thank you for again doing CPU benchmarks with a "normal" GPU. Too many people forget that unless you ahve a top of the line GPU you are likely going to be GPU bound in every game.
Yeah, I'm only using a 12100F with my 3060 Ti/2560x1080 resolution and in most AAA/new games I'm GPU bound anyway since I tend to crank up the settings as long as I have the GPU headroom for it. I don't play competitive games tho so I can't talk about those, but in any other game I've played/playing atm I see little to no point in upgrading for my use case.
Even with a top of the line GPU, if you play on maximum graphics settings and resolution higher than 1080p, you'll also be GPU bound. CPUs are just not gaming parts.
@@garrusvakarian8709 That is simply not true, in almost every actually GPU demanding game its my 3060 Ti that maxes out first and not my 12100 at my 2560x1080 res. Currently playing Cyberpunk and my GPU sits at 98-99% all the time on high-ultra settings with some RT on top and DLSS Quality enabled. Same story with Spider Man and God of War or any AAA game from the past few years. Spider Man is probably the closest I've seen both maxed out at the same time with RT enabled.
I understand why these tests are all done at 1080p, but I would still like to see even just a quick summary of 1440p or even 4k just to further drive home the point that, once you go past 1080p even to just 1440p, the CPU starts to matter A LOT LESS.
It really depends on the quality settings and the game itself. You'd see much the same in the heavily CPU limited games such as CSGO and ACC for example.
@@Hardwareunboxed Exactly. I would often reduce the graphics settings on 1440p to improve performance somewhat with my 6650XT, so I get slightly worse visuals (but higher fidelity just because it's 1440p). The effective CPU load would be near the same as 1080p high. I'd always try to target 165fps and that doesn't care much about GPU (as you can tone down settings) but does care about CPU as in some games you WILL hit a wall regardless. A lot of people are sleeping on the fact that a CPU bottleneck means you're SOL and with a GPU bottleneck you have a plethora of ways to improve performance by tweaking settings. Actually, maybe you could do a short follow-up to illustrate this, e.g. 1080p high/ultra vs 1440p medium. It will depend on the games, but just a couple would validate 1080p testing.
@@Hardwareunboxed You can keep the tests 1080p, but you really should enable ray tracing in games that support it. It's by far the most CPU choking setting I'm aware of in games like Cyberpunk.
because this is CPU scaling BENCHMARK, the point is to show here where you gain absolutely nothing from higher end CPU. Naturally, absolute majority of RTX 4090 owners would play at 4K - but still, you won't be pairing it with budget R5 5600 anyhow, because who pairs $2000 GPU with $140 CPU?
Also nobody using a 4090 or a 6950 XT would be gaming at 1080p, so those results are more or less irrelevant. I would like to see how much difference there is between the CPUs when running a resolution that is appropriate for the GPU. Most likely you'll find the 7600 is a waste of money in the current situation (I know there is always this mysterious "future-proofing").
I'd like to have seen more Nvidia cards there, but the upper range of cards is kinda useless to 5600 users. Maybe the 4070ti. It's so annoying not having a bloody 4060/rx7700 up there to compare against, mid range cards. Overclocking the 5600x can give 5950x levels of speed though.
@@Alvin853. It’s not so much “futureproofing “ but more like it shows what it would be like a couple gens down the road. Also, some might save on the cpu and put the money into the gpu.
When the 5600x first came out in November 2020, my first impression was that it was a phenomenal CPU beating even the i9 10900k at gaming and matching the 10700k in productivity. The 7600 has overtaken the 5600 by more than 40% better performance in some games and matching the 5800x in productivity. This is what competition brings and we really need to see this kind of progress in the GPU market.
@@nikosensei1258 The multi core score in cinebench R23 of a maxed out 5600 was 12400ish points, with a maxed out 7600 you score 16350 (higher than a maxed out 5800x which maxes out in the low 16k range with a solid overclock. Pretty impressive for having 2 less cores.
A good comparison would be the 5800X3D, 7600, 5600, and 3600 at GPU limited situations (1440p) to see if upgrading to AM5 is worth it. Currently using a 3600 and a 5600 XT. Considering upgrading my GPU to a 6800 XT and sticking with my R5 3600 until prices come down.
@@MadClowdz brother has it already. Better to upgrade to a 5800x3D for a more substantial upgrade but not soon. Too many non gaming expenses. Having a working rig is already a luxury.
@Andre I don't think the price of the 5800X3D will come down until the 7000X3D series comes out. Heck, it's already going out of stock at times since it's so popular.
As an "average consumer" (no high end parts, 3+ years between upgrades) those benchmarks are by far the most interesting to me. Comparing CPUs with 4090s is interesting for sure, but testing and comparing low to mid range systems is the most relevant when you seak reasonnable "cost per frame" upgrades. So thanks again and please keep it up :) cheers
You also need to keep in mind that the higher performance of the 7600 with a 4090 does in fact apply to casual gamers too. Reasoning for that is future GPUs. You might not see a 20 - 30% increase with your current PC setup, but in a few years time when it is time to upgrade your GPU the more budget cards will be closer to the 4090 performance, which would mean you wouldn't need to upgrade your CPU/build a new PC.
You are not an "average consumer". You are a mid-high one. I am an average consumer. I got a low end GPU, a low end GPU (tho, that was from my mistake, I could had gotten a better one) and an 8GB ram (while at the time, 16 was considered HIGH). And my PC build was from almost 5 years ago in 2018 and I don't have plans to upgrade very soon.
Just built a Zen 4 PC with my nephew. He got a 7600 paired with a b650 200$ mobo and a 140$ set of 32gb 6000mhz 36-36-36 memory. The whole build was 1200$ before sales tax with a 6700xt, and he’s happy as can be playing all his games in 1440p high settings. Was a HUGE upgrade over his hand-me-down 1660ti laptop he’s been using for for the past 2 years.
Beast! That’s last years, just upgrade the GPU in 3-4 years time. Just built my daughter a Ryzen 5 5600, had an old AM4 board kicking about with 16GB DDR4 3200 and my sons old 3060 12GB and it gets on the Red Dead 2 Benchmark Avg 59fps and low of 39fps on a 3440x1440p monitor. 👌🏻
@@papalazarou7880 Yeah, I’m really happy with the build and the performance for the money he spent, but I’m even happier that he’s being responsible enough with his money that he’s managed to save enough to buy himself a decent used car and a modestly specced gaming PC since he graduated HS last year. Had fun teaching him how to fix up his car, and had more fun teaching him how to build a PC. He’s a good kid.
thanks for the great tests! I upgraded from r5 1600 to r5 5600 on the same b350 motherboard I had and I dont think ill upgrade for quite a while until AMD comes up with at least double the performance boost. kudos to AMD for the great value motherboard chipset from 1st gen.
Yeah upgrading should only be worth when changing for double the performance at the same price, using cpuz as reference the single core performance for the r5 1600 is around 370, while the 5600 is like 609, meaning that the difference wasn't exactly double the performance but close to it and at least the 5600 was cheaper than the 1600.
@@rawdez_ Because of the same reason both nvidia and amd have prevented the generational progress on the low end and even mid low, just look at amd selling that 6500xt that performs exactly the same as the 480/580 or /470/570 on pcie 3.0 and the 5500/590 in some games for the same cost despite the 5 years difference, nivida does the same the 3050 should have performed like the 2060 and the same 3060 should have performed like the 2080 and not like the 2060s those companies don't want to release strong performance for the low segments cause then nobody would upgrade their gpu's if the low ends are capable of 100 fps at high settings, so they prefer to keep low and mid segment at stagnation so you are forced to change your gpu at the next gen.
Too bad they didn't include other popular graphics cards like a RTX 2080, 3060 Ti or RX 6700 XT! Also the testing method of "only 1080p" and medium details (for example CP2077) is ausgelegt to be CPU bound! I don't know about you but when I invest in something like a RX 6700 XT or 3060 Ti (from something like a GTX 1060, RX 580 or 1660 super) I would definitely crank up the details to high or even ultra if not go 1440p and step up to "next gen" performance (PS5 ish with 1440p, high details) targeting 60 fps+ in story games like GoW, HZD or CP2077. That takes the focus back from CPU to GPU only! Hardware unboxed had another video with comparison of Ryzen 2600, 3600 and 5600 which were similar.
I think it is better in the future to also add a more in between card. 6650XT to 6950XT has a large performance gulf. Maybe a 6700XT and/or 3070 should be enough to fill that gap. Also preferably to also test 1440p since it is a popular resolution, thus can give the viewers better information on deciding whether to upgrade their CPU or GPU when playing at that resolution. I picked 6700XT/3070 instead of 6800 or 3080 because afaik those GPU will still be CPU limited at 1440 most of the time, thus 6700XT/3070 should be the better choice for more variety on the data.
In addition, I am very interested in both VR Scaling and Ray Tracing scaling with CPUs... I would love if HBU could throw just one VR and Ray Tracing bench into the CPU testing.
I have 6800xt with 5600 (with pbo2 though, basically a 5600x but with better all-core boost), in majority of cases I am GPU limited. The only exception in my use case, apart of 2d games, is Path Of Exile. I play at 1080p, though.
To add a data point to the GPU/CPU bottleneck question I offer my anecdotal experience with my 5700x and RX 6750 XT Red Devil. I play at 1440p with as high settings as I can and am GPU limited. Not sure what 1080p would look like except max settings would probably still give high FPS. However, at almost max 1440p graphics with RDR2, Baldur's Gate 3, Horizon Zero Dawn, The Callisto Protocol, and World of Tanks I simply don't care if my rig isn't giving me all the bells and whistles at 100%. Why would I with average FPS of ~165 on some of those titles? Besides I'm getting too old and way too beat up to be competitive in 1st person shooter type games.
I would really appreciate 1440p results as that's the harder to answer question on paper. At 1080p a newer CPU will obviously benefit but if I'm running games at 1440p or 4k how much would I really be getting out of a new CPU?
I wonder about it too. I have 7900XT with 5600X, what am I gaining by switching to new gen if I game at 4K? I am already getting 60+ FPS in most games with high settings. What about 5800X3D, is it that much better than 5600X in 4K?
@@dougalmacburnie6510 It's negligible, the amount people try to achieve by upgrading to a better cpu is usually lost by overlooking the fact that you do need the fastest possible ram also as shown on benchmarks.
That's pointless, as it would be entirely dependent of what GPU you have. A 4090 might still be greatly bottlenecked by the 5600 at 1440p in some titles, but this doesn't mean your GPU would.
I've just built a new PC based around a 5600x CPU. My board may be ddr4 and pcie4 limited but I feel like there is heaps of life left in those two specs. My upgrade plan when I go to buy a new GPU in a few years time is to find a good 5700x/5800x/5900x on the second hand market as I'm guessing new ones would be all but gone. I'm taking a bit of a gamble here also in noping the they won't be too expensive either on the second hand market
Can we take a moment to appreciate the extra time it takes them to do the GPU scaling for these videos? It's amazing real-world valuable info that many people under-appreciate
Testing with these high tier gpus at 1080p just feels pointless because pretty much no one buys a 6950 xt (let alone a 4090) to play at 1080p. Think it would be much better to keep the 1080p results with 6650 xt but include 1440p and 4k with 6950 xt and 4090. I know the cpu matters less at those resolutions but at least you would be able to see how much difference the upgrade makes. Also dlss/fsr should really be included at this point
I'm impressed with the 7600 but can't wait to see the 7000x3d's. I just got a 5800x3d a week ago, and with a 3080 at 1440p I think I'm good for a few years.
Hell I have a 5600x and a 3080 now (up from a 2070 S) and the difference is night and day. At 1440p, I too think I'm good for years now. Or atleast I better be. Genuinely looking like we're plateauing on performance boosts these days, but can those gains outpace the poor console ports we keep getting is the real problem.
