A confession to make - I've been sitting on these for quite some time as I've recorded the comparison back in April and just couldn't get myself to put it all together. It is less relevant now than it was back then I think, but I figured it's still better to put it out late than not at all. The promised simracing focused part 2 will hopefully come in a week or two (my schedule for the next week or so is a bit packed, so I'm not sure right now).
of these tests, you have an indicative result only in the banisher, in the dogma in places gpu bottleneck, and in cyberpunk from the very beginning, vram loading is 100%, so there is a drop to 35 fps at 10:19. for the cpu test, it is better to set 1080p dlss performance, it is clear that no one uses it, but then why this test
@@ovlbo In DD2, the only part where you could suggest there being a GPU bottleneck is once outside of the main town, and even then not all the time. And that is also absolutely intentional, because the main CPU focus was on the famously very CPU bottlenecked main town area. I simply wanted to also show how the game performs outside of town where there's much less of a CPU bottleneck, and during some fighting, even if it meant getting closer to a GPU bottleneck. As for Cyberpunk, being close to the VRAM limit doesn't really matter, there's little you can do about that, that's how the game is with these settings (for example, I consider at least some of the RT settings mandatory for playing Cyberpunk, and that has a big impact on that). Those random framerate drops in Dogtown can not be mitigated, at least not on my setup. Dropping to 1080p achieves nothing in that regard except making the game look like ass. That's how the Dogtown area is on my setup. I suspect the drops might be related to parts of the area loading. And yes, if this was supposed to be a pure CPU benchmark/test, then you would obviously drop the graphics settings as low as possible. That's not what this video was intended to be, though, it was simply meant to be a more "realistic" comparison using settings you might (well, I might, anyway) use during normal gameplay, meaning not using ridiculously high graphics settings that would completely negate any CPU changes, but also not using settings so low nobody would actually play the game like that. I think this is something I should've made more clear in the video, because it keeps coming up in comments like yours.
Good video. I also have this setup (5600x and 3070) and I also plan to update my CPU so this for sure helps. The longer showcases are Nice. More vídeos like this with other games would be nice
As mentioned, there will be a more comprehensive video focused on simracing titles coming relatively soon. I will cover pretty much all of the current simracing titles. Unfortunately nothing beyond that right now.
You have a really good cpu as of now and zen4 is only 20% faster AT MOST. You would see much more of an increase in fps and way less frame dips if you upgrade you gpu to something faster and with more vram. 7900gre or 4070 super would massively increase your performance(40-45%) and your cpu is still strong enough to keep up with either of those gpus. You can also probably get a good amount of money for your 3070. If your budget constrained like me, you should strongly consider upgrading your gpu before cpu in your specific case.
Guys, 3070, 5600x and 16 GB of RAM 3733 SL 16 on hynix (or 18, I don't remember...) Need advice. I only play competitive games (mainly cs2), I need minimal latency. I feel like I have problems with this, since there are constant lags, "viscosity", the hertz is falling apart, etc. 2 options: take 5700x3d, or 32 GB of RAM b-die DDR4. I play full HD, I don't seem to need a new GPU, I only have enough money for one thing now)
From what I know, people playing the so-called competitive games tend to be obsessed about framerate, in which case CPU should be the most important thing, provided you're not GPU limited. But I find the world of competitive games to be a bit bizarre and rather illogical quite often, so your best bet is probably asking people focused on the specific games you are interested in.
@@Case_ I agree, I also can't find a 100% answer, although I seem to know quite a bit. In any case, I just want to see some advice, perhaps with a little explanation, because my opinion about whether to take a processor or RAM changes literally every hour)
Good video. I needed a 5600x vs 5700x3d comparison for a cheap upgrade to pair with my 4070super. I can upgrade with 220€ and stay with AM4 for other 2 years, almost. So 1440p was perfect, and perfect was also the test on that shitty game I love (dragons dogma 2), that's the most unoptimized game today. Just a tip: put the resolution on the screen, or in the title or at least in the description
It's somewhat explained in the video description, and it's also been discussed multiple times before. In other words: because "reasonable" settings for 60+ fps on a 3070 in 1440p, and also somewhat because it's a CPU-focused test, not a GPU-focused test.
