's 1080p) indeed shows 100% CPU utilization and only 67% utilization on GPU which is not far from what the page said 29.2% vs 33% he actually sees. In 1080p you really need lot of MHZ on the CPU (probably big L1/L2 cache, possibly fast RAM CL) so CPU can request GPU generated frames more often.
Easy back of the hand bottleneck calc: Compare CPU reviews under a 4090. Take the max fps at your desired res, then seek out GPU reviews done under a 14900K or 7800X3D, observe which one lines up with the first fps target. (you can also do this in reverse. take your chosen GPU fps numbers and look up the CPU that achieves them) Bam, balanced build. For the most part, any GPU priced for normies won't be held back by any CPU down to a 5600/10400. So just aim for a graphics bottleneck and be happy with a build for a few years. :)
Yup. This is accurate. When I built my PC I searched which CPU can give me 60 FPS in the most demanding games I would play. It was the 5700x. I looked up which was the GPU that can give me 60 FPS in 2K in the most demanding game I would play. It was the RX 6800. So. I got the CPU, and 2 years later, I just got the RX 7700XT, since it's the same thing but with AV1 encoder (which benefits me). The result is basically a console that can play in 2K, but with RGB and Windows. So I can also work from it. My needs were pretty vanilla though. There's people who play Counterstrike in 640x480. So I see them use 7800x3D with a RTX4060 so that they can get 550FPS. They would play in FHD, but you would need a 4090 for FHD Ultra to pump up those FPS. Or. There's some people who want to play in 8K30FPS. So they get an i3-10100 with a 4090. The spectrum is large ... My rule of thumb is ... just don't pair a 6 year old CPU with a current GPU, or viceversa ... Unless the old one is a really high-end product.
If i remember correctly, r5 2400g and i5 7400 are similar from what i could see on yt but intel one was like 10-20% weaker, more bottlenecking, forgot which gpu i paired with on that site
I have checked some “ Bottleneck Calculators “ online and according to them my 9700K is a 30% bottleneck for RX 7800XT. I am playing all my games on 1440p and have yet to find one that is bottlenecked by my CPU. I always have my AMD overlay while playing those games and CPU is never used 100% and GPU is working fully depending on a game . My RAM is DDR4 3600Mhz 16GB
I'm getting 20.1% on one page, 26.76 on another and 24.5 on a third one that seemed more accurate and showed differences in performance in various games with things like expected fps based on different resolutions and it said it would be playable or not... One of the tabs said a game in 4k would run at 59.6 fps so it was "not playable" lol what the fuck. Seems like bottleneck calculators are a snake oil kind of thing to make more people buy the higher end and expensive processors that they don't actually need for gaming. 12100F & 6750XT in case you wondered.
Ok, so what they mean in the 2077 example is that the cpu isnt being used to its full potential. If i look at the cpu usage in the video they are actually really close with the cpu saying 70 something %! No hate but just wanting to clear things up. It also probably says that when you scroll down, you can see the percentages utilised and if it is a cpu or gpu bottleneck. And yes, it does say some bullsh!t sometimes 😅
Thanks for the video, I was shocked when I used the calculator, so I'm at ease now. Quick question what is the tool you use to measure the GPU and the CPU? Thanks in advance. 🙂 Btw, Subbed your channel, nice to see some honest peeps on the net 😁
I think there is a misunderstanding of how this calculator is working. If you play around with it, you will notice that with the same GPU/CPU combo the lower the resolution you go the higher it will say the bottle neck is and the opposite is true of you change it to higher resolutions. I believe the calculator is saying that the more frames your GPU is putting out (which it would at lower resolution) the CPU will not be able to keep up to the shear volume of frames the GPU is able to produce i.e. bottlenecking. Once you slap a better CPU in the calculator, boom bottleneck goes down because that CPU can handle the volume of frames better. This is also supported by the additional information provided under the test result, but I could be wrong. just how I was understanding what I was reading. Now as for the accuracy of the calculator, who knows...
Techically you can download swap files which are slow alternatives to ram that are stored on your ssd when you go over your ram limits. Somewhat similar to resizable bar.
Bottleneck and futureproofing are some of the buzzwords that keep coming up from the masses that should be eradicated to be honest. They just make newbies spend more than they need to with problems that mostly doesn't affect them
Yeah, bottleneck calculators make for some interesting people trying to argue points that aren't entirely accurate. Most recent was someone telling me that my CPU was creating a bottle neck for my GPU and I should be on a 7900X3D. Had to explain why, even if it was creating a bottleneck, the 7800X3D would have been a better pairing given the 8core 3D chip vs a 6core+6core3D cache. Better to know what and how a bottleneck is created and build your system around those expectations.
I got a 6700xt used last month. Put it in my rig that has a 12400, 32gb of 3600mhz ram. I have a 165hz 1080p monitor and there was Apperently supposed to be a bottle neck. I honestly haven't seen it yet(mind you I have it undervolted cause it's summer and if I can get away with keeping my room a little cooler I will). Most demanding games I played were control and guardians of the galaxy and at that point my gpu was pushing out 60 fps on both with the highest settings with frame gen too cause I like single player games looking pretty and my gpu was pegged at 98% the entire time. Same for overwatch which is a lighter game but I still get almost max utilisation on my gpu.
I just checked on the bottleneck site for PC-builds. I have a 7800X3D and a 4070 Super. Here is the result: The processor might become a bottleneck for the graphics card's performance. While the graphics card is capable of handling intensive graphical workloads, the processor's processing power may not be sufficient to fully utilize the graphics card's potential. This imbalance could restrict the system's overall performance, leading to slower processing and potentially reduced graphics quality. To achieve a more balanced setup, it would be advisable to consider upgrading to a more powerful processor that can keep up with the demands of modern applications and games.
I'm not sure just using benchmark is telling you the whole story. I'm getting a bottlenecked using Ryzen 7600 (UV, PBO, offset to all core 5,3ghz) + 4080 super on 3440x1440 ray tracing ultra + DLSS quality without FG. In busy areas the GPU drops down to 68-74% usage. So in my specific scenario the calculator is on point showing around 30% bottleneck.