Im in the same boat, I get my 5800x3d delivered today (upgrading from a 3900x), and I have a 6900xt and game at 1440. I think I'll be happy for a few years. Hopefully, the market will come down by then though lol.
@@Dynasty1818 I had a 5800x before which is very close to the 5600x so I know that works well with the 3080. I knew the 5800x3d was a lot faster but what I didn't expect was how it's pushing my 3080 really hard now. I had to bump up the fan curve because it runs about 10 degrees hotter now and plays so much smoother in heavy AAA games. I recommend for you to eventually get a 5800x3d if you can because it's amazing, but I would have been happy with the 5800x, it's just that I wanted to max out my pc and not have to upgrade for years and I think I've accomplished that.
5600 is a very good option when it comes to 1440p gaming because there is a much smaller difference in frames compared to 7600, we must also have results for 1440p here.
@@jotunheim5302 unless he’s using RT the even his 3090 is capable of pushing more frames at 2k ultra than his 5600 can supply, though I agree does it really matter if he’s getting 150 vs 165? The more interesting question to me is how to measure the smoothness and stuttering that many games have with weaker cpus. Average fps and even .01 lows don’t tell that story. Minimum frames does and minimum frames is heavily cpu dependent, yet we never see that in review videos.
I’ll give an example. My old system was a 1300x and a gtx1060. Online cpu calculator tells you that is a great match for 1080p, very little bottleneck either way. Massive stuttering at ultra settings. Recently I change out 1300x for a 5700x and everything else same. Frames are still low 30s average because gpu is maxed out, but the stuttering in hitman and assassins creed is now gone and it’s now a smooth 30fps and playable. That’s the power of a good cpu and what all the reviews ignore. Sure it doesn’t matter now when your system is new, but it matters to those who want to stretch their system life as long as possible.
Great content! I would love to see graphs which show where any of the graphics-cards reach the cpu-bottleneck. In other words: What cpu do I have to use at minimum with which graphics-card to not hold back the card. For all AMD and Nvidia-cards...
Its very dependent on the game your playing. Most games media benchmark run great on any modern cpu. But other games that are less "mainstream" is where cpu performance really makes or breaks your game play.
@@ramair325 it's more about having a tendency I think, as we clearly saw with these benchmarks, that there is usually a huge improvement going from 5600 to 7600 with a high-end graphics card. And then, you would have to define what is your target framerate/resolution that, if not met, would be considered bottlenecking. Personally, I would say 1080p@90fps average in order to get at least a 1% low of 60fps, but I am pretty sure the demand range would scale all the way from 50fps to 200fps average. Going by the charts, even a 6650XT is plenty for me, making any CPU viable, meaning, for me, we can't be possibly talking about CPU bottleneck at the moment.
@@sassuki ya and thats my point, these benchmarks make it look like cpu performance does not matter. but what they dont show is how bad "81fps average" can really be in a game because at some point in that game you might be getting 140fps but at other points you maybe down in the 30's. And most games are very cpu demanding in many places if not most places other than a few AAA games that have been really well optimized. games can also be very gpu demanding in many places as well but at least with a gpu bottleneck you can reduce game settings and gain massive fps boosts. its not really the same thing for cpu's. This is also why i hate these type of reviews as it shows the 7600x as the god gamer cpu when in reality the 5800x3d is going to be far better across far more games. dont get me wrong i know why steve is doing these benchmarks but its just making the water muddy for consumers.
@@ramair325 Naah, it's not that bad, you're overcomplicating it; the rule of thumb of "Average is usually about 1,5x the 1% low" holds true. Some exceptions are games like Fortnite where the average is roughly 2x the 1% Low. So even if you take Fortnite into account, nobody should need more than 130fps average to get a minimum of 60fps, which is enough to get smooth gameplay even on old shitty 60hz monitors with v-sync. With a 1st gen FreeSync monitor, even 100fps would be enough to keep the gameplay tearing-free. So at the end, there is no CPU bottleneck with any of the CPUs presented here, just overperforming GPUs for overinflated egos, that need to justify why they just spent a 1000 bucks on a f*cking graphics card!! lol People tend to forget where we came from: in 1996, you needed a special 3D card to get solid 30fps in Tomb Raider... at 640x480! Talk about some demands nowadays: they get 3 times the framerate at 6 times the resolution and dare talk about bottlenecks! lol
@@sassuki it can be far worse than that in some games. and thats my point. the games used for these tests do not represent a worse case scenario rather quite the opposite.
Every CPU Upgrade I do usually includes an increase in both Thread counts and clocks when some game just starts being too limited. I switched from i5 4690 to i7 4790 because of Battlefield in 2017 and to R5 5600X in 2020 because of Watch Dogs Legion and CP2077. So I think this Gen the 5600 is definitely still a keeper. (I am also using a 2080 ti) Maybe new UE5 games will change that but until then I'll wait and probably do a completely new build when Cards with 4080-equivalent Performance are available for 500/600$. It will be a long time til then but as long as the PS5/XSEX are the current console generation, the 5600 won't be obsolete.
@11DaT11 the 4060 ti is supposed to be on 3080 performance but apparently it’s just a 3070 lol which is kinda a joke considering a 3060 ti from last gen is basically the same as a 3070 maybe 5-10% slower
Really like this Scaling benchmark series, would like to see more videos. We get to see the real world usage of budget CPU with appropriate budget GPU too. Thanks Team HUB.
At 4K high/ultra settings, except for some older games and FPS, most games I'm under 100fps with my 3080 12gb. So for me the upgrade from my 5600X doesn't really make sense, my 3080 is always pegged at 100%. Point being, unless you are CPU bound or just mildly GPU bound, whether it's at 1080p with a 3060 Ti or 1440p with a 3080 or 4K with a 4090, you shouldn't really be looking at upgrading from ZEN 3. At 100% GPU utilisation performance will be the same.
true , maybe your 1% lows would be a bit better, but thats such a minimal performance increase for so much money. you are better off just buying a better card :p
@@Bastyyyyyy agree unless you mostly play multiplayer or cpu imtemsive simulation/sandbox games like cities skylines and teardown cpu upgrade is pointless unless the system is too old that it would bottleneck triple a titles which is not normally cpu heavy, unless that triple a games is terribly optmized like gotham knights,
Knowing that I don’t need a newer cpu to pair with my 3060 ti makes me happy. I’m using 5600 and it’s been the best upgrade I did from 2600. 5800X3D would’ve been great but I’m already satisfied with performance I get till another new gen consoles hit. (if newer PC AAA games are optimized…)
I upgraded my old R3 3100 and gtx1060 3gb pc on Christmas. Using old RAM and a520m motherboard, I got R5 5600 and RTX3060ti for like $400. The money well spent. The performance were jumped, like doubled or even tripled the FPS, and now I got a PC that I can play games easily for 3 more years.
Yes! Would love to see a test at 1440p and 4K for different CPUs such as 5600X, 5800X, 5800X3D, 5900X, and the Zen variants, but also Intel variants and see how they stack up when using 4090. Of course, it is a lot of work, but these things can affect buying decisions.
I’m assuming when we move to 1440p and 4k, the differences between 5600 and 7600 would reduce further. Which means if I’m on 4k atm, I really don’t “need” to upgrade. Thanks for the content!
@@johnboylan3832 what games do you mean? On high settings 4K you dont got much further than 100FPS anyway and in most games, the 5600 is enough for that
@@otozinclus3593 - Cyberpunk, AC Origins, AC Odyssey, Far Cry 5 all were CPU bottlenecked in certain areas at 4k/high settings on a 3090ti with a 10900K which is why I moved to a 5800X3D. For example, Alexandria and Memphis in AC Origins and many areas like Athens or even in the wilderness in Odyssey bottleneck pretty severely well under 100fps on the 10900K even at 4K. The bottleneck goes away on a 5800X3D. So even at 4K, CPU choice definitely matters if your GPU is strong enough.
@@otozinclus3593 The two most recent I was playing after I got my 4080 and before I upgraded the CPU were The Witcher 3, which was about 35-40 in the towns; Spider-Man was 65-80. The former is now 45-60, the latter 100-105. I upgraded to a 5800x3D. The 5600x struggles a bit at 4K with ray-tracing.
@@terraincognitagaming No they don't you are projecting. The ability that he _might_ be able to slap on a 9000 series CPU into his AM5 board is what he is really referencing. That kind of longevity is not common at all with intel, and if Intel stay true to the course, a Meteor Lake CPU will need a new platform. In fact, one will need a new board for what comes next after Arrow Lake if Intel keeps the 2 generations per socket march. It is conceivable, that AMD may even release 10000 series CPUs on AM5, but I do have my doubts about that. All that being said, I have always felt that by the time an Intel system is considered _"too slow"_ for most, it is probably time for a new motherboard anyway, case in point how long has the 9900K been working for *many* users? But the thought of putting 3 maybe 4 generations of CPU on one board is now very important to a lot of consumers. Also, AM4 and Socket are for all intents and purposes a dead end, no more upgrades for them.
@@terraincognitagaming I honestly don't understand people like you. Have you lived away from civilization the past few years? Here in the comments, people wrote hundreds of times that they changed their Ryzen 1000 and 2000 processors to 5000, while receiving a huge performance boost for quite modest money. I have personally done the same on two different systems, which are noticeably faster now. In fact, even switching to the Ryzen 3000 series was already a significant step towards increasing performance, especially considering that the maximum number of cores has doubled - from 8 to 16, and all this without junk cores, like Intel has - Atom cores never intended for desktop solutions.
@@terraincognitagaming I don't think anyone is really _rockin_ that CPU in modern games and certainly not anything where they want to be ultra-competitive online. That CPU is the perfect example of a CPU that is *too slow* and the platform it is on too old for the newest and most demanding triple A releases. Nobody is putting a GPU of any substance on that 9-year-old CPU, well they shouldn't. Sure, it may be good if you want to play games from 5 years ago, but it will get decimated by a 12100/13100 in today's software. Don't get confused with casuals holding on to hardware because they don't play demanding games or games with higher fidelity vs enthusiasts here that hold on to hardware for a few years with a GPU upgrade along the way.
@@terraincognitagaming Dude, nobody but ultra-casuals are holding on to their system for 10 years to build or buy a new one. Enthusiasts will have a GPU upgrade along the way and usually build a new system within 6 years. This is why I have seen a number of 9900K users talking about their build either this year or for Meteor Lake, it is right around that 6 year mark. You are talking extremes that do not reflect the mainstream enthusiast level. That is one reason that having a platform that can accept 3 CPU generations or maybe even 4 is such a positive for them.
Float plane member here and glad to contribute ever since the green team silliness a while back. Keep up the good work. I appreciate independent tech coverage.
I think the review suffered from not including 1440p data. I think the RX 6950XT data will look even less extreme in terms of differences among the processors.
I went with the 5600 because I mainly needed a second PC besides my notebook for as cheap as possible but at good performance. This worked great for me because I haven't had a desktop PC for many many years, so the jump in performance from zero (well, actually Core 2 Duo) to Ryzen 5600 was huge! :D I also preferred the more stable and known platform over the brand new one with DDR5 and the new processor generations. And buying into the new processor generation (for example Intel 13400F) but sticking to DDR4 seemed wrong. And DDR5 is too pricey for now. Also, the new Ryzen CPUs are still too new to have good prices and I wanted a Ryzen and not an Intel for now. And the Ryzen 5600 seems very good overall. Also, I usually don't do partial upgrades of my PC hardware. So, the end of life for Socket AM4 is no problem for me.
I get this was to show CPU scaling, but it would have been nice to see different resolutions as well to provide a large overall picture of the importance of CPU performance as I don't think most people are building new PCs with 1080p in mind.