Did you already tried ACC with these two? Can you tell me the difference? I have a 3600 with a rx6700xt and i dont know which one i should buy? The 5600x or the 5700x3d. I play mostly sim racing games like ACC.
@@Moerk-vn4zq I did, as mentioned, the simracing-focused part of the comparison is coming soon. There's pretty much no scenario where the 5700X3D does not give you better performance over the 5600X, so it's more of a question of price difference between the two. The 5600X was fine for me to run ACC at around 90-100 fps, with the 5700X3D, it's more like 130-140 with no changes in graphics settings. That's on my 3070 in 1440p with DLSS balanced.
@@Moerk-vn4zq I use customized settings. It will be shown in the video (though I've since upped some as I had some spare headroom thanks to the 5700X3D). I vaguely remember describing my settings under one of my recent ACC videos, but don't remember which one exactly, sorry. The Paul Ricard one, maybe?
I have 3080, do you think the upgrade will make a difference in 3440x1440? Bear in mind I have the 5600x overclocked to 4.4Ghz which is noticeable difference.
Difficult to tell. Depends on what your desired framerate is, what resolution and obviously, most importantly, what game are we talking about. If you're frequently GPU limited, upgrading CPU will not help you much.
@@aleksdeveloper698Check GPU utilization during gaming. If it regularly stays above say 90 percent, you're very close to being GPU limited. Above say 95 you're pretty much GPU limited right now. On the other hand, if your GPU utilization is lower than that, but lowering resolution doesn't really help with improving framerate or doesn't help as much as could be expected, you're likely to benefit from a more powerful CPU. It's not easy to predict how a certain game will respond to CPU or GPU upgrades, especially with modern titles these days, but these guidelines should help to give you at least a general idea.
Need advice... Will I notice a difference in games if I replace my 5600x with a 5700x3d? I only play AAA games - single-player ones. My configuration: rx 7800xt, Asus B550 motherboard, 32gb ram 3600MHz, 1080p monitor. I will be grateful for your answer.
I'm afraid I won't be of much help. I have no real experience with AMD GPUs, and whether or not you are currently CPU limited varies from game to game and also depends a lot on your specific settings (and your target framerate). Apparently the 7800xt is a considerably more powerful GPU than my 3070, so i would assume it is likely you might be CPU limited in quite a few titles by the 5600X, especially if you're only running at 1080p, but that's just guessing. Check your current GPU utilization in games that perform worse than you would like. If you're not reaching close to 100%, then you're likely CPU limited and a CPU upgrade would probably help.
hey man thanks for the video im looking to upgrade to the 5700x3d i have a rtx 3080ti and a 4k panel which i mostly use dlls on quality so i guess it would be 2k real resolution. in your eyes do you think is worth it?
@@ВиталийПоплевкин Дружище, я вот только поменял 5600х на 5700x3d. Приведу один пример. Киберпанк с отражениями лучами, сразу на выходе в город из квартиры Ви и в окрестностях при максимальной толпе у меня система проваливалась в 50+ кадров, 5600х разогнанный был загружен на 80-90%. На 5700x3d у меня в тех же условиях стабильные 75 кадров теперь и напрочь пропали статеры. Видяха у меня 6950xt. Обмен стоит выделки на все 200%
The better performance is due to more cache..... If they just made another version of the 5600x or the 5700x but with more l3 cache they wouldn't NEED the 3d cc,... Just my thought, lol
I wouldn't sum it like that, because it's not really specifically about 1440p, more about what using that resolution and the other graphics settings achieves, but if you insist on phrasing it like that, then yes. The other comparisons I've seen and referred to were flawed because they were GPU limited and/or using impractical/unreasonable graphics settings, which resulted in not showing the potential this specific CPU upgrade actually has. If you set your games so you're always GPU limited, the CPU makes no difference.