Of course your are, you're running 1440P ultrawide dlss quality, which means you are running 960P ultrawide lol On top of that you most likely have a poor expo ram profile
What I say to friends is that, bottleneck is not important at all, what is important is to have a decent or strong base to build on top. The PSU should provide enough power to enable you to make upgrades in the future without the need to upgrade the PSU itself. The Motherboard should be decent enough to give you all the features you need to suit your needs without breaking the bank, a cheap mobo would lack in features and be poorly constructed, while a very expansive mobo doesn't justify it's high price for the amount of features it provides compared to decent cheaper options having most of the same features. And the CPU should have at least 6 cores for gaming, 8 cores for "future proofing" (if that exists) and more than that for a workstation (provided the work in question will need/benefit from the extra cores) With the holy trifecta settled, you are much more free to choose your hardware now, or upgrades in the future.
i think for the reason that the 7700X has a more higher boost clock speed and it is more flexible because it can be overclocked. the 7800X3D is only faster in games because of its 3D Cache and there is no room for overclocking. i think the results are actually good because in terms of single core and multi core performance the 7700X is a bit faster.
Hello Fabio. 100 percent load means nothing, as does maximum consumption. Because for clarity, in the place where you have a load of 7700x -100 percent. Try installing 7800x3d and see how it is distributed. The load gauge often shows complete nonsense. The chips are apparently very complex things and the sensors cannot show the full load because they are apparently not able to take parameters from all computing units. I roughly compare 2 parameters of the video card 1) Consumption 2) load sensor, this will only show the approximate load. PS Oh, Fabio, I learned from you that there is such a bottleneck calculator. 🤣🤣🤣🤣🤣🤣
I know and I explaines that in the video. Consumption and load together are the key, bjt they also change according to the resolution, clocks, etc. You need to know your shit to be sure haha
My bottleneck is my wallet.
🤣🤣🤌
That's usually the biggest bottle neck of almost
/Threaded
aren't we all
The bottlenecks of bottlenecks🤣🤣🤣💀
The biggest bottleneck in a computer sits in front of the desk.
😂👌
Cap the cpu matters more
Got a 3060 last year from a guy for $100 who used a bottleneck calculator and was upgrading to 3060ti to get rid of his "bottleneck"
Looooooool, that's crazy xD
Well 3060Ti is a lot more powerful.
Perhaps everything else was in place and 3060Ti was indeed the missing link.
Until you know the rest of his system , it's easy judge .
@@GameslordXY 3060ti ain't that much powerful, 30% more performance tops. I'd understand, if he got a 13600k and wanted to upgrade to a 3080ti/4070TiS, but 3060ti really is a bad decision
Crazy is that a 3060 is barely faster than a regular GTX 1080, and the GTX 1080 Ti is even faster than the 3060 lol
You only benefit from DLSS with a 3060
well, it wasnt a guy, it was a gen Z child
Bottleneck calculators are close to the new snake oil, at least according to what I tested of course!
Their algorithms has probably not been updated since 9900K released.
@@me-df9re based on the bottleneck calculator site.
for 1080p gaming 12400F and 4060 have a 26% bottle neck
but in some test videos on youtube it doesn't look that much ~
🤣🤣
I tested them about a month ago to see if going to a 7800xt from a 6600xt on a 5900x would be a serious bottle neck and the result just didn't add up so I just discounted them as rubbish but as it happens its academic as the games I play (Wow classic world of tanks, and ships) getting a 7800xt is massive over kill and just a waist of £500. so just going to save for zen 6 and upgrade then.
Fabio, I think you are right about the calculators being useless, but you have a misunderstanding in the percentages that they show. For example: A bottleneck percentage of 22% can be the same using different GPUs and the same CPU. Even if the GPU2 is faster than GPU2. It just shows the percentage of the FPS that is slower to their max potential performance.
So a 3600x with 7800xt can have the same percentage of bottleneck like 3600x with 7900xtx even if the FPS is much faster using the 7900xtx.
You want to see nuts?
Try the DCS:World bottleneck calculation on the pc-builds bottleneck calculator with a Ryzen 7 5800X3D and an RX7900XT.
It says 43.3% CPU bottleneck.
It claims that at 3440x1440 I would get on:
low settings
Maximums of 120 fps, Averages of 100 and lows of 80.
medium settings
Maximums of 95fps, Averages of 80 fps, and lows of 63 fps
High settings
Max fps of 75, average Fps of 63, and lows of 51
Ultra Settings
Maximums of 47 fps, averages of 39 fps and lows of 31 fps.
The reality, with my optimized settings (some very high, some high, some medium and some off, like a lot of post-processing.) and with me running into the 180 fps engine frame limit.
On the Caucasus map (free map) in a combat scenario with the Su-25T (free aircraft) and with me dropping clusterbombs, aka the ultimate performance killer in DCS:
P99 fps 304.6 fps, Median 179 fps, Average 177 fps, P5 (5%) 135.7, P1 (1% fps) 118.6 fps, P0.2 (0.2% lows) 106 fps and P0.1 (0.1% lows) 97.3.
On the Marianas Map in a free flight (non-combat scenario) in the F-14, mostly low level.
P99 fps 301 fps, median 152.2 fps, Average 151.8, P5 104 fps, P1 93.5, P0.2 79.7 and P0.1 of 66.2 fps.
So in short in most cases with optimized settings my lows were beyond the highs of the medium settings and mostly above the highs of the low settings, as claimed by that bottleneck calculator.
My highs were doubling and tripling the claims of the bottleneck calculator and that is on settings that ran the gamut from very high to medium.
And even on the infamous Syria map on in the F-14 I was bouncing off the 180 fps limiter most of the time, and the F-14 is the biggest performance hog of all aircraft.
That's not just off by a little, that is not off by a mile, that is off by the length of a maximum range Phoenix shot, aka 120 nautical miles.