PERFECT!!!! That's exactly what I needed! You should mention to buyers that, for higher resolutions (all other things Equal) the differences % in FPS will be even smaller, for example, i am looking for an 1440p 34" monitor, the difference in the 5600 vs 7600 its even smaller and in my opinion its better to get the cheaper option and invest that money into a better GPU, so this 200$ difference in your example should go to a better GPU , to run the higher resolution monitor!! Thanks Steve!!!
My monitor was a ROG 27" IPS 165hz unit. I replaced my 4K TV with a Hisense 55U8G which is wall mounted. My GPU was an EVGA 1080Ti FTW3, 5800X, X570 and 4x8 3600cl16. My TV is now my monitor and my ROG is back in its' box. Yep, no RT with a 1080Ti, and with "reasonable" eye candy settings, the 1080Ti performed very well with what I do. WoW, elder scrolls and GW2. If you're thinking of 34", wall mounting a 55" (U8H $700 USD via Amazon) could be an option.
Like the video, but I’d really like to see the 4K tests that others have requested. My particular question is: how much, if any, performance would I be leaving on the table by running a 5800x with the 4080 or 4090 at 4K?
I upgraded my Ryzen 2600 to a 5600X and my RX 580 to an RX 6700 XT for a total of $300 around 2 months ago. It's like using a brand new build and I'm still on my Tomahawk B450 MAX motherboard. No regrets.
Recently got a 2nd hand 6600xt for an AMD combo with 5600x, and it looks like I don't have to spend any more money on a PC this year. Thanks for the benchmarks!
This has been one of the most helpful videos for my use case.. 6600 xt @ 1080p and mostly single player games... the upgrade ATM would just not be worth it.
I really wish the 3600 had been included here as a baseline. If someone is pondering whether a 5600 or 7600 upgrade would be worth their money, being able to compare either of them to their current CPU would be very useful. Perhaps the best upgrade is no upgrade! For instance, I have a 3600xt-based system. And a 120Hz 4k display. (I currently do 90% of my gaming on a console.) If "good" GPUs ever come down in price, I might actually want to buy one. Do I even need a new CPU? Probably? Maybe? I'd like to see the 5800X3D included in these comparos as well.
3600 is a pretty good rough proxy for current gen console performance. It should hold up for a while yet for a midrange rig, but it's not a rocketship for sure.
I hope you reach 1m subscriber this year! Love your content! Do you still feel like cpu don't make much difference at 4k still? i feel like its starting to make a difference considering how fast all those new CPU are catching up to the GPUs over the past few years, not as much as lower res but still , would be interesting to double check!
I would like to see 1440p and 4k scaling as well with new processors to see if there is any real improvements. I do understand you will be mostly if not all GPU bound but sometimes there is a difference depending on the game.
It would tell anybody that you dont need to upgrade a cpu @ 2160p and mostly 1440p, not in interest of Hardware Unboxed, the goal here is to keep all hype alive, clicks and visitors count at any cost. These guys are money oriented as all youtube, just more truthlike acting.
Thank you for benchmarking GPUs and CPUs for the general public. Just look at steam's hardware stats to see what hardware the general public is using. Many tech youtubers live in a cloud and only benchmark with the top of the line. There should be more video testing more modest hardware
It would be good to show a couple higher resolution examples or mention it since that really does equalize differences in CPUs even more. So if you prefer 1440, then you'll see even less of a difference between the 5600 and 7600. 4k even more so, but requires a beefier GPU anyways which may make you want a better CPU to utilize the GPU better when changing settings or playing less demanding games.
Would be intersting if you make an analysis about which GPU class is start to make difference between 5600 and 7600 My assumption is around RTX 3060 or RX 6700
As an owner of 5600 and RTX 3060, I can assure you that this card is not powerful enough for 5600 to introduce any significant CPU bottlenecks at 1080p, at least in (relatively) modern titles. Some 10 year old games limited to single threaded CPU load (like CS GO) might be exceptions though. But I would not call it CPU bottleneck in this case, as 200+ fps is hard to call a "limit".
I assume more like at least RTX 3080 or RX 6800 when looking at 1440p high details. The testing is this video is also taken CPU-bound -ish since it's only tested in 1080p, sometimes even medium details! But when spending so much money at a RX 6700 XT or better I assume most people would also like to play at ultra details (if not with RT) and 1440p to experience graphic gems like GoW, Horizon Zero Dawn, Witcher 3 or Cyberpunk 2077 in very high quality.
@@starstreamgamer3704 nice to know, my point is showing the diminishing return graph. They already have the data. When they reviewed 5600 they use should already have data when used by low spwc gpu to high tier. Then make a graph of diminishing return. Compare graph of diminishing return from 3600 and 5600, then you can compare the gradient of the graph and decide where is the sweer spot or where is the biggest performance jump
I'm running a 13600K on an ASUS ROG DDR4 mobo with 4x32GB G.Skill Royal 4000 CL15 RAM clocked at 3200MHz and a fast NVMe SSD with a FE 3090 ($700 used from eBay). This system cranks all day at 1440P, so the value proposition of a DDR5 or AM5 platform just isn't there. Given that I paid WAY too much for this RAM ~3 years ago for an AM4 gaming system I was running, I couldn't not use it and I'm happy I stuck with it. For kicks, I put my GPU, CPU and SSD on an ASUS ROG Z690 MATX DDR5 mobo with 2x16GB T Force 6000 CL30 RAM and it had identical performance in Dyson Sphere Program and CP2077 as my DDR4 platform... However, running a 3080 and 5800x3D on an AsRock x570 Aqua motherboard with the same G.SKill DDR4 and fast NVMe SSD in a custom loop, the 1440P ultra FPS performance is ~20% BEHIND the 13600K in the games I play (regardless of RAM type on a Z690 motherboard). This is due in part to the 3080/3090 uplift I'm sure, but the 13600K is a beast of a CPU. Again, if you still have decent DDR4 RAM, AM5 costs a lot to adopt, but I completely agree that the sun has finally set on AM4. I think these tests really highlight the need for a current gen CPU with the RTX 4000 series cards and that for midrange CPUs, last gen GPUS are still a great value in price / performance. If you want to stretch your dollars, the 5800x3D is awesome, but I'd love to have seen it added to these tests to ensure the RTX 4000 GPUs show the same (expected) uplift at 1080p. I assume they would, but it'd be a great confirmation. Thanks!
Exactly the video I was looking for, thanks guys!! I want to move onto a new platform but want to keep my GPU because I love the way it looks and it already performs the way I want
Im curious: at what level of GPU performance having a Ryzen 5600 would be a severe bottleneck? Im talking about future GPU upgrades down the line. It seems that the furthest a 5600 can keep up with is something like a 4070ti, which would mean a rtx5060 in a few years Sounds fine to me
I really appreciate this kind of scaling review. I want to go pretty high end with my next GPU, but I think I'm better off replacing my 5600x/b450 before doing so. Thanks!
The lowest end GPU results are the only ones that matter here. Any real world advantage of the 7600 make sense with a higher end GPU which immediately begs the question: if you can afford a top of line GPU, why settle for any of these CPUs? It’s a paradox that renders most of these comparisons moot.
It's the difference between a GPU vs a GPU, a motherboard, CPU, DDR5 RAM. Maybe you'll need a new power supply with this upgrade. Hell, let's just get a new case while we're at it and just like that you bought a whole new PC
Killer love this video. I have 5600 and haven’t struggled yet in any game but I do get the itch to upgrade . But 5800x3d just seems like better route for me if I ever do. Using 6700xt and 32gigs ram
@@Felix_Schmitz well I agree but when I see specs on some of these new games like hogwarts and forspoken requiring 8 cores for 1440 p ultra it just makes me wonder if that’s bs or just terrible ports
@@bmdshred77 Yeah, the consoles also have more threads than 5600 but I wouldn't upgrade until it is apparent that 6 Cores aren't enough for Console equivalent settings. But I am using a 2080 ti with my 5600X so for me the GPU will probably be the bottleneck most of the time anyway.
I think most people who would go for 5600 are those who are building a Budget System. And based on that idea and my local pricing (non-US), the GPU's (only considering GDDR6 variants) which would fall under the Budget category would be either "GTX 1650" or "RX 6500 XT" or "GTX 1660 Super" or "GTX 1660 Ti" or "RX 6600" (for people in my region, sub $300 GPU's would fall under Budget, not sure about other regions).
It would been nice to include a 3600 and 5800X3d into the mix. My 3700x seems to be bottlenecking the 3080Ti at 4K, and am thinking if I should upgrade. But not sure to 5700 or 5800X3D as there a massive price difference.
Really depends on what you play, for most games I play (primarily strategy type games) the 5800x3d vs 5700 would make a pretty big difference, but for many popular games it really isn't that different.
@@h.b.5577 I mostly play FPS and still see the GPU usage howering around 92% many a times with DLSS on as its rendering at lower res with DLSS. But I guess it will either have to be CPU or GPU bottleneck, they never go hand in hand, so I will leave it for now and see how it goes. I am getting good 100+ fps anyway.
@@coolvinay That's really what it ultimately boils down to, always going to be bottleneck-ed one way or the other. I recently upgraded my couch gaming pc from a rx 580 to a 6700 xt while keeping its 9600k which now often results in cpu bottlenecks, but if performance is still 60+ fps I don't really see the need to worry about it so much.
Also got a 3060 Ti but went with 7600x after the prices dropped. Definitely doesn't make sense in games today (particularly as I mostly play at 4K!) so I've basically taken a punt that DDR5 / PCIe 5.0 will eventually become relevant in games. Will likely upgrade the GPU in two or so years, then the CPU further down the road (presumably AM5 will have an equivalent "swansong" like the 5800X3D). TBH the main reason though was because I had a bad experience with my previous Intel system (which was very old but was still just about suiting my needs), the motherboard died and replacements just didn't exist at non-crazy prices, so I was without a working desktop for a few months. Definitely prefer AMD's approach of supporting a platform long term, instead of just for a generation or two. Buying into a new platform hopefully means I'll have good options for upgrades / replacement parts for the next 5-6 years.
@@tyre1337 Didn't feel like that when I tried running the Witcher 3 with RT on a 3060 Ti. Feels like some kind of conspiracy to get me to buy a new GPU.
While a good indication of raw CPU performance I don't think this video really works as an indicator of GPU scaling for the next couple of years. The 6650XT is a more entry level card and is still pushing nearly 100FPS across all title in this vide at 1080p and even 1440P gaming at medium settings around 60FPS is doable at entry level. The days of 1080P are clearly limited and going to shift to higher resolution gaming in the next 2-3 years now that high end GPU's are deliver 200-300FPS at lower resolutions. It will take at least 1-2 generations of new GPU for a lower end card to deliver the performance of a RTX 4090, so if we really wanted to answer the question of this video "Is Zen 4 worth it" the testing should have been done at higher resolutions and see how much bottlenecking actually takes place and how the 5600 will performance in a few years.
Its great to see the difference but would have loved to also see a 3070 or a 3060 there so that ppl on a budget that update a few years down the road could have used this video to make their choice
I ended up upgrading from Zen+ 2700x to Ryzen 5 5600 for in socket upgrade with my Asus X470-i on the super cheap. If I was to build a new Ryzen system then would factor in a higher core count 7900 series chip. Besides, I already had 32 gigs of DDR4 and didn't need to reset or start over to upgrade. Good review Steve on another excellently written comparison.
Wow. First i’d like to say great work. I really love these ‘scaled’ benchmarks. However, i’d recommend including a GPU in-between the 6650xt and the 6950xt too. Maybe a 6750xt/6800 or 3080? Because we can see a pretty noticeable difference in some games between the CPUs when we use a 6950xt or 6650xt - but I think it would be cool to see the result of something that sits in the middle of that spectrum too. Just a suggestion! I know this stuff is hard work lol.