@@vincenthughes5795 I'm not sure what you mean by "GPU becomes the bottleneck". Banishers and Dragon's Dogma are almost never GPU limited in my video, even with the 5700X3D (and even less so with the 5600X), and while Cyberpunk certainly spends more time being GPU limited here, especially with the 5700X3D, there's still plenty of it not being GPU limited. As to why test these specific settings (which happen to be at 1440p for all three games on this setup), that's simple - I wasn't trying to do a "pure" CPU test (in which case you'd run the games with lowest detail at something like 720p), but more of a "reasonable settings" test, meaning using settings that are not pointlessly cranked up to the max, but still have some headroom and correspond to what someone with that setup might use (specifically, I, because they're more or less my normal settings for those games, settings that balance performance and visuals while leaving some headroom for some more intense scenes).
@@Case_ mate, in 2 of 3 tests for 5600x fps is less than 60, to me it doesn't necessarily indicate "reasonable settings" and "someone with that setup might use". I'd say that's a 1080p setup for 144Hz monitor. In any case, to avoid all those questions it's better to just do at least two resolutions. IMHO.
@@vincenthughes5795 Dragon's Dogma 2 was notoriously CPU limited (no idea if they improved it by now), especially in this town, that was the entire reason it was included in the video, and lowering resolution did not really improve framerate. So dropping to 1080p would not really help anything except making the game look horrible. I only really drop resolution below 1440p in games as a last resort when everything else fails and when I really can't deal with the framerate at all, because 1080p looks horrible, especially with DLSS (which is often pretty much mandatory with modern games) And while Cyberpunk has drops below 60 in Dogtown area (which I included in the video quite intentionally to point out that Dogtown runs worse than the rest of the game), the performance shown seems quite reasonable to me for normal playing. With the 5600X, dropping resolution further would again result in no real performance improvement and would just make the game more of a blurry mess than it already is with these settings. And while some performance could be gained by turning off RT, that's also something I don't think people should do in this case (unlike many other games). So some drops to 50s are IMO a perfectly acceptable compromise, especially given that most actual missions run a fair bit better than that anyway You are obviously free to disagree or prefer something else, but those settings are not random and there's logic behind them. And, more importantly, they serve the purpose of the CPU comparison fairly well and actually show what could be expected of those CPUs. They could certainly be improved, but you know what they say about hindsight. I don't have the luxury of swapping the CPUs back and forth and optimizing the settings to perfection, it all had to be done as best as possible in one go and with quite a few games to go through, too (there's part 2).
It is one of the worst cases of CPU bottleneck recently, yeah. I should check to see if they actually improved it since (even though this video was uploaded two months ago, the recording reflects the game's state fairly shortly after release).
Try 10-20% instead of absolute figures (because you can see especially in Banishers that it really isn't 5 fps). And even if those 10-20 percent only give you 5(ish) fps, it might make a noticeable difference in playability. But sure, if you don't care about higher performance, then don't get a better CPU, that's only logical. Keep what you have and enjoy.
@@antimafiotbg for me this one of the reasons why I switched from 5600x to 5700x3d. Now I don't have any deeps into 50fps, I got stable 75+ in the same scenarios. And this is just one example, you can easily find more, for example Dragons Dogma 2 etc.
A confession to make - I've been sitting on these for quite some time as I've recorded the comparison back in April and just couldn't get myself to put it all together. It is less relevant now than it was back then I think, but I figured it's still better to put it out late than not at all. The promised simracing focused part 2 will hopefully come in a week or two (my schedule for the next week or so is a bit packed, so I'm not sure right now).
Здравствуйте. Какая частота оперативной памяти используется в видео? Слышал кеш L3 который в 5700x3d очень хорошо работает с памятью высокой частоты.
of these tests, you have an indicative result only in the banisher, in the dogma in places gpu bottleneck, and in cyberpunk from the very beginning, vram loading is 100%, so there is a drop to 35 fps at 10:19. for the cpu test, it is better to set 1080p dlss performance, it is clear that no one uses it, but then why this test
@@ovlbo In DD2, the only part where you could suggest there being a GPU bottleneck is once outside of the main town, and even then not all the time. And that is also absolutely intentional, because the main CPU focus was on the famously very CPU bottlenecked main town area. I simply wanted to also show how the game performs outside of town where there's much less of a CPU bottleneck, and during some fighting, even if it meant getting closer to a GPU bottleneck.