Fabio chooses to mention different tech outlets that do better benchmarking than these "Bottleneck Calculators" and mentions his own channel last. Can't get more humble than that, that's why I love this channel and will always support its content! Lots of respect for Fabio.
He's the best. He knows what he's talking about and he isn't sponsored by Nvidia and AMD reading a script they give him like SOME reviewers.
Edit: I will say he is sponsored by GVGMall, which he makes abundantly clear, and I used it because of him for Windows 11. That was around a year ago. It's good to go..... his code and the key.
You're doing gods work.
I'm going to send this video to any noobie who falls victim to one of these sites 😅
@@me-df9replease explain how to use it...
Was going to hit you up on Twitter since you were missing for 3 days😂.
Adding to the subject, you could have all settings set correctly and yet still have a bottleneck somewhere due to either a game or OS bug and you can't do much there until the bug is fixed.
its a gift that keeps on giving haha. Tomorrow we have the Adrenalin drivers review :D
Don’t twomad us
Can u tell me if i5 4570 is good with rx 5700xt with 8gb ram😢
@@AbdulRehman-nf1we that cpu is too old mate, will work, but not great
@@AncientGameplays how bout an rx 6500 xt or 1660 super or 1070 ti
Your biggest bottleneck will be the software you're running.
The title suggested I should use a bottleneck calculator to increase my RAM performance.
I always had my doubts about those calculators and you've gone ahead and confirmed them, great content as always brother.
I'm glad you did this alongside the UserBenchmark video. I'm sure those guys unsubscribed & disliked the video. Shots Fired!🤣
So if its like downloading RAM is good!! Last week i downloaded 32GB just fine! 😎
That’s insane, I’ve been having issues with my ram so maybe I should download some too, do you know any good places to download some? 😊
FINALLY SOMEONE CALLING OUT THE B.S!!
Great vid! That first bottleneck calculator was a “tool” I’d used in the past before finding out it was bogus, great to see proof it’s definitely terrible!
Good thinking. I chose my GPU and CPU based on other important factors like, what I actually use them for, as well as energy efficiency and ability to cool, both of which contribute to longevity, and things like VRAM size and memory bandwidth that help with future proofing (within reasonable boundaries). So bottlenecking is the least of my problems.
Great video as always.
An idea you might be interested in since you're already on the "fake things in pc world" streak videos. I recently seen more and more people that are afraid of bottlenecks when ordering parts, a video showing no combo is actually perfect in some scenarios (like on certain games and on certain settings) and there's always a "bottleneck" (cpu or gpu depending on the case) could be a nice idea (basically you pairing a few cpus with a few gpus and showing "bottlenecks" in game scenarios even when the combo is a really decent one to show nothing is actually perfect).
I had many people try and tell me my 7600 CPU I paired with my 7900XTX was a waste and a big bottleneck. Yes, it is...if I play at 1080p. One doesn't buy a 7900XTX to play at 1080p though. lol
@@Gofr5people buy high priced Nvidia for 1080p 360hz monitors competitive games tho.
@pR0ManiacS True, but that's more of a niche case.
@@Gofr5 îs it tho? Checking twitch streamer theyve been playing competitive games and for competitive many players think 360hz is becoming a standard and a must have
@@Gofr5 i dunno how niche it îs. There îs plenty of cs fortnite dota lol pubg etc players. I would even Argue that the majority of gamers are actually playing competitive games online and there's a niche of single players.
So what’s the best way to keep from bottlenecking if the calculators don’t work I kinda feel like the marketing is a set up for failure
Learn to PC.
This is a world class video! There so many people on TikTok quoting these cpu bottleneck calculators and advising people the wrong advice!
are u sharing a braincell with ur pillow?@@me-df9re
Mate bottleneck calculators are useless. The fact of the matter is they never accurate or even close to what they report. So people are making buying decisions on misinformation just because a fake calculator says so. I messed around with a few min/max setups and everytime it was wrong. @@me-df9re
@@TechTusiast there are a lot of smooth brained people
When recommending/building a pc to clients, I always get bombarded with "I don't think this CPU is enough for my GPU". Now I have the perfect weapon (your video) to just shut them up and complete their builds lol
Wow, this actually makes sense so much sense. I have for the longest time thought that my rig was bottlenecked when its not.... Thx for de-bunking and clearing it up what it actually means and how to know if you actually got it!!! More of these type of vids is what we need!!!
❤️❤️
My son uses the 7700X CPU with his 6950 XT. Seems to be plenty fast enough, at least for an 11 year old.
Ofc it is
excellent content with valid points and solid logic, thanks! I'm glad you mentioned the Portuguese part, I was wondering about the accent. Cheers.
Thanks as well
Man GTA 5 has a engine limit of 188fps. And also if you hit engine limit it stutters like crazy.
Watch your tach...
7800X3D is so powerful that it broke that limit.
Glad to hear the ram speed mentioned in the context of a CPU bottleneck. I heard a short hand idea once that 200 in frequency is equivalent to two in cast latency. Everything is specific use case and it really depends what settings and it what resolution with what hardware and everything can be manipulated to make anything else the bottleneck.
The PC Builds one is insane! Told me that with 7800X3D / 7900XTX is 32% @1080p, 25% @1440p and 14.5% @4k!!! So, my high-end parts from the same generation are incompatible out of the box!!! NUTS
Did u use the graphics intensive settings
What? Hahhaha, big lol
@AncientGameplays Yo Fabioooooo!!! Do you have a 7900GRE? I Think Overclocking the Shyte out of it and the VRAM would make great content. An AIB card would be preferable, since you would likely reach higher sustainable clocks. It just seems to be gimped out of box.
@ShayNoMore1 it was just the general calculations. No specific game.
@@stephanhart9941 noo, whne u go to calculate the bottleneck you need to choose between general task and GPU or Cpu intense tasks
U probably chose general tasks
You are absolutely correct. As simple as playing two different games can make the difference. One game can choke the CPU and another is throttled by the GPU.
The new non-paywall Userbenchmark lol.
6:50 Good work with the reference card there, really good it tuned well.