Hoping the 5800X3D will drop when the new X3D chips come so I can get it at a better price than it has been. It seems to have started dropping now (in Norway), so hopefully that will continue and I can replace my 5600X with it (using the 6900 XT).
idk about that. i think they will just thottle down supply for it to make space for the new x3d so the price wont really go down because a lot of people think about upgrading to a 5800x3d that are on am4 still. it could even get a bit more expenive imo but we will see
my building philosophy is: get the best gpu for your budget and pair it with the cheapest cpu that wont bottleneck, glad tô see that it still the way tô go
I just got a 5600X a couple months ago on my AsRock b450m Pro 4 and I plan on sticking with that for at least 2-3 years before I even think about upgrading my total system. Paired with an RTX 3060 it handles everything at 1440p easily and DLSS is awesome.
Sat content with my 5600 and 1070ti. The 5600 will remain viable for at least 2 more GPU upgrades for me before either switching to 5800x3d or going AM5
Curious, how is your 1070TI holding up at the moment? I have a 1660TI, but it is hamstrung by a very slow CPU, a 5600 is what is on my radar as a drop-in replacement.
@@davidandrew6855 perfectly fine for my needs at 1080p 144hz depending on quality settings. Jumped from a 3100 to a 5600 which made a decent difference in FPS and overall PC feel
I do not see the point to make a comparison in 1080p resolution with the most expensive graphics card (4090). I doubt that if somebody is having such graphics card that he intend to use in for 1080p and chose the CPU for that....
You know there are 2 other GPU performance levels in the test, right? You *always* use the most powerful GPU and lower resolutions to see just how much performance a CPU has.
Interesting how the value of the 6700xt was listed as near one of the best options, but we rarely see it used as a midrange option despite its performance improvement compared to the 6650. That said, I love these comparisons. C: You guys rock!
The RX 6600, 6650 and their XT models are often the best midrange cards with an attractive price point. When someone has more money spend I assume often goes more "all in" and chooses the RX 6800 or better right away. I have the same feeling about RTX xx70 products! Many people buy RTX 2060/ 3060 or go 2080/ 3080 right away. The success or the RTX 4070 at the moment comes from a missing RTX 4060 Ti option, I guess. But I would also see more midrange in those tests as well as with high details and not only 1080p medium, since RTX xx60 and xx70 as well as RX x600 and x700 are more often bought and the border line when higher CPU power would really matter would be more visible.
To some degree tests like this are set up to be a bit more academic, in that we want to see how these systems behave at the limits. The 6650XT being generally regarded as a great 1080p card makes a better starting point. The data is still helpful though for anyone buying a 6700XT, combined with previous GPU comparisons you can interpolate a pretty good idea of about where it would have benchmarked.
@@kenshirogenjuro873 I just meant on the whole, with as much data as there is showing how amazing of a deal the 6700XT is (cue every single monthly GPU value video since its launch), there's *shockingly* little attention given to it. I'd have figured that would be used as well. Not hating, just find it interesting that they point to one thing being super high value, but never show it in any video ever when they do these comparisons.
Without even watching video - with my budget PC (RX 6600 XT) it wasn't worth waiting for Zen4 anyhow when I owned AM4 (with R5 2600) already. For higher end GPUs (and generally systems) and new builders - there is a bit different price and value perspective - where you can justify paying quite a bit more, for not so much more performance. Also - it's a bit different situation now since Zen4 launch prices fall a bit of a cliff. Also with cards like RX 6900 XT and up, it would be silly to to go for very budget CPU as R5 5600 which is just great value CPU for people on budget systems.
These are the videos more people need to pay attention to. Too many people want to argue about one CPU being 5% faster than another without thinking about the fact that's only when it's paired with a $2000 GPU. It's always important to build your system evenly, otherwise you're just straight wasting money.
Could you guys do a test like this for 4K gaming too? I'd love to see the difference between a 5600 and 7600/7900 running a 4090, especially in CPU bound games like Spiderman Remastered.
funny how Spiderman is very playable even on my i5 3470. i get 40 plus in city swinging at all maxed rt off. 60 locked inside building where missions happen. with rt on rest on high i get very playable 30 locked. that cpu is a dinosaur from 2012
4k is GPU bound in virtually every game. 1440p is GPU bound on low-end and mid-range GPUs, so CPU will only matter if you have a high-end GPU where 1440p can start being CPU bound. It would be a large time waste for Steve to run those tests.
@@DarthChewie for instance, I am thinking of upgrading to 6800xt, I use 3440×1440 ultra wide display. Will I need to upgrade my 3600x or will it be sufficient for single player RPGs with higher quality graphics
@@Selandr Is it a 1080p ultrawide or a 1440p ultrawide? If 1080p, then it might depend on which games you play if you would become CPU bottlenecked with a 6800xt and 3600x, but on a 1440p ultrawide you'd definitely be GPU bound. Either way, I imagine that a GPU upgrade would still yield a good uplift in performance, even if you were leaving a few % on the table as some games start to become more CPU bound on a 1080p ultrawide.
Very nice as usual , I upgraded from a 1600 to 5700 and with a 2080 and this was like end of line facelift for my pc I understand after this there is no point in buying any older parts , literally a new system build from that point for me and planning on 7700x with a xtx for my next pc
I am fairly certain that if you play at 1080p, you will not have to upgrade in the next 10 years if you have something from last generation (amd 5600, intel 12600, 3060 or above, or similar graphics card). I think we have mostly plateaued in terms of graphic fidelity and resolution is the only frontier where they can actually make enough gains to sell newer products. The number of people who gain something from having 200fps instead of 160fps at 1080p is probably very limited.
Thank you for this video. Helped me to decide on what to upgrade to from my i7-4770 to pair it with a used 2060 super I recently bought. Looks like Ryzen 5600 would be more than adequate for my case. And it should see me through to a potential 4060-like gpu upgrade in a few years time.
if you're using a 3090Ti or 4090 for 1080p, you're a bloody nutcase. This isn't an issue whatsoever if you are using the card with a resolution appropriate for the horsepower on tap.
Would love to see scaling on this as well. I'm hoping to skip AM5 and keep this 5800X3D until next socket is out. Currently on a 3070Ti and thinking I might upgrade to the next gen when they appear in a couple of years maybe. If the current price insanity keeps up I can at least spread the cost more and get a new GPU in between the mobo+cpu upgrades.
thank you for again doing CPU benchmarks with a "normal" GPU. Too many people forget that unless you ahve a top of the line GPU you are likely going to be GPU bound in every game.
Yeah, I'm only using a 12100F with my 3060 Ti/2560x1080 resolution and in most AAA/new games I'm GPU bound anyway since I tend to crank up the settings as long as I have the GPU headroom for it.
I don't play competitive games tho so I can't talk about those, but in any other game I've played/playing atm I see little to no point in upgrading for my use case.
Even with a top of the line GPU, if you play on maximum graphics settings and resolution higher than 1080p, you'll also be GPU bound.
CPUs are just not gaming parts.
@Garrus Vakarian The Intel 12100 and 13100 proves you wrong, especially with high details and 1440p.
@@garrusvakarian8709
That is simply not true, in almost every actually GPU demanding game its my 3060 Ti that maxes out first and not my 12100 at my 2560x1080 res.
Currently playing Cyberpunk and my GPU sits at 98-99% all the time on high-ultra settings with some RT on top and DLSS Quality enabled.
Same story with Spider Man and God of War or any AAA game from the past few years.
Spider Man is probably the closest I've seen both maxed out at the same time with RT enabled.
@Garrus Vakarian depends which i3, the alder lake one is slightly faster than a 3700x so it can easily saturate a wide variety of gpus
I understand why these tests are all done at 1080p, but I would still like to see even just a quick summary of 1440p or even 4k just to further drive home the point that, once you go past 1080p even to just 1440p, the CPU starts to matter A LOT LESS.
It really depends on the quality settings and the game itself. You'd see much the same in the heavily CPU limited games such as CSGO and ACC for example.
@@Hardwareunboxed Exactly. I would often reduce the graphics settings on 1440p to improve performance somewhat with my 6650XT, so I get slightly worse visuals (but higher fidelity just because it's 1440p). The effective CPU load would be near the same as 1080p high. I'd always try to target 165fps and that doesn't care much about GPU (as you can tone down settings) but does care about CPU as in some games you WILL hit a wall regardless. A lot of people are sleeping on the fact that a CPU bottleneck means you're SOL and with a GPU bottleneck you have a plethora of ways to improve performance by tweaking settings.
Actually, maybe you could do a short follow-up to illustrate this, e.g. 1080p high/ultra vs 1440p medium. It will depend on the games, but just a couple would validate 1080p testing.
@@Hardwareunboxed You can keep the tests 1080p, but you really should enable ray tracing in games that support it. It's by far the most CPU choking setting I'm aware of in games like Cyberpunk.
KSP requires a powerful CPU regardless of resolution...
because this is CPU scaling BENCHMARK, the point is to show here where you gain absolutely nothing from higher end CPU. Naturally, absolute majority of RTX 4090 owners would play at 4K - but still, you won't be pairing it with budget R5 5600 anyhow, because who pairs $2000 GPU with $140 CPU?
I find the 6650xt results far more informative as someone rocking a 5600 is far more likely to be using this level of gpu.
Also nobody using a 4090 or a 6950 XT would be gaming at 1080p, so those results are more or less irrelevant. I would like to see how much difference there is between the CPUs when running a resolution that is appropriate for the GPU. Most likely you'll find the 7600 is a waste of money in the current situation (I know there is always this mysterious "future-proofing").
Precisely my case. It really showed me that the upgrade is just not worth it.
I'd like to have seen more Nvidia cards there, but the upper range of cards is kinda useless to 5600 users. Maybe the 4070ti. It's so annoying not having a bloody 4060/rx7700 up there to compare against, mid range cards. Overclocking the 5600x can give 5950x levels of speed though.
@@igaovieira its not worth it to upgrade last year build but its worth to upgrade to am5 if your build is lower than am4.
@@Alvin853. It’s not so much “futureproofing “ but more like it shows what it would be like a couple gens down the road. Also, some might save on the cpu and put the money into the gpu.
When the 5600x first came out in November 2020, my first impression was that it was a phenomenal CPU beating even the i9 10900k at gaming and matching the 10700k in productivity. The 7600 has overtaken the 5600 by more than 40% better performance in some games and matching the 5800x in productivity. This is what competition brings and we really need to see this kind of progress in the GPU market.
What game give 40%? Maybe in some game + 4090. Other than that 5600 are just fine.
@@nikosensei1258 Modern Warfare 2 (+42%), The Riftbreaker (+40%), A plague tale: requiem (+36%), Spiderman Remastered (+35%), Cyberpunk 22 (+35%),
The improvement largely comes from TSMC 5nm process. Just like Apple bragging about their M chips every year when it was TSMC who did the hard part
@@nikosensei1258 The multi core score in cinebench R23 of a maxed out 5600 was 12400ish points, with a maxed out 7600 you score 16350 (higher than a maxed out 5800x which maxes out in the low 16k range with a solid overclock. Pretty impressive for having 2 less cores.
and with what GPU?? people who get 5600 wont buy 4090/6950xt GPU@@dragonl4d216
A good comparison would be the 5800X3D, 7600, 5600, and 3600 at GPU limited situations (1440p) to see if upgrading to AM5 is worth it. Currently using a 3600 and a 5600 XT. Considering upgrading my GPU to a 6800 XT and sticking with my R5 3600 until prices come down.
Why not just grab a 5600?