As for Cyberpunk, being close to the VRAM limit doesn't really matter, there's little you can do about that, that's how the game is with these settings (for example, I consider at least some of the RT settings mandatory for playing Cyberpunk, and that has a big impact on that). Those random framerate drops in Dogtown can not be mitigated, at least not on my setup. Dropping to 1080p achieves nothing in that regard except making the game look like ass. That's how the Dogtown area is on my setup. I suspect the drops might be related to parts of the area loading.
And yes, if this was supposed to be a pure CPU benchmark/test, then you would obviously drop the graphics settings as low as possible. That's not what this video was intended to be, though, it was simply meant to be a more "realistic" comparison using settings you might (well, I might, anyway) use during normal gameplay, meaning not using ridiculously high graphics settings that would completely negate any CPU changes, but also not using settings so low nobody would actually play the game like that. I think this is something I should've made more clear in the video, because it keeps coming up in comments like yours.
Thanks for the comparison!
Great video mate, thank you!
Good video. I also have this setup (5600x and 3070) and I also plan to update my CPU so this for sure helps. The longer showcases are Nice. More vídeos like this with other games would be nice
As mentioned, there will be a more comprehensive video focused on simracing titles coming relatively soon. I will cover pretty much all of the current simracing titles. Unfortunately nothing beyond that right now.
You have a really good cpu as of now and zen4 is only 20% faster AT MOST. You would see much more of an increase in fps and way less frame dips if you upgrade you gpu to something faster and with more vram. 7900gre or 4070 super would massively increase your performance(40-45%) and your cpu is still strong enough to keep up with either of those gpus. You can also probably get a good amount of money for your 3070. If your budget constrained like me, you should strongly consider upgrading your gpu before cpu in your specific case.
Guys, 3070, 5600x and 16 GB of RAM 3733 SL 16 on hynix (or 18, I don't remember...) Need advice. I only play competitive games (mainly cs2), I need minimal latency. I feel like I have problems with this, since there are constant lags, "viscosity", the hertz is falling apart, etc. 2 options: take 5700x3d, or 32 GB of RAM b-die DDR4. I play full HD, I don't seem to need a new GPU, I only have enough money for one thing now)
From what I know, people playing the so-called competitive games tend to be obsessed about framerate, in which case CPU should be the most important thing, provided you're not GPU limited. But I find the world of competitive games to be a bit bizarre and rather illogical quite often, so your best bet is probably asking people focused on the specific games you are interested in.
@@Case_ I agree, I also can't find a 100% answer, although I seem to know quite a bit. In any case, I just want to see some advice, perhaps with a little explanation, because my opinion about whether to take a processor or RAM changes literally every hour)
I bought a better processor, RAM does not play such an important role for FPS.
Yeah, I forgot to mention more RAM would play a miniscule role, if any.
@@Case_ ty, i buy r5700x3d)
Good video.
I needed a 5600x vs 5700x3d comparison for a cheap upgrade to pair with my 4070super. I can upgrade with 220€ and stay with AM4 for other 2 years, almost.
So 1440p was perfect, and perfect was also the test on that shitty game I love (dragons dogma 2), that's the most unoptimized game today.
Just a tip: put the resolution on the screen, or in the title or at least in the description
The resolution is right there in the Afterburner overlay visible at all times, among many other things ;)
@@Case_ omg forgive me! Whatching vid on a smartphone and didn't see that resolution was written down there 😅
@@Stinox91 😂No worries.
220€?? i got my 5700x3d for 136€..
@@s4yto streer price in EU is around 200
Why Dlss performance?
It's somewhat explained in the video description, and it's also been discussed multiple times before. In other words: because "reasonable" settings for 60+ fps on a 3070 in 1440p, and also somewhat because it's a CPU-focused test, not a GPU-focused test.
Did you already tried ACC with these two? Can you tell me the difference? I have a 3600 with a rx6700xt and i dont know which one i should buy? The 5600x or the 5700x3d. I play mostly sim racing games like ACC.