I have a 5600x and a 7900 xt, should I jump to a 5800x3d/5700x3d? Playing at 1440p UW
it would help in some games for sure, the 5600x will not be enough in some parts of recent games even at 1440P UW, especially if you use FSR
@@AncientGameplays bought a 5700x3d for 199 euros, adds 5-10fps in normal games and like 15-20 in CPU intensive games. I've also OC and UV the GPU for a 15% gain. I'm really satisfied right now.
Yup.
I'm old enough to remember when the bottleneck was the northbridge.
Now the bottleneck is either a Lamborghini on a motocross track, or a rock-crawler on the Autobahn...like one wasn't enough.
I'm getting too old for this shizzz.
Bro sounds like:" Im gonna break my Monitor dude, I swear " with deeper voice.
I’ve done some ram tests for older Zen CPU’s (ie; Ryzen 5-2600) when I lowered the speed and tuned the ram with tighter timings. I was able to achieve higher throughput at lower speeds and lower latency overall by significant margins. I used TCreate Classic DDR4-3200 running at 2933 CL14. Primaries, secondaries, and tertiary timings tightened and 1.4v to the sticks. Results were very impressive. So there are things you can do to increase performance on low budget rigs.
Missing by your sponsor GVGmal the voice from S.A. „Just do it“ - as always tnks for your effort and information you did here. Greetings from Austria 🇦🇹
Great video, i really had some doubs about that website. Now you made that clear. Keep up the good work 👍 Thank you!
Ill admid when i build my first pc 2 months ago i had no idea that these sotes arent good representations of real world since i always heard people talk about bottlenecks. It made sense to me so i just googled and found the calculators, thankfully it didnt make me do any stupid decision (7600X & 7700XT) i hope
There are a lot of people on foruns advising others about bottlenecks without even understanding what a bottleneck actually is, they see those rubbish calculators once and Dunning-Kruger does the rest
Very notorious especially on local facebook gaming groups, the moment people spit out but this calculator says I need a better gpu than your recomended CPU/GPU I just call them out and say that they are dumb and cannot be helped. It doesn't help that they are also almost always fanboys of nvidia/intel/amd zzz.
show them this video then haha
Or that you are AMD fanboy.
Not saying you are, but isn't that a possibility?
I noticed that AMD crowd doesn't care to take some things into consideration.
Like Raytracing, regardless of how powerful their GPUs.
1 of the reasons might be I guess the AMD not having the alternative to DLSS3.5 Ray Reconstruction.
They should have called it just Ray Reconstruction or DLSS Ray Reconstruction.
Some think that it Works only on RTX4000 , which is fauls as it Works on 2000 and 3000 as well.
Natively.
@GameslordXY What about frame generation?
@@GameslordXY They don''t care to take about raytracing because they don't care about raytracing. With my 7900 xtx or 4090 i'd always go for 165+fps @ 1440p max settings / 4k100-120 fps over raytracing at 60 . Every time. Have you ever played something at max settings at 170-200 fps ? It's glorious.
@@GameslordXY why should a amd crowd take into consideration a feature made by nvidia into their gpu-s? And its not like rt is a life changer settings
Looks like it's more of a "Panic Calculator". You panic - you buy new stuff.
I know i got a bottleneck (Ryzen 5 1600x on a X370 Board + RX 6600XT + 16 gb Trident Z 14CL B-Die) and need to upgrade my CPU, flash the BIOS, install Win 11 and run for 3 more years without concerns.
But as @13loodLoost said: my real bottleneck is my wallet.
PS: just checked the Board too. It has PCIe 3.0, so my bottleneck gets worse. RIP to my wallet.
put 3600 and u are GOOD! My 6700xt was paired with 1600af for a while, and it's not too bad combo as ppl would think...so 6600xt with 3600 is golden! Pci 4.0? u should not care for that also... Wallet should be fine!
pcie 3.0 is only a bottleneck if you have an x4 or x8 gpu, but a 1080 ti (roughly 6600xt perf) has no issues
@damasterpiece08 yeah, but I have to think about the future. Buying now a PCIe 3.0 is a dead end.
@@Orimbar youre not buying a new mb, just cpu and gpu like you said. Maybe look up how a 6800xt does on pci3
@damasterpiece08 I my case I would just upgrade the cpu to a ryzen 5 5700 or 5700x and flash the BIOS. For gaming on mid settings, it's just fine.
me: build the top setup
calculator: you suck! © Shao Khan
As someone who just upgraded from a Ryzen 5 5600X to a Ryzen 7 5800X, and is also planning to do an upgrade from a 3060 ti to a 7800 XT (Already ordered) , I was a bit concerned when I saw the bottleneck test saying my 5800X would be too weak and I would get like 20-25% bottleneck. It even says my 5800x is too weak for 3060 ti as well. But thanks to this video I guess I can say that the calculator is wrong, so I appreciate this video.
Thsnk you as well, but the 5800x was a bad choice, you should have went with the 5700x3d
@@AncientGameplays Ah ok I didn't know that and that sucks because if I did I would have then chosen the 5700x3d. However as I did see a decent improvement with the 5800X over the 5600X I am not that disappointed, as it is still a upgrade over my current setup. I am also sure that the 5800X will be able to handle the 7800 XT, atleast better than what the calculator said.
You're absolutely right. I'm using a 2680v4 with a 6800 and there is absolutely no bottleneck and I'm using all the games that have the absolute best graphics. Furthermore I've seen a bottleneck on a 5700x with a 3060 which shouldn't exist according to the calculator.
Cooling matters too.
If you have a 13900k but with a 30$ amazon cooler the bottleneck due to throttling will be higher than if you had a 7900x with a 360 AIO
I was once gaming off an external hard drive and it worked before, so I didn't question it. I was gaming in my parent's pc and it was unplayable on games that seem like they should be easy to run, like overcooked. I just assumed it was lack of ram, or the 5300g I gave them...
out of desperation, I tried plugging my hard drive into the motherboard usb3 ports (because the case's front ports were usb2) and it turns out, my bottleneck was the fact I was gaming off of 20 year old technology.