@@MadClowdz brother has it already. Better to upgrade to a 5800x3D for a more substantial upgrade but not soon. Too many non gaming expenses. Having a working rig is already a luxury.
@Andre I don't think the price of the 5800X3D will come down until the 7000X3D series comes out. Heck, it's already going out of stock at times since it's so popular.
Yeah, I'm really tired of all these 1080p benchmarks to sensationalize the results. We need more 1440p benchmarks.
@@keithgnx 1080p is better to check performance of the cpu, at higher resolutions the difference is lower
My 5600 set-up cost:
R5 5600: $92
Asus Strix B450 Gaming F II: $90
32GB (2x16GB) Silicon Power 3200 CL16: $75
TOTAL: $257
My 5600 cost (new)
5600: $120
16gb 3600mhz Corsair vengeance: $35
B550m ds3h $71
Total: $225 but I got less ram
My 2022 5600 cost :
5600g 355$
Asus Strix B450 Gaming F II 251$
2x16 corsair 3200mhz 125$
TOTAL: $731
7800X3D new $220
Gigabyte Aorus Elite AX B650 open box $88
Gskill 32gb 6000MT / CL30 $93
$401
I was totally going to upgrade to 7000 series. But after seeing it... The price .... I bought a 5800x3d to pair with my 6950xt. Very satisfied.
As an "average consumer" (no high end parts, 3+ years between upgrades) those benchmarks are by far the most interesting to me.
Comparing CPUs with 4090s is interesting for sure, but testing and comparing low to mid range systems is the most relevant when you seak reasonnable "cost per frame" upgrades.
So thanks again and please keep it up :)
cheers
You also need to keep in mind that the higher performance of the 7600 with a 4090 does in fact apply to casual gamers too. Reasoning for that is future GPUs. You might not see a 20 - 30% increase with your current PC setup, but in a few years time when it is time to upgrade your GPU the more budget cards will be closer to the 4090 performance, which would mean you wouldn't need to upgrade your CPU/build a new PC.
You are not an "average consumer". You are a mid-high one.
I am an average consumer. I got a low end GPU, a low end GPU (tho, that was from my mistake, I could had gotten a better one) and an 8GB ram (while at the time, 16 was considered HIGH). And my PC build was from almost 5 years ago in 2018 and I don't have plans to upgrade very soon.
Just built a Zen 4 PC with my nephew. He got a 7600 paired with a b650 200$ mobo and a 140$ set of 32gb 6000mhz 36-36-36 memory. The whole build was 1200$ before sales tax with a 6700xt, and he’s happy as can be playing all his games in 1440p high settings. Was a HUGE upgrade over his hand-me-down 1660ti laptop he’s been using for for the past 2 years.
Beast! That’s last years, just upgrade the GPU in 3-4 years time. Just built my daughter a Ryzen 5 5600, had an old AM4 board kicking about with 16GB DDR4 3200 and my sons old 3060 12GB and it gets on the Red Dead 2 Benchmark Avg 59fps and low of 39fps on a 3440x1440p monitor. 👌🏻
@@papalazarou7880 Check out Hardware Unboxed’s optimized settings for Red Dead.
@@papalazarou7880 Yeah, I’m really happy with the build and the performance for the money he spent, but I’m even happier that he’s being responsible enough with his money that he’s managed to save enough to buy himself a decent used car and a modestly specced gaming PC since he graduated HS last year. Had fun teaching him how to fix up his car, and had more fun teaching him how to build a PC. He’s a good kid.
Same setup ordered few days ago - coming from AM3+ 😄 - kinda future proof.
Listing RAM speed. That’s how you know someone at least knows something about PCs. Most people laser focus on the CPU and GPU only.
thanks for the great tests! I upgraded from r5 1600 to r5 5600 on the same b350 motherboard I had and I dont think ill upgrade for quite a while until AMD comes up with at least double the performance boost. kudos to AMD for the great value motherboard chipset from 1st gen.
Get a 5800x3d. If you game at 1080p then you will see a BIG difference.
Same here bro. I saw massive performance boost and full utilization of my 1080 Ti. Next build will be of CPU with double the performance of 5600.
@@JamesSmith-sw3nk yes that is true. But I was talking about R5 vs R5 generational boost.
Yeah upgrading should only be worth when changing for double the performance at the same price, using cpuz as reference the single core performance for the r5 1600 is around 370, while the 5600 is like 609, meaning that the difference wasn't exactly double the performance but close to it and at least the 5600 was cheaper than the 1600.
@@rawdez_ Because of the same reason both nvidia and amd have prevented the generational progress on the low end and even mid low, just look at amd selling that 6500xt that performs exactly the same as the 480/580 or /470/570 on pcie 3.0 and the 5500/590 in some games for the same cost despite the 5 years difference, nivida does the same the 3050 should have performed like the 2060 and the same 3060 should have performed like the 2080 and not like the 2060s those companies don't want to release strong performance for the low segments cause then nobody would upgrade their gpu's if the low ends are capable of 100 fps at high settings, so they prefer to keep low and mid segment at stagnation so you are forced to change your gpu at the next gen.
Those 6650 XT results are truly an eye opener.
Fist time realising that if you dont have a top of the line GPU, your CPU doesn't matter much? If that graphs was 1440p they would all tide xD.
@@neoperol not 1440p, but 4K would make them a lot closer. If not almost same.
@@eugenijusdolgovas9278 1440p already gets a big shift from cpu to gpu bound doe
Too bad they didn't include other popular graphics cards like a RTX 2080, 3060 Ti or RX 6700 XT!
Also the testing method of "only 1080p" and medium details (for example CP2077) is ausgelegt to be CPU bound!
I don't know about you but when I invest in something like a RX 6700 XT or 3060 Ti (from something like a GTX 1060, RX 580 or 1660 super) I would definitely crank up the details to high or even ultra if not go 1440p and step up to "next gen" performance (PS5 ish with 1440p, high details) targeting 60 fps+ in story games like GoW, HZD or CP2077.
That takes the focus back from CPU to GPU only!
Hardware unboxed had another video with comparison of Ryzen 2600, 3600 and 5600 which were similar.
@@eugenijusdolgovas9278 Most gamers still don't target real 4K. 1080p is still a thing.
would love to see a scaling benchmark video featuring previous flagship cpus. would like to see my 9900k vs current flagships.
ya me too, i have a i9-9900KS and a 4090, playing at 4K. I really want to see some benchmarks to see if the CPU is worth upgrading at all or not.
I think it is better in the future to also add a more in between card. 6650XT to 6950XT has a large performance gulf. Maybe a 6700XT and/or 3070 should be enough to fill that gap. Also preferably to also test 1440p since it is a popular resolution, thus can give the viewers better information on deciding whether to upgrade their CPU or GPU when playing at that resolution. I picked 6700XT/3070 instead of 6800 or 3080 because afaik those GPU will still be CPU limited at 1440 most of the time, thus 6700XT/3070 should be the better choice for more variety on the data.
In addition, I am very interested in both VR Scaling and Ray Tracing scaling with CPUs... I would love if HBU could throw just one VR and Ray Tracing bench into the CPU testing.
yeah always though the 70-cards are the sweetspot so I'm surprised they didn't go with RX 6700XT/3070 here
I have 6800xt with 5600 (with pbo2 though, basically a 5600x but with better all-core boost), in majority of cases I am GPU limited. The only exception in my use case, apart of 2d games, is Path Of Exile. I play at 1080p, though.
Lot of us still using 1080p.
To add a data point to the GPU/CPU bottleneck question I offer my anecdotal experience with my 5700x and RX 6750 XT Red Devil. I play at 1440p with as high settings as I can and am GPU limited. Not sure what 1080p would look like except max settings would probably still give high FPS. However, at almost max 1440p graphics with RDR2, Baldur's Gate 3, Horizon Zero Dawn, The Callisto Protocol, and World of Tanks I simply don't care if my rig isn't giving me all the bells and whistles at 100%. Why would I with average FPS of ~165 on some of those titles? Besides I'm getting too old and way too beat up to be competitive in 1st person shooter type games.
I would really appreciate 1440p results as that's the harder to answer question on paper. At 1080p a newer CPU will obviously benefit but if I'm running games at 1440p or 4k how much would I really be getting out of a new CPU?
I wonder about it too. I have 7900XT with 5600X, what am I gaining by switching to new gen if I game at 4K? I am already getting 60+ FPS in most games with high settings.
What about 5800X3D, is it that much better than 5600X in 4K?
in higher resolution the load is even more on the gpu and difference will get even lower.
@@freaky425 That is the point though, will the lesser difference be worth it is the question
@@dougalmacburnie6510 It's negligible, the amount people try to achieve by upgrading to a better cpu is usually lost by overlooking the fact that you do need the fastest possible ram also as shown on benchmarks.
That's pointless, as it would be entirely dependent of what GPU you have. A 4090 might still be greatly bottlenecked by the 5600 at 1440p in some titles, but this doesn't mean your GPU would.
I've just built a new PC based around a 5600x CPU. My board may be ddr4 and pcie4 limited but I feel like there is heaps of life left in those two specs. My upgrade plan when I go to buy a new GPU in a few years time is to find a good 5700x/5800x/5900x on the second hand market as I'm guessing new ones would be all but gone. I'm taking a bit of a gamble here also in noping the they won't be too expensive either on the second hand market
Can we take a moment to appreciate the extra time it takes them to do the GPU scaling for these videos? It's amazing real-world valuable info that many people under-appreciate
chat goes so fast no one noticed I am 37 yo VIRGIN
They under appreciate it? Isn't that why we are all watching this video?
Testing with these high tier gpus at 1080p just feels pointless because pretty much no one buys a 6950 xt (let alone a 4090) to play at 1080p.
Think it would be much better to keep the 1080p results with 6650 xt but include 1440p and 4k with 6950 xt and 4090. I know the cpu matters less at those resolutions but at least you would be able to see how much difference the upgrade makes.
Also dlss/fsr should really be included at this point
I'm impressed with the 7600 but can't wait to see the 7000x3d's. I just got a 5800x3d a week ago, and with a 3080 at 1440p I think I'm good for a few years.
Hell I have a 5600x and a 3080 now (up from a 2070 S) and the difference is night and day. At 1440p, I too think I'm good for years now. Or atleast I better be. Genuinely looking like we're plateauing on performance boosts these days, but can those gains outpace the poor console ports we keep getting is the real problem.
Im in the same boat, I get my 5800x3d delivered today (upgrading from a 3900x), and I have a 6900xt and game at 1440. I think I'll be happy for a few years.
Hopefully, the market will come down by then though lol.
@@Dynasty1818 I had a 5800x before which is very close to the 5600x so I know that works well with the 3080. I knew the 5800x3d was a lot faster but what I didn't expect was how it's pushing my 3080 really hard now. I had to bump up the fan curve because it runs about 10 degrees hotter now and plays so much smoother in heavy AAA games. I recommend for you to eventually get a 5800x3d if you can because it's amazing, but I would have been happy with the 5800x, it's just that I wanted to max out my pc and not have to upgrade for years and I think I've accomplished that.
5800x rtx3080 1440p gamer here, no need to upgrade anything. Even Witcher 3 runs 60fps with all RT options enabled.
@@mikeclarke3990 Nothing wrong with the 5800x at all.
5600 is a very good option when it comes to 1440p gaming because there is a much smaller difference in frames compared to 7600, we must also have results for 1440p here.
For 4k even better.
For 1440p and up your gpu of choice will improve your fps alot more than the cpu. Due to being gpu limited, not cpu.
@@jotunheim5302 i have 5600 and rtx3090, and i want to see how much i will get whit 7600 on 1440p, it is worth to buy or not.