I use those old Mickey's beer wide mouth bottles. Hardly any neck at all.
I was rendering an animation at 1080p...while simultaneously playing WarThunder at 1440p maxed out and hitting 200+ fps.
Sometimes it got slightly jerky and dropped to around 150fps. (when it started a new animation frame)Temps in mid 60s. What more do you want? (I might try a 5950X)
3080 10g - 3700X - 4x8 Ballistix - B550M/ac - RM1000x - 8 Sickleflows @ 65cfm each - CM240L push/pull
😂😂
Good to see you doing a bit different videos! Very helpful! Thanks
According to the vidéo 12.4% is a bottleneck ? 💀
Idk, I just recommend people by just asking them what CPU theyre buying new
If R3/i3 - anything GPU performance number of 50 or 60 (eg. RTX3050 or RX7600
If R5/i5 - anything GPU performance number of 60 or 70 (eg. RTX4060 or RX7700)
If R7/i7 - anything GPU performance number of 70 or 80 (eg. RTX4070 or RX7800)
If R9/i9 - anything GPU performance number of 80 or 90 (eg. RTX4080 or RX7900)
...and the PSU is just simply asking them using 650W / 750W / 850W and 1000W
Nothing gives real insight as much as actually testing the parts.
That's why i always say Don't look for calculations, look for benchmarks.
Thank you!!! I was already wondering wtf are all cpus now the Bottleneck...
I just downloaded more ram, and now my bottlebecks are gone
That's why I always recommend buying a very good Mobo+CPU+Ram and pairing it with a weaker GPU, because you can upgrade the GPU in the next generation quite cheaply if you sell your old one and gain a lot of FPS, while upgrading a CPU won't happen so often and is more expensive.
So, I can upgrade my 4790k to a 7800x3d and still keep my 1080ti would give me better performance. I would like to have less lag/stuttering in Fallout 4 and Skyrim SE maybe this combo would be better.
😅 don't use a 1080 ti on a 7800X3D
@@Savethepandabears Why? It was a suggestion by someone on Discord to do that had already done that last year. So, I can the PC Parts now and upgrade the GPU later.
@@jodiepalmer2404 ah I see
@@Savethepandabears He said the fps had increased a constant 100fps and 70fps 1% low and that there was hardly any stuttering/lag in his game. He went on to say that he got a rtx 4080 in November last year and with a new 4k 120hz monitor now he can play Fallout 4 at 4k with 140+ fps and a 120 fps 1% low.
No one thinks its 100% accurate but it just gives you an idea of what to fit together
If you watched the video you would understand they aren't even good for an idea in most cases
@AncientGameplays I hear you, but like i said, just an idea of what would work together is that most people do more research, but a lot of people will just follow the bottleneck calculator to heart. So it's good you made this video, but at the end of the day, it all depends on the games or software you are trying to run. For example, if you play cs having a much stronger cpu, it will help even when a less strong cpu won't bottleneck your gpu.
@@no5ense401 the thing is that their results arent trustable in 95% of cases man
@@no5ense40195% they are useless they do give you an idea but you can still get a better idea by watching yt videos with certain build cpu gpu combs and see if theres bottleneck and depending on resolution,its just takes more time
In Harrison Ford's voice: THAT'S NOT HOW THE PERCENTAGE WORKS!
No, but seriously. I hate all the magical online calculators that don't disclose the method and just spit out a random percentage. Sadly, you also aren't blameless here. Way too many percentage comparisons are thrown around, and luckily I'm too lazy to do all the maths, so just consider the example below.
There are 3 graphics cards:
Card A can deliver 30fps,
Card B - 60fps,
Card C - 100fps.
Card C is our baseline. The simplified graph would look something like this:
C |=|=|=| 100%
B |=|=| | 60%
A |=| | | 30%
Would you say card B is only 30% faster than card A? No, it's twice as fast, so it's 100% faster (or 200% as fast).
Is card A 30% slower than card B? No, it's half as fast, so it's 50% slower.
Similarly, card C is ~66% faster than card B.
Thus, for example, the comparison between 7900XTX and 7900XT at ~ 18:30 is slightly off. There is no 15% difference between those cards. There is a 22fps difference, which means (according to the graph) that 7900XTX is about 11% faster and 7900XT is about 13% slower.
the doctors said im a bottleneck to my life ...
If amd offer you a job as head of marketing would you take it. Because you are already doing an amazing job
I would take it just to try and make the brand better, as a challenge haha
very nice video I now know how to download ram and vram I also learned how to upload my clock speeds
Question: I admit I'm not experienced with the topic of bottlenecks. And I understand this might not be enough detailed information to give an accurate answer. But I use an i7-10700 CPU (8 cores-16 threads @ approximately 4.8 GHz), 32GB of RAM (2933 MHz), and an RTX 4070. Do you think there is any significant bottleneck going on with my PC?
The i7-10700 is a really good CPU, but I've assumed that it's not exactly *optimal* for a GPU as powerful as the 4070. Just from my own experience, I honestly don't notice any performance issues in games. All games run very smoothly on my PC. In fact, it seems my PC often runs games smoother than other people who also have powerful PCs. So I guess you could say "Well, there's your answer. If your PC runs everything smoothly, then there is no significant bottleneck at all."
And of course, it would be fair to tell me: "Stop being lazy. Install MSI Afterburner, look at the GPU utilization while playing games, and see for yourself." Anyways, sorry for the stupid question. I was just curious. Thanks for any feedback you can give.
Yes, the 10700k will be a bit slow for that card in some cases for sure, but the biggest bottleneck there is that crappy ram
@@AncientGameplays Thanks for the feedback, Fabio. I appreciate it. In any case, I don't notice any sort of performance issues when playing games, like I was saying. So that's the good thing. Thanks again!
@tGameplays PS: I just read an article written by a person that performed several benchmarks with specifically the i7-10700 paired with a 4070. And according to this article, it's a bottleneck of approximately 15.1 on average. And I'm losing about 30-45 fps on average in games. Damn, that's actually pretty significant. But the good thing is that I use a 60Hz monitor, so my max framerate is always 60fps anyways. And the 4070 is so powerful that even with that reduction of frames due to the bottleneck, the game never drops below 60fps. So I'm not being impacted by the bottleneck. So that's the 'bright side' of this situation, at least.
Anyways, sorry for the wall of text. Thanks again for the feedback man. Have a good one!
No issues, just getting a better 3600MHz CL16 ram kit will make your performance go much higher in some scenarios@@DeadPixel1105
@@AncientGameplays Thanks again for the feedback and advice, Fabio! Much appreciated 👍
Bottleneck calculators are good for a Basic understanding. You need to do your own research afterwards ofc. You cannot trust the Result of a single Website. You should always do your own research afterwards.
Thank you bottleneck calculator for making me think I needed a i7-13700K for the 3080Ti at 1440p
Loool
i dont care if my 5800x will "bottlenecking" the future gpu i will grab but i will move to 4k@144Hz instead of 1080p@144Hz upgrading just my monitor and my GPU (currently on 6750xt).
Since i dont have money to spend on a new PC i prefer to put my money on a future rdna5 or 6, or maybe even rtx60 or intel arc 3rd gen. it has to be an 80-class card with at least 4090 raster performance.
All i know is that i play my fav games with video settings that ALWAYS makes the gpu "a bottleneck" so 99% utilization like it's designed to be: in 20 years of gaming i noticed that the bottleneck it's all placebo effect _(edit: if you choosed the right tier parts to balance ur build)._
Maybe in this or that specific tasks i need more cpu or gpu or ram or vram but they are quite rare.
Enjoy ur PC guys stop looking at these things that melts ur brain :)
I just bought a "broken" rx 580 8gb for 15$ , the person who I bought said he can't start his pc, when I tested it, it worked but it was full of artifacts, I switched the bios from the gpu and boom everything works flawlessly 👍👍
Loooool, great haha
A proper bottleneck calculator would be extremely complex and AT LEAST needs to consider some basic RAM information and ingame settings or at least an FPS target.
1:23 I would like to add that some games only bottleneck in specific areas, making the calculation even more difficult.
EDIT: lol never mind you point this out at 5:47
This is why I only use Userbenchmarks, especially when thinking about buying AMD.
UB hates it, then I get interested in it.
You got me in the first half XD
My setup: GTX 1080, i7-7700k @4.2, 2 x 16 GB RAM @3000 XMP1, Samsung 990 1TB. I only play one game - Rust, and get avg 90fps on low settings
cool
Pretty educational video, this one is gold 🥇
Thanks leo
There was a channel called 'Testing Games' on TH-cam that was banned a week ago. Now that they're banned a bunch of copycat channels have popped up using their name, a similar name, and logo. Maybe you should do a video about this.
What? That's odd
Maybe the real bottleneck are the friends we made along the way
pc builds use user feedback to calculate the bottlenecks, so some of the bottleneck if, not every single test, is contaminated by bad people
finally i understand what bottleneck is :)
I haven't built a computer since 2017 and haven't kept up with many released. So my knowledge was severely out of date a couple of months ago when I started planning another build. I was checking a lot of these bottleneck calculators and was just baffled by what CPUs and GPUs I was supposed to be pairing together. But several of the tools were giving me the same bottlenecks, so I just assumed they were right and my intuition was simply wrong or just out of date.
Thanks for making a video like this and showing why the bottleneck calculators are often wrong.
Glad I could help
They wouldn't be _completely_ useless, if they were just more accurate. If they were it could at least give a rough estimate for someone that knows absolutely NOTHING about PCs. Like you don't want to see anyone pair a 7800X3D or 13900k with a 3060. But the problem with that is anyone that knows nothing about PCs isn't going to be doing that kind of research in the first place and they'll just buy a prebuilt. Unfortunately prebuilts are often really lopsided too because they don't want to give you a competent GPU without making a fat margin on it.
If someone could make one more accurate it might be a decent source to point people to that are completely alien to PC hardware. The REAL problem is these sites have absolutely _nothing_ to do with actually building a balanced system and everything to do with coercing you to spend more while using their affiliate links.
It's not a calculator, it's an AD. And I wouldn't be even slightly surprised to find out they're probably owned by a seller. I doubt they've done even a single benchmark, probably just punch random stats into a deeply flawed calculator.
they definitely have affiliate links
3060 ?
12GB model?
Why?
It can serve as Nice placeholder for those with relatively limited budget until a gamer can afford 7900GRE(finaly availible outside of China) or RTX4070Ti Super or something.
Getting something more balanced can mean MUCH sooner upgrade of both CPU and GPU.
With DLSS 3060 12GB can cover most players needs especially when serving as a placeholder for something much better.
@@GameslordXY If you have to cherry pick a placeholder you're missing the point entirely because you _still_ have to spend more down the road. Ironic you claim a balanced build will require an upgrade sooner, when using a placeholder means you still need an upgrade RIGHT NOW.
Besides, if you do that staggered buildout you don't use a $300 GPU as a placeholder, you use a $100 CPU. Unless you only play @1080p low an entry level GPU will _always_ choke a CPU, but with real world settings a low end CPU won't be nearly as restrictive to a decent GPU. Besides, you're choosing to buy 2 of the most expensive component of the system making your overall costs much higher.
If you spend all your time waiting for the "bigger better thing" you'll always be waiting. The Gimped Rabbit Edition is a terrible example as it isn't even a price to performance upgrade from the 7800xt you could have gotten long ago. May as well wait until RDNA4 later this year that WILL give you better value than buying that GPU AMD dug out of the trash.
It is very interesting how the "average PC builder" feelings on bottlenecking kind of went full circle.
At the Core 2 era, most decently new PCs had strong enough CPUs to match or exceed 60fps on console games and most players were satisfied with that experience, so it didn't matter that much. The general knowledge was that most CPUs were fine for gaming, all you needed was a strong GPU. Hell I even remember some channels from my region testing 980Tis with Pentium G4560s, "proving" that CPU "didn't even matter" at 4K.