@@jotunheim5302 unless he’s using RT the even his 3090 is capable of pushing more frames at 2k ultra than his 5600 can supply, though I agree does it really matter if he’s getting 150 vs 165? The more interesting question to me is how to measure the smoothness and stuttering that many games have with weaker cpus. Average fps and even .01 lows don’t tell that story. Minimum frames does and minimum frames is heavily cpu dependent, yet we never see that in review videos.
I’ll give an example. My old system was a 1300x and a gtx1060. Online cpu calculator tells you that is a great match for 1080p, very little bottleneck either way. Massive stuttering at ultra settings. Recently I change out 1300x for a 5700x and everything else same. Frames are still low 30s average because gpu is maxed out, but the stuttering in hitman and assassins creed is now gone and it’s now a smooth 30fps and playable. That’s the power of a good cpu and what all the reviews ignore. Sure it doesn’t matter now when your system is new, but it matters to those who want to stretch their system life as long as possible.
Great content!
I would love to see graphs which show where any of the graphics-cards reach the cpu-bottleneck.
In other words: What cpu do I have to use at minimum with which graphics-card to not hold back the card.
For all AMD and Nvidia-cards...
Its very dependent on the game your playing. Most games media benchmark run great on any modern cpu. But other games that are less "mainstream" is where cpu performance really makes or breaks your game play.
@@ramair325 it's more about having a tendency I think, as we clearly saw with these benchmarks, that there is usually a huge improvement going from 5600 to 7600 with a high-end graphics card.
And then, you would have to define what is your target framerate/resolution that, if not met, would be considered bottlenecking. Personally, I would say 1080p@90fps average in order to get at least a 1% low of 60fps, but I am pretty sure the demand range would scale all the way from 50fps to 200fps average.
Going by the charts, even a 6650XT is plenty for me, making any CPU viable, meaning, for me, we can't be possibly talking about CPU bottleneck at the moment.
@@sassuki ya and thats my point, these benchmarks make it look like cpu performance does not matter. but what they dont show is how bad "81fps average" can really be in a game because at some point in that game you might be getting 140fps but at other points you maybe down in the 30's. And most games are very cpu demanding in many places if not most places other than a few AAA games that have been really well optimized. games can also be very gpu demanding in many places as well but at least with a gpu bottleneck you can reduce game settings and gain massive fps boosts. its not really the same thing for cpu's. This is also why i hate these type of reviews as it shows the 7600x as the god gamer cpu when in reality the 5800x3d is going to be far better across far more games. dont get me wrong i know why steve is doing these benchmarks but its just making the water muddy for consumers.
@@ramair325 Naah, it's not that bad, you're overcomplicating it; the rule of thumb of "Average is usually about 1,5x the 1% low" holds true. Some exceptions are games like Fortnite where the average is roughly 2x the 1% Low.
So even if you take Fortnite into account, nobody should need more than 130fps average to get a minimum of 60fps, which is enough to get smooth gameplay even on old shitty 60hz monitors with v-sync. With a 1st gen FreeSync monitor, even 100fps would be enough to keep the gameplay tearing-free.
So at the end, there is no CPU bottleneck with any of the CPUs presented here, just overperforming GPUs for overinflated egos, that need to justify why they just spent a 1000 bucks on a f*cking graphics card!! lol
People tend to forget where we came from: in 1996, you needed a special 3D card to get solid 30fps in Tomb Raider... at 640x480! Talk about some demands nowadays: they get 3 times the framerate at 6 times the resolution and dare talk about bottlenecks! lol
@@sassuki it can be far worse than that in some games. and thats my point. the games used for these tests do not represent a worse case scenario rather quite the opposite.
4K results please. I have a 4090 and obviously I'm not using it for 1080P, would like to see if its worth upgrading my cpu.
The 6950xt is a monster
I think it is the best "pro gaming" GPU right now. As long as you can still buy it :D
Every CPU Upgrade I do usually includes an increase in both Thread counts and clocks when some game just starts being too limited.
I switched from i5 4690 to i7 4790 because of Battlefield in 2017 and to R5 5600X in 2020 because of Watch Dogs Legion and CP2077.
So I think this Gen the 5600 is definitely still a keeper. (I am also using a 2080 ti)
Maybe new UE5 games will change that but until then I'll wait and probably do a completely new build when Cards with 4080-equivalent Performance are available for 500/600$.
It will be a long time til then but as long as the PS5/XSEX are the current console generation, the 5600 won't be obsolete.
Well, we don't have 3080 equivalent for 500/600 and you want 4080 which is like 50% faster. Good luck with waiting.
@11DaT11 the 4060 ti is supposed to be on 3080 performance but apparently it’s just a 3070 lol which is kinda a joke considering a 3060 ti from last gen is basically the same as a 3070 maybe 5-10% slower
@@Amzyy If 4060ti does have less than 10Gb of VRAM it's not on 3080 level period. 8GB on 500+ card is unacceptable.
@@dat_21that too good point mate people need to wake up and not get scammed
@@dat_21 Yup. It's gonna be a loooong time.
Really like this Scaling benchmark series, would like to see more videos. We get to see the real world usage of budget CPU with appropriate budget GPU too. Thanks Team HUB.
Thanks for testing more typical sweet spot configurations! Comparisons using sub $2000 gpu's are so much more useful for the vast majority of us 😂
At 4K high/ultra settings, except for some older games and FPS, most games I'm under 100fps with my 3080 12gb. So for me the upgrade from my 5600X doesn't really make sense, my 3080 is always pegged at 100%. Point being, unless you are CPU bound or just mildly GPU bound, whether it's at 1080p with a 3060 Ti or 1440p with a 3080 or 4K with a 4090, you shouldn't really be looking at upgrading from ZEN 3. At 100% GPU utilisation performance will be the same.
true , maybe your 1% lows would be a bit better, but thats such a minimal performance increase for so much money. you are better off just buying a better card :p
@@Bastyyyyyy agree unless you mostly play multiplayer or cpu imtemsive simulation/sandbox games like cities skylines and teardown cpu upgrade is pointless unless the system is too old that it would bottleneck triple a titles which is not normally cpu heavy, unless that triple a games is terribly optmized like gotham knights,
Knowing that I don’t need a newer cpu to pair with my 3060 ti makes me happy. I’m using 5600 and it’s been the best upgrade I did from 2600. 5800X3D would’ve been great but I’m already satisfied with performance I get till another new gen consoles hit. (if newer PC AAA games are optimized…)
Same but I'm on a 3600 and can't see the benefit of going to a 5600 playing at 1440p
I upgraded my old R3 3100 and gtx1060 3gb pc on Christmas. Using old RAM and a520m motherboard, I got R5 5600 and RTX3060ti for like $400.
The money well spent. The performance were jumped, like doubled or even tripled the FPS, and now I got a PC that I can play games easily for 3 more years.
Would love if you would do the GPU scaling test on 1440p as you also say in your videos that its the sweet spot for most gamers.
Yes! Would love to see a test at 1440p and 4K for different CPUs such as 5600X, 5800X, 5800X3D, 5900X, and the Zen variants, but also Intel variants and see how they stack up when using 4090. Of course, it is a lot of work, but these things can affect buying decisions.
I’m assuming when we move to 1440p and 4k, the differences between 5600 and 7600 would reduce further. Which means if I’m on 4k atm, I really don’t “need” to upgrade.
Thanks for the content!
It depends on your GPU and the game. I have a 4080 and had a 5600x and I was CPU bottlenecked in some games, even at 4K.
@@johnboylan3832 what games do you mean? On high settings 4K you dont got much further than 100FPS anyway and in most games, the 5600 is enough for that
If you play simulation or MMO type games CPU can still matter a lot even in 4k, but yea it depends on the games you play. For most, it should be ok.
@@otozinclus3593 - Cyberpunk, AC Origins, AC Odyssey, Far Cry 5 all were CPU bottlenecked in certain areas at 4k/high settings on a 3090ti with a 10900K which is why I moved to a 5800X3D. For example, Alexandria and Memphis in AC Origins and many areas like Athens or even in the wilderness in Odyssey bottleneck pretty severely well under 100fps on the 10900K even at 4K. The bottleneck goes away on a 5800X3D. So even at 4K, CPU choice definitely matters if your GPU is strong enough.
@@otozinclus3593 The two most recent I was playing after I got my 4080 and before I upgraded the CPU were The Witcher 3, which was about 35-40 in the towns; Spider-Man was 65-80. The former is now 45-60, the latter 100-105. I upgraded to a 5800x3D. The 5600x struggles a bit at 4K with ray-tracing.
Please do a CL timing test for DDR5, we need to see the CAS latency has an impact on gaming performance
watch buildzoid's (actually hardcore overclocking) "the cas latency timing doesn't matter as much as you think it does" video
@@tyre1337 thanks for the suggestion, I will check it out
I went for the 7600x. Pretty happy with it so far. The system was a bit pricey for sure but its hopefully got a long upgrade path ahead of it.
@@terraincognitagaming No they don't you are projecting.
The ability that he _might_ be able to slap on a 9000 series CPU into his AM5 board is what he is really referencing. That kind of longevity is not common at all with intel, and if Intel stay true to the course, a Meteor Lake CPU will need a new platform. In fact, one will need a new board for what comes next after Arrow Lake if Intel keeps the 2 generations per socket march. It is conceivable, that AMD may even release 10000 series CPUs on AM5, but I do have my doubts about that.
All that being said, I have always felt that by the time an Intel system is considered _"too slow"_ for most, it is probably time for a new motherboard anyway, case in point how long has the 9900K been working for *many* users? But the thought of putting 3 maybe 4 generations of CPU on one board is now very important to a lot of consumers.
Also, AM4 and Socket are for all intents and purposes a dead end, no more upgrades for them.
@@terraincognitagaming I honestly don't understand people like you. Have you lived away from civilization the past few years? Here in the comments, people wrote hundreds of times that they changed their Ryzen 1000 and 2000 processors to 5000, while receiving a huge performance boost for quite modest money. I have personally done the same on two different systems, which are noticeably faster now.
In fact, even switching to the Ryzen 3000 series was already a significant step towards increasing performance, especially considering that the maximum number of cores has doubled - from 8 to 16, and all this without junk cores, like Intel has - Atom cores never intended for desktop solutions.
@@terraincognitagaming I don't think anyone is really _rockin_ that CPU in modern games and certainly not anything where they want to be ultra-competitive online.
That CPU is the perfect example of a CPU that is *too slow* and the platform it is on too old for the newest and most demanding triple A releases. Nobody is putting a GPU of any substance on that 9-year-old CPU, well they shouldn't.
Sure, it may be good if you want to play games from 5 years ago, but it will get decimated by a 12100/13100 in today's software.
Don't get confused with casuals holding on to hardware because they don't play demanding games or games with higher fidelity vs enthusiasts here that hold on to hardware for a few years with a GPU upgrade along the way.
@@terraincognitagaming Dude, nobody but ultra-casuals are holding on to their system for 10 years to build or buy a new one.
Enthusiasts will have a GPU upgrade along the way and usually build a new system within 6 years. This is why I have seen a number of 9900K users talking about their build either this year or for Meteor Lake, it is right around that 6 year mark.
You are talking extremes that do not reflect the mainstream enthusiast level. That is one reason that having a platform that can accept 3 CPU generations or maybe even 4 is such a positive for them.
@@terraincognitagaming 4770k is dead.
My 5600x is perfect for my 6650xt for 1080p gaming, no need to upgrade for now👍
Float plane member here and glad to contribute ever since the green team silliness a while back. Keep up the good work. I appreciate independent tech coverage.
I think the review suffered from not including 1440p data.
I think the RX 6950XT data will look even less extreme in terms of differences among the processors.