I guess close to the 2015 era and after, high refresh gaming became more and more popular and people became more and more aware of the need for more CPU performance to reach high frame-rates. Yet, the general knowledge was still that a mid-range CPU from a couple years ago was enough to run most games at high resolutions at decent FPS. After all, how could your CPU bottleneck your system if it isn't at 100% usage? nonsense.
I don't know exactly when it changed, maybe it was when current-gen consoles became a thing, but it felt like overnight a lot of people kept throwing "CPU bottleneck" at anyone that had any performance issue in games. Either you had the highest-end CPU possible or you had a CPU bottleneck. I feel like after one or two years of constant "CPU bottleneck" discussions, the calculators came around. And they never ever had any sort of basis in reality. I feel like people had mixed feelings that every system they put in the calculator returned a CPU bottleneck. It both proved them right that all systems were apparently CPU bottlenecked and also pissed them off because their system was also bottlenecking.
That's when the discussion changed again and it feels like people are consciously rejecting the thought of CPU bottlenecking. Like they accepted it as an inevitability and are aggressively fine with it and very hostile to anyone that dares to point out a bottlenecking scenario. Specially if performance is already high or if CPU usage is very low.
It is a complex topic. Most people simply lack the knowledge about game rendering to understand how each part of the system can affect the whole. People want a "blanket" term: If X then Y, to fix A improve B, and hardware is almost never that simple. Every piece of software utilizes hardware differently and every single system will have a bottleneck in some part of the process, if not, we would have infinite FPS.
Before watching past the first minute mark, I googled "bottleneck calculator", entered my hardware, selected "graphics card intense tasks" (i.e., gaming) and got:
"This configuration has 0.0% of bottleneck . Everything less than 5% should not be concerned major bottleneck."
I say these calculators work perfectly fine. For me. :)
When you pointed out the "Game calculator" existed, I tried that one with Cyberpunk 2077 - 1.9%. Still well below that 5% threshold, and I'm still happy.
That's because bottlenecks have all kinds of weird shapes and sizes so it's hard to tell. lol
As explained in the intro. The inconsistencies inside the same website are easy to find though
@@AncientGameplays I was talking about actual bottles. :D
But the analogy fits imho. lol
When your Cyberpunk test starts, your CPU has 30% less usage than your GPU, so it seems the calculator got the right number, but has the metric backwards or something. They're interpreting their data incorrectly.
that's not how you caculate a CPU bottleneck...at all... as explained across the video
I mean in the introduction he is right to point out being cautions with these calculators
But it's funny the first calculation when he opens the recording (assuming it's 1080p) indeed shows 100% CPU utilization and only 67% utilization on GPU which is not far from what the page said 29.2% vs 33% he actually sees. In 1080p you really need lot of MHZ on the CPU (probably big L1/L2 cache, possibly fast RAM CL) so CPU can request GPU generated frames more often.
That's not how bottlenecks work mate. Its not about the cpu usage
Easy back of the hand bottleneck calc:
Compare CPU reviews under a 4090. Take the max fps at your desired res, then seek out GPU reviews done under a 14900K or 7800X3D, observe which one lines up with the first fps target.
(you can also do this in reverse. take your chosen GPU fps numbers and look up the CPU that achieves them)
Bam, balanced build.
For the most part, any GPU priced for normies won't be held back by any CPU down to a 5600/10400. So just aim for a graphics bottleneck and be happy with a build for a few years. :)
Yup. This is accurate.
When I built my PC I searched which CPU can give me 60 FPS in the most demanding games I would play. It was the 5700x. I looked up which was the GPU that can give me 60 FPS in 2K in the most demanding game I would play. It was the RX 6800.
So. I got the CPU, and 2 years later, I just got the RX 7700XT, since it's the same thing but with AV1 encoder (which benefits me). The result is basically a console that can play in 2K, but with RGB and Windows. So I can also work from it.
My needs were pretty vanilla though.
There's people who play Counterstrike in 640x480. So I see them use 7800x3D with a RTX4060 so that they can get 550FPS. They would play in FHD, but you would need a 4090 for FHD Ultra to pump up those FPS.
Or. There's some people who want to play in 8K30FPS. So they get an i3-10100 with a 4090. The spectrum is large ...
My rule of thumb is ... just don't pair a 6 year old CPU with a current GPU, or viceversa ... Unless the old one is a really high-end product.
Put this bottleneck calculator right up there with userbenchmark EFPS XD
If i remember correctly, r5 2400g and i5 7400 are similar from what i could see on yt but intel one was like 10-20% weaker, more bottlenecking, forgot which gpu i paired with on that site
I have checked some “ Bottleneck Calculators “ online and according to them my 9700K is a 30% bottleneck for RX 7800XT. I am playing all my games on 1440p and have yet to find one that is bottlenecked by my CPU. I always have my AMD overlay while playing those games and CPU is never used 100% and GPU is working fully depending on a game . My RAM is DDR4 3600Mhz 16GB
Depends on what FPS you are trying to hit, if aiming for 240FPS the CPU may very well be the bottleneck.
@@magnusnilsson9792 fair enough. However I only got 144Hz monitor and I get 144FPS in most games.
I'm getting 20.1% on one page, 26.76 on another and 24.5 on a third one that seemed more accurate and showed differences in performance in various games with things like expected fps based on different resolutions and it said it would be playable or not...
One of the tabs said a game in 4k would run at 59.6 fps so it was "not playable" lol what the fuck.
Seems like bottleneck calculators are a snake oil kind of thing to make more people buy the higher end and expensive processors that they don't actually need for gaming.
12100F & 6750XT in case you wondered.
I saw that bottleneck calculator , and I agree is it overstating bottlenecks by about 20 %. That site is useless.
Ok, so what they mean in the 2077 example is that the cpu isnt being used to its full potential. If i look at the cpu usage in the video they are actually really close with the cpu saying 70 something %! No hate but just wanting to clear things up. It also probably says that when you scroll down, you can see the percentages utilised and if it is a cpu or gpu bottleneck. And yes, it does say some bullsh!t sometimes 😅
No, that's not how bottlenecking works. A CPU bottleneck means that the CPU can't process the same amount of data the GPU can in a specific scenario.