I went with the 5600 because I mainly needed a second PC besides my notebook for as cheap as possible but at good performance. This worked great for me because I haven't had a desktop PC for many many years, so the jump in performance from zero (well, actually Core 2 Duo) to Ryzen 5600 was huge! :D
I also preferred the more stable and known platform over the brand new one with DDR5 and the new processor generations. And buying into the new processor generation (for example Intel 13400F) but sticking to DDR4 seemed wrong. And DDR5 is too pricey for now. Also, the new Ryzen CPUs are still too new to have good prices and I wanted a Ryzen and not an Intel for now. And the Ryzen 5600 seems very good overall.
Also, I usually don't do partial upgrades of my PC hardware. So, the end of life for Socket AM4 is no problem for me.
I get this was to show CPU scaling, but it would have been nice to see different resolutions as well to provide a large overall picture of the importance of CPU performance as I don't think most people are building new PCs with 1080p in mind.
PERFECT!!!! That's exactly what I needed!
You should mention to buyers that, for higher resolutions (all other things Equal) the differences % in FPS will be even smaller, for example, i am looking for an 1440p 34" monitor, the difference in the 5600 vs 7600 its even smaller and in my opinion its better to get the cheaper option and invest that money into a better GPU, so this 200$ difference in your example should go to a better GPU , to run the higher resolution monitor!!
Thanks Steve!!!
My monitor was a ROG 27" IPS 165hz unit. I replaced my 4K TV with a Hisense 55U8G which is wall mounted. My GPU was an EVGA 1080Ti FTW3, 5800X, X570 and 4x8 3600cl16. My TV is now my monitor and my ROG is back in its' box. Yep, no RT with a 1080Ti, and with "reasonable" eye candy settings, the 1080Ti performed very well with what I do. WoW, elder scrolls and GW2. If you're thinking of 34", wall mounting a 55" (U8H $700 USD via Amazon) could be an option.
Like the video, but I’d really like to see the 4K tests that others have requested. My particular question is: how much, if any, performance would I be leaving on the table by running a 5800x with the 4080 or 4090 at 4K?
This is exactly what I want to know, I also have a 5800x
So at 1080P you can build an AM5 system with a 6950XT for the price of a 4090 alone and get similar performance to an AM5 system with the 4090!
I upgraded my Ryzen 2600 to a 5600X and my RX 580 to an RX 6700 XT for a total of $300 around 2 months ago. It's like using a brand new build and I'm still on my Tomahawk B450 MAX motherboard. No regrets.
As a 3060 12GB user there’s clearly no benefit to upgrading my 5600X at this time. This video is was very helpful.
Yeah that's a capable combo for 1080p-1440p for most titles.
Recently got a 2nd hand 6600xt for an AMD combo with 5600x, and it looks like I don't have to spend any more money on a PC this year.
Thanks for the benchmarks!
This is the reason why I always return to this channel. Tests considering real world use cases are top notch and well taken in to consideration.
If you don't have a powerful gpu, the upgrade to 7600 makes sense in a couple of years, when the prices will be discounted.
This has been one of the most helpful videos for my use case.. 6600 xt @ 1080p and mostly single player games... the upgrade ATM would just not be worth it.
Really happy with my 5800X3D/3080Ti combo. No need for me to upgrade for quite some time.
That’s a perfect 1440p combo right there mate
@@Amzyy yes you’re right mate. That’s exactly what I do - game at 1440p and it’s brilliant.
I really wish the 3600 had been included here as a baseline. If someone is pondering whether a 5600 or 7600 upgrade would be worth their money, being able to compare either of them to their current CPU would be very useful. Perhaps the best upgrade is no upgrade!
For instance, I have a 3600xt-based system. And a 120Hz 4k display. (I currently do 90% of my gaming on a console.) If "good" GPUs ever come down in price, I might actually want to buy one. Do I even need a new CPU? Probably? Maybe?
I'd like to see the 5800X3D included in these comparos as well.
3600 is a pretty good rough proxy for current gen console performance. It should hold up for a while yet for a midrange rig, but it's not a rocketship for sure.
I hope you reach 1m subscriber this year! Love your content! Do you still feel like cpu don't make much difference at 4k still? i feel like its starting to make a difference considering how fast all those new CPU are catching up to the GPUs over the past few years, not as much as lower res but still , would be interesting to double check!
Let's get some love for all the people playing at 1080p in 2023 big up to those people
I would like to see 1440p and 4k scaling as well with new processors to see if there is any real improvements. I do understand you will be mostly if not all GPU bound but sometimes there is a difference depending on the game.
It would tell anybody that you dont need to upgrade a cpu @ 2160p and mostly 1440p, not in interest of Hardware Unboxed, the goal here is to keep all hype alive, clicks and visitors count at any cost. These guys are money oriented as all youtube, just more truthlike acting.
Thank you for benchmarking GPUs and CPUs for the general public. Just look at steam's hardware stats to see what hardware the general public is using. Many tech youtubers live in a cloud and only benchmark with the top of the line. There should be more video testing more modest hardware
modest hardware like RTX 2070/ 3060, RX 6700 XT or RTX 3080.
Why does the 6950xt have uh.. 80% better 1% lows in hitman 3? That's an insane difference in consistency
Seems to be a driver issue with the RTX 40 series.
It would be good to show a couple higher resolution examples or mention it since that really does equalize differences in CPUs even more. So if you prefer 1440, then you'll see even less of a difference between the 5600 and 7600. 4k even more so, but requires a beefier GPU anyways which may make you want a better CPU to utilize the GPU better when changing settings or playing less demanding games.
Would be intersting if you make an analysis about which GPU class is start to make difference between 5600 and 7600
My assumption is around RTX 3060 or RX 6700
You mean 3070, the 6650 XT is already in 3060 Ti territory.
As an owner of 5600 and RTX 3060, I can assure you that this card is not powerful enough for 5600 to introduce any significant CPU bottlenecks at 1080p, at least in (relatively) modern titles. Some 10 year old games limited to single threaded CPU load (like CS GO) might be exceptions though. But I would not call it CPU bottleneck in this case, as 200+ fps is hard to call a "limit".
I assume more like at least RTX 3080 or RX 6800 when looking at 1440p high details.
The testing is this video is also taken CPU-bound -ish since it's only tested in 1080p, sometimes even medium details!
But when spending so much money at a RX 6700 XT or better I assume most people would also like to play at ultra details (if not with RT) and 1440p to experience graphic gems like GoW, Horizon Zero Dawn, Witcher 3 or Cyberpunk 2077 in very high quality.
@@starstreamgamer3704 nice to know, my point is showing the diminishing return graph.
They already have the data.
When they reviewed 5600 they use should already have data when used by low spwc gpu to high tier.
Then make a graph of diminishing return.
Compare graph of diminishing return from 3600 and 5600, then you can compare the gradient of the graph and decide where is the sweer spot or where is the biggest performance jump
Seems easy when I wrote it but I know it is hard
I'm running a 13600K on an ASUS ROG DDR4 mobo with 4x32GB G.Skill Royal 4000 CL15 RAM clocked at 3200MHz and a fast NVMe SSD with a FE 3090 ($700 used from eBay). This system cranks all day at 1440P, so the value proposition of a DDR5 or AM5 platform just isn't there. Given that I paid WAY too much for this RAM ~3 years ago for an AM4 gaming system I was running, I couldn't not use it and I'm happy I stuck with it. For kicks, I put my GPU, CPU and SSD on an ASUS ROG Z690 MATX DDR5 mobo with 2x16GB T Force 6000 CL30 RAM and it had identical performance in Dyson Sphere Program and CP2077 as my DDR4 platform... However, running a 3080 and 5800x3D on an AsRock x570 Aqua motherboard with the same G.SKill DDR4 and fast NVMe SSD in a custom loop, the 1440P ultra FPS performance is ~20% BEHIND the 13600K in the games I play (regardless of RAM type on a Z690 motherboard). This is due in part to the 3080/3090 uplift I'm sure, but the 13600K is a beast of a CPU. Again, if you still have decent DDR4 RAM, AM5 costs a lot to adopt, but I completely agree that the sun has finally set on AM4. I think these tests really highlight the need for a current gen CPU with the RTX 4000 series cards and that for midrange CPUs, last gen GPUS are still a great value in price / performance. If you want to stretch your dollars, the 5800x3D is awesome, but I'd love to have seen it added to these tests to ensure the RTX 4000 GPUs show the same (expected) uplift at 1080p. I assume they would, but it'd be a great confirmation. Thanks!
Exactly the video I was looking for, thanks guys!! I want to move onto a new platform but want to keep my GPU because I love the way it looks and it already performs the way I want
Im curious: at what level of GPU performance having
a Ryzen 5600 would be a severe bottleneck?
Im talking about future GPU upgrades down the line.
It seems that the furthest a 5600 can keep up with is something like a 4070ti, which would mean a rtx5060 in a few years
Sounds fine to me
It depends on which resolution you like to play at. I like playing at 4k so my RTX 3070 bottlenecks my system much more than my 5600x.
I really appreciate this kind of scaling review. I want to go pretty high end with my next GPU, but I think I'm better off replacing my 5600x/b450 before doing so. Thanks!
The lowest end GPU results are the only ones that matter here. Any real world advantage of the 7600 make sense with a higher end GPU which immediately begs the question: if you can afford a top of line GPU, why settle for any of these CPUs? It’s a paradox that renders most of these comparisons moot.
It's the difference between a GPU vs a GPU, a motherboard, CPU, DDR5 RAM. Maybe you'll need a new power supply with this upgrade. Hell, let's just get a new case while we're at it and just like that you bought a whole new PC
7900xtx instead of 6950.
It would be nice to see how it scales with a weaker CPU.
Killer love this video. I have 5600 and haven’t struggled yet in any game but I do get the itch to upgrade . But 5800x3d just seems like better route for me if I ever do. Using 6700xt and 32gigs ram
As long as PS5 and Series X are the current gen consoles your 5600 will do fine.
if you are on am4, just get 5800x3d and enjoy it for now unless you want to upgrade in another 3 or 4 year...where am5 starts to make sense.
@@Felix_Schmitz well I agree but when I see specs on some of these new games like hogwarts and forspoken requiring 8 cores for 1440 p ultra it just makes me wonder if that’s bs or just terrible ports
@@hiimcortana1568 I'm on a 5900x. Will i be fine for another 5 years???
@@bmdshred77 Yeah, the consoles also have more threads than 5600 but I wouldn't upgrade until it is apparent that 6 Cores aren't enough for Console equivalent settings.
But I am using a 2080 ti with my 5600X so for me the GPU will probably be the bottleneck most of the time anyway.
thanks for benchmarking them with diferent gpus :)
I think most people who would go for 5600 are those who are building a Budget System.
And based on that idea and my local pricing (non-US), the GPU's (only considering GDDR6 variants) which would fall under the Budget category would be either "GTX 1650" or "RX 6500 XT" or "GTX 1660 Super" or "GTX 1660 Ti" or "RX 6600" (for people in my region, sub $300 GPU's would fall under Budget, not sure about other regions).
Got a 5600x,in 1-2 years i get the 5800x3d used for close to nothing 🤗
It would been nice to include a 3600 and 5800X3d into the mix. My 3700x seems to be bottlenecking the 3080Ti at 4K, and am thinking if I should upgrade. But not sure to 5700 or 5800X3D as there a massive price difference.
Really depends on what you play, for most games I play (primarily strategy type games) the 5800x3d vs 5700 would make a pretty big difference, but for many popular games it really isn't that different.
@@h.b.5577 I mostly play FPS and still see the GPU usage howering around 92% many a times with DLSS on as its rendering at lower res with DLSS. But I guess it will either have to be CPU or GPU bottleneck, they never go hand in hand, so I will leave it for now and see how it goes. I am getting good 100+ fps anyway.