@@AncientGameplays yeah i know but for some reason the website doesn't. And besides that it is still bad...
Nice! Was just thinking of you doing a video about this the other day & here it is! =D
Thanks for the video, I was shocked when I used the calculator, so I'm at ease now.
Quick question what is the tool you use to measure the GPU and the CPU?
Thanks in advance. 🙂
Btw, Subbed your channel, nice to see some honest peeps on the net 😁
I think there is a misunderstanding of how this calculator is working. If you play around with it, you will notice that with the same GPU/CPU combo the lower the resolution you go the higher it will say the bottle neck is and the opposite is true of you change it to higher resolutions. I believe the calculator is saying that the more frames your GPU is putting out (which it would at lower resolution) the CPU will not be able to keep up to the shear volume of frames the GPU is able to produce i.e. bottlenecking. Once you slap a better CPU in the calculator, boom bottleneck goes down because that CPU can handle the volume of frames better. This is also supported by the additional information provided under the test result, but I could be wrong. just how I was understanding what I was reading. Now as for the accuracy of the calculator, who knows...
Techically you can download swap files which are slow alternatives to ram that are stored on your ssd when you go over your ram limits. Somewhat similar to resizable bar.
What are you talking about lol. Swap files are a windows thing since ever and they're nothing like rebar lol
very helpfull, thanks for your work, i always found reliable and usefull your content. (sorry for my bad english)
your english is fine, thank you!
Bottleneck and futureproofing are some of the buzzwords that keep coming up from the masses that should be eradicated to be honest. They just make newbies spend more than they need to with problems that mostly doesn't affect them
I think as long as your cpu isn't the bottleneck you're fine
Even when it is, depends on the bottleneck
Yeah, bottleneck calculators make for some interesting people trying to argue points that aren't entirely accurate. Most recent was someone telling me that my CPU was creating a bottle neck for my GPU and I should be on a 7900X3D. Had to explain why, even if it was creating a bottleneck, the 7800X3D would have been a better pairing given the 8core 3D chip vs a 6core+6core3D cache. Better to know what and how a bottleneck is created and build your system around those expectations.
I got a 6700xt used last month. Put it in my rig that has a 12400, 32gb of 3600mhz ram. I have a 165hz 1080p monitor and there was Apperently supposed to be a bottle neck. I honestly haven't seen it yet(mind you I have it undervolted cause it's summer and if I can get away with keeping my room a little cooler I will). Most demanding games I played were control and guardians of the galaxy and at that point my gpu was pushing out 60 fps on both with the highest settings with frame gen too cause I like single player games looking pretty and my gpu was pegged at 98% the entire time. Same for overwatch which is a lighter game but I still get almost max utilisation on my gpu.
I just checked on the bottleneck site for PC-builds.
I have a 7800X3D and a 4070 Super.
Here is the result:
The processor might become a bottleneck for the graphics card's performance. While the graphics card is capable of handling intensive graphical workloads, the processor's processing power may not be sufficient to fully utilize the graphics card's potential. This imbalance could restrict the system's overall performance, leading to slower processing and potentially reduced graphics quality. To achieve a more balanced setup, it would be advisable to consider upgrading to a more powerful processor that can keep up with the demands of modern applications and games.
Well, if you run the CPU at base clock with no v-cache it might be true, since that would perform worse than a 2700X.
😂
Really a lot like User benchmark but without the hate speech. They even missed the monitor Hz, it's just as important as the resolution.
Monitor doesn't matter at all though, that's a different topic
@tGameplays It matters if you bottleneck a system. EDIT: But actually you're right lol anyway many variables to think about.
I'm not sure just using benchmark is telling you the whole story.
I'm getting a bottlenecked using Ryzen 7600 (UV, PBO, offset to all core 5,3ghz) + 4080 super on 3440x1440 ray tracing ultra + DLSS quality without FG.
In busy areas the GPU drops down to 68-74% usage. So in my specific scenario the calculator is on point showing around 30% bottleneck.
Of course your are, you're running 1440P ultrawide dlss quality, which means you are running 960P ultrawide lol
On top of that you most likely have a poor expo ram profile
What I say to friends is that, bottleneck is not important at all, what is important is to have a decent or strong base to build on top.
The PSU should provide enough power to enable you to make upgrades in the future without the need to upgrade the PSU itself.
The Motherboard should be decent enough to give you all the features you need to suit your needs without breaking the bank, a cheap mobo would lack in features and be poorly constructed, while a very expansive mobo doesn't justify it's high price for the amount of features it provides compared to decent cheaper options having most of the same features.
And the CPU should have at least 6 cores for gaming, 8 cores for "future proofing" (if that exists) and more than that for a workstation (provided the work in question will need/benefit from the extra cores)
With the holy trifecta settled, you are much more free to choose your hardware now, or upgrades in the future.
i think for the reason that the 7700X has a more higher boost clock speed and it is more flexible because it can be overclocked. the 7800X3D is only faster in games because of its 3D Cache and there is no room for overclocking. i think the results are actually good because in terms of single core and multi core performance the 7700X is a bit faster.
Another issue is ethier having smart access memory or rebar turned off or on.
Those should be on by default in any pc nowadays
Hello Fabio. 100 percent load means nothing, as does maximum consumption. Because for clarity, in the place where you have a load of 7700x -100 percent. Try installing 7800x3d and see how it is distributed. The load gauge often shows complete nonsense. The chips are apparently very complex things and the sensors cannot show the full load because they are apparently not able to take parameters from all computing units. I roughly compare 2 parameters of the video card 1) Consumption 2) load sensor, this will only show the approximate load.
PS Oh, Fabio, I learned from you that there is such a bottleneck calculator. 🤣🤣🤣🤣🤣🤣
I know and I explaines that in the video. Consumption and load together are the key, bjt they also change according to the resolution, clocks, etc.
You need to know your shit to be sure haha
@@AncientGameplays 🤣🤣🤣