@@coolvinay That's really what it ultimately boils down to, always going to be bottleneck-ed one way or the other. I recently upgraded my couch gaming pc from a rx 580 to a 6700 xt while keeping its 9600k which now often results in cpu bottlenecks, but if performance is still 60+ fps I don't really see the need to worry about it so much.
Thank God I made the right choice of not spending too much. I bought 5600 + 3060 ti and it's perfect for my 1080p 180hz monitor.
i am using ryzen 5600 with gtx 1050ti and in Valorant I am getting 600+ fps
Also got a 3060 Ti but went with 7600x after the prices dropped. Definitely doesn't make sense in games today (particularly as I mostly play at 4K!) so I've basically taken a punt that DDR5 / PCIe 5.0 will eventually become relevant in games. Will likely upgrade the GPU in two or so years, then the CPU further down the road (presumably AM5 will have an equivalent "swansong" like the 5800X3D).
TBH the main reason though was because I had a bad experience with my previous Intel system (which was very old but was still just about suiting my needs), the motherboard died and replacements just didn't exist at non-crazy prices, so I was without a working desktop for a few months. Definitely prefer AMD's approach of supporting a platform long term, instead of just for a generation or two. Buying into a new platform hopefully means I'll have good options for upgrades / replacement parts for the next 5-6 years.
that's going to last you for YEARS
@@tyre1337 Didn't feel like that when I tried running the Witcher 3 with RT on a 3060 Ti. Feels like some kind of conspiracy to get me to buy a new GPU.
@@liamness that conspiracy is RT, turn it off, on a 60 class card it's essentially a demo
2:20 thanks steve for asking yourself my never ending question. Keep these scaling contents coming!
While a good indication of raw CPU performance I don't think this video really works as an indicator of GPU scaling for the next couple of years. The 6650XT is a more entry level card and is still pushing nearly 100FPS across all title in this vide at 1080p and even 1440P gaming at medium settings around 60FPS is doable at entry level. The days of 1080P are clearly limited and going to shift to higher resolution gaming in the next 2-3 years now that high end GPU's are deliver 200-300FPS at lower resolutions.
It will take at least 1-2 generations of new GPU for a lower end card to deliver the performance of a RTX 4090, so if we really wanted to answer the question of this video "Is Zen 4 worth it" the testing should have been done at higher resolutions and see how much bottlenecking actually takes place and how the 5600 will performance in a few years.
Would be interesting to include 5800X3D to those graphs as the 'final' step of AM4
Check the Hardware Unboxed video from December 10th for a similar comparison involving the X3D 🙏
Its great to see the difference but would have loved to also see a 3070 or a 3060 there so that ppl on a budget that update a few years down the road could have used this video to make their choice
I ended up upgrading from Zen+ 2700x to Ryzen 5 5600 for in socket upgrade with my Asus X470-i on the super cheap. If I was to build a new Ryzen system then would factor in a higher core count 7900 series chip. Besides, I already had 32 gigs of DDR4 and didn't need to reset or start over to upgrade. Good review Steve on another excellently written comparison.
Wow. First i’d like to say great work. I really love these ‘scaled’ benchmarks. However, i’d recommend including a GPU in-between the 6650xt and the 6950xt too. Maybe a 6750xt/6800 or 3080? Because we can see a pretty noticeable difference in some games between the CPUs when we use a 6950xt or 6650xt - but I think it would be cool to see the result of something that sits in the middle of that spectrum too.
Just a suggestion! I know this stuff is hard work lol.
ACC scaling as usual is good craic, as always thanks for including it.
Hoping the 5800X3D will drop when the new X3D chips come so I can get it at a better price than it has been. It seems to have started dropping now (in Norway), so hopefully that will continue and I can replace my 5600X with it (using the 6900 XT).
idk about that. i think they will just thottle down supply for it to make space for the new x3d so the price wont really go down because a lot of people think about upgrading to a 5800x3d that are on am4 still. it could even get a bit more expenive imo but we will see
@@theplayerofus319 You might be right. I'll be monitoring the price. With the 5800X3D I can keep the AM4 system for even longer.
@@Alexandra-Rex yeah i got the 5800x3d right after launch and i can recommend it! Keeps am4 alife even after ryzen 7000
my building philosophy is: get the best gpu for your budget and pair it with the cheapest cpu that wont bottleneck, glad tô see that it still the way tô go
I swear the 5800x3D keeps being the best thing AMD did for AM4
it would be better to upgrade to the 5800x3D
I just got a 5600X a couple months ago on my AsRock b450m Pro 4 and I plan on sticking with that for at least 2-3 years before I even think about upgrading my total system. Paired with an RTX 3060 it handles everything at 1440p easily and DLSS is awesome.
Sat content with my 5600 and 1070ti. The 5600 will remain viable for at least 2 more GPU upgrades for me before either switching to 5800x3d or going AM5
At that point, I think the AM5 CPU will be best way to go over the 5800X3D.
Curious, how is your 1070TI holding up at the moment? I have a 1660TI, but it is hamstrung by a very slow CPU, a 5600 is what is on my radar as a drop-in replacement.
@@davidandrew6855 perfectly fine for my needs at 1080p 144hz depending on quality settings. Jumped from a 3100 to a 5600 which made a decent difference in FPS and overall PC feel
I do not see the point to make a comparison in 1080p resolution with the most expensive graphics card (4090). I doubt that if somebody is having such graphics card that he intend to use in for 1080p and chose the CPU for that....
You know there are 2 other GPU performance levels in the test, right? You *always* use the most powerful GPU and lower resolutions to see just how much performance a CPU has.
Interesting how the value of the 6700xt was listed as near one of the best options, but we rarely see it used as a midrange option despite its performance improvement compared to the 6650.
That said, I love these comparisons. C: You guys rock!
The RX 6600, 6650 and their XT models are often the best midrange cards with an attractive price point. When someone has more money spend I assume often goes more "all in" and chooses the RX 6800 or better right away.
I have the same feeling about RTX xx70 products! Many people buy RTX 2060/ 3060 or go 2080/ 3080 right away. The success or the RTX 4070 at the moment comes from a missing RTX 4060 Ti option, I guess.
But I would also see more midrange in those tests as well as with high details and not only 1080p medium, since RTX xx60 and xx70 as well as RX x600 and x700 are more often bought and the border line when higher CPU power would really matter would be more visible.
To some degree tests like this are set up to be a bit more academic, in that we want to see how these systems behave at the limits. The 6650XT being generally regarded as a great 1080p card makes a better starting point.
The data is still helpful though for anyone buying a 6700XT, combined with previous GPU comparisons you can interpolate a pretty good idea of about where it would have benchmarked.
@@kenshirogenjuro873 I just meant on the whole, with as much data as there is showing how amazing of a deal the 6700XT is (cue every single monthly GPU value video since its launch), there's *shockingly* little attention given to it. I'd have figured that would be used as well.
Not hating, just find it interesting that they point to one thing being super high value, but never show it in any video ever when they do these comparisons.
Without even watching video - with my budget PC (RX 6600 XT) it wasn't worth waiting for Zen4 anyhow when I owned AM4 (with R5 2600) already. For higher end GPUs (and generally systems) and new builders - there is a bit different price and value perspective - where you can justify paying quite a bit more, for not so much more performance. Also - it's a bit different situation now since Zen4 launch prices fall a bit of a cliff. Also with cards like RX 6900 XT and up, it would be silly to to go for very budget CPU as R5 5600 which is just great value CPU for people on budget systems.
Great video. Now we need one with x3D added.
who is still rocking a haswell? 4770k here .... Need a new pc !
The video i was waiting for...
This video is gonna sell a lot of RX 6950 XT GPUs because of people confused thinking the 6950 XT = RTX 4090 performance.
These are the videos more people need to pay attention to. Too many people want to argue about one CPU being 5% faster than another without thinking about the fact that's only when it's paired with a $2000 GPU. It's always important to build your system evenly, otherwise you're just straight wasting money.
In other words, if your target video card is the the RX 6800XT or better, you should opt for Zen4, but if not, then the older Zen3 is fine
Could you guys do a test like this for 4K gaming too? I'd love to see the difference between a 5600 and 7600/7900 running a 4090, especially in CPU bound games like Spiderman Remastered.
funny how Spiderman is very playable even on my i5 3470. i get 40 plus in city swinging at all maxed rt off. 60 locked inside building where missions happen. with rt on rest on high i get very playable 30 locked.
that cpu is a dinosaur from 2012
@@sharathvasudev Mainly just want a confirmation from them, along with another video to watch 😉.
Would be awesome to see 1440 and 4k comparison too here, or even better ultra wide display.
4k is GPU bound in virtually every game. 1440p is GPU bound on low-end and mid-range GPUs, so CPU will only matter if you have a high-end GPU where 1440p can start being CPU bound. It would be a large time waste for Steve to run those tests.
@@DarthChewie thanks, I understand that but think it's still worth at least briefly mentioning as 1 combo chart.
@@DarthChewie for instance, I am thinking of upgrading to 6800xt, I use 3440×1440 ultra wide display. Will I need to upgrade my 3600x or will it be sufficient for single player RPGs with higher quality graphics
@@Selandr Is it a 1080p ultrawide or a 1440p ultrawide? If 1080p, then it might depend on which games you play if you would become CPU bottlenecked with a 6800xt and 3600x, but on a 1440p ultrawide you'd definitely be GPU bound.
Either way, I imagine that a GPU upgrade would still yield a good uplift in performance, even if you were leaving a few % on the table as some games start to become more CPU bound on a 1080p ultrawide.
@@DarthChewie 1440p, we don't deal with half measures :) Thanks mate
Very nice as usual , I upgraded from a 1600 to 5700 and with a 2080 and this was like end of line facelift for my pc I understand after this there is no point in buying any older parts , literally a new system build from that point for me and planning on 7700x with a xtx for my next pc
You don’t need to upgrade from 5700 to 7700x.
It’s a one generation upgrade, cmon.
👍 Thanks, recently did a 5600 / 6600 build to hold me for a while.
Im sorry but who the hell is gonna play cyberpunk2077 on 1080p medium settings? lol even with a 6650xt lol
I am fairly certain that if you play at 1080p, you will not have to upgrade in the next 10 years if you have something from last generation (amd 5600, intel 12600, 3060 or above, or similar graphics card). I think we have mostly plateaued in terms of graphic fidelity and resolution is the only frontier where they can actually make enough gains to sell newer products. The number of people who gain something from having 200fps instead of 160fps at 1080p is probably very limited.
Thank you for this video. Helped me to decide on what to upgrade to from my i7-4770 to pair it with a used 2060 super I recently bought. Looks like Ryzen 5600 would be more than adequate for my case. And it should see me through to a potential 4060-like gpu upgrade in a few years time.
Why did you include an i3 on there? Just already had the data done recently?
Hitman 3 3090TI Quality setting, nice ;) The overhead is quite nuts honestly.
if you're using a 3090Ti or 4090 for 1080p, you're a bloody nutcase. This isn't an issue whatsoever if you are using the card with a resolution appropriate for the horsepower on tap.
Can you test the 7900xtx versus the 4090 with a 5800x3d. I Iwant to know how much nvidas driver overhead will affect the 5800x3d.
Would love to see scaling on this as well. I'm hoping to skip AM5 and keep this 5800X3D until next socket is out.
Currently on a 3070Ti and thinking I might upgrade to the next gen when they appear in a couple of years maybe.
If the current price insanity keeps up I can at least spread the cost more and get a new GPU in between the mobo+cpu upgrades.
I am missing benchmarks with RT enabled - would be courious if CPU performance makes a difference with RT on.
Excellent video. My only question is: how bad is the "Nvidia driver overhead" on lessor cards, 3060,3070,etc.
It will be less present but still there. You can probably calculate a few extra percent less.