@@Moerk-vn4zq I did, as mentioned, the simracing-focused part of the comparison is coming soon. There's pretty much no scenario where the 5700X3D does not give you better performance over the 5600X, so it's more of a question of price difference between the two. The 5600X was fine for me to run ACC at around 90-100 fps, with the 5700X3D, it's more like 130-140 with no changes in graphics settings. That's on my 3070 in 1440p with DLSS balanced.
@@Case_ thanks for your Feedback! Do you use high or epic graphic settings in ACC?
@@Moerk-vn4zq I use customized settings. It will be shown in the video (though I've since upped some as I had some spare headroom thanks to the 5700X3D). I vaguely remember describing my settings under one of my recent ACC videos, but don't remember which one exactly, sorry. The Paul Ricard one, maybe?
I have 3080, do you think the upgrade will make a difference in 3440x1440?
Bear in mind I have the 5600x overclocked to 4.4Ghz which is noticeable difference.
Difficult to tell. Depends on what your desired framerate is, what resolution and obviously, most importantly, what game are we talking about. If you're frequently GPU limited, upgrading CPU will not help you much.
@@Case_ how would i know if I am frequently GPU limited?
@@aleksdeveloper698Check GPU utilization during gaming. If it regularly stays above say 90 percent, you're very close to being GPU limited. Above say 95 you're pretty much GPU limited right now.
On the other hand, if your GPU utilization is lower than that, but lowering resolution doesn't really help with improving framerate or doesn't help as much as could be expected, you're likely to benefit from a more powerful CPU.
It's not easy to predict how a certain game will respond to CPU or GPU upgrades, especially with modern titles these days, but these guidelines should help to give you at least a general idea.
@@Case_ Yeah I am GPU limited, 95%+ in all games.
Need advice... Will I notice a difference in games if I replace my 5600x with a 5700x3d? I only play AAA games - single-player ones. My configuration: rx 7800xt, Asus B550 motherboard, 32gb ram 3600MHz, 1080p monitor. I will be grateful for your answer.
I'm afraid I won't be of much help. I have no real experience with AMD GPUs, and whether or not you are currently CPU limited varies from game to game and also depends a lot on your specific settings (and your target framerate). Apparently the 7800xt is a considerably more powerful GPU than my 3070, so i would assume it is likely you might be CPU limited in quite a few titles by the 5600X, especially if you're only running at 1080p, but that's just guessing. Check your current GPU utilization in games that perform worse than you would like. If you're not reaching close to 100%, then you're likely CPU limited and a CPU upgrade would probably help.
@@Case_ Ok,thanks
@@Ruslan-Night конечно будет разница для 1080p
hey man thanks for the video im looking to upgrade to the 5700x3d i have a rtx 3080ti and a 4k panel which i mostly use dlls on quality so i guess it would be 2k real resolution. in your eyes do you think is worth it?
For 4k it makes no sense to upgrade.
And if is for 1080p?? Whit the same config 5600x 3080ti
@@guidoxs941 if you have a panel up 144hz then go for it.
Здравствуйте. Скажите пожалуйста есть ли смысл сменить 5600х на 5700x3d? Монитор 3440х1440. Оперы 32 гб 3600 мг 2х16.ssd 980pro, rtx 3070ti.
Что решил?
Так вроде это видео как раз и сделано чтобы ответить на твой вопрос :))) нет?
@@b4dman взять rtx 4070ti super
@@ВиталийПоплевкин Дружище, я вот только поменял 5600х на 5700x3d. Приведу один пример. Киберпанк с отражениями лучами, сразу на выходе в город из квартиры Ви и в окрестностях при максимальной толпе у меня система проваливалась в 50+ кадров, 5600х разогнанный был загружен на 80-90%. На 5700x3d у меня в тех же условиях стабильные 75 кадров теперь и напрочь пропали статеры. Видяха у меня 6950xt. Обмен стоит выделки на все 200%
@@ВиталийПоплевкинДля твоей новой видяхи это будет просто бомбический апгрейд!
The better performance is due to more cache.....
If they just made another version of the 5600x or the 5700x but with more l3 cache they wouldn't NEED the 3d cc,...
Just my thought, lol
I mean that's literally what their 3D cache is - a way to include more L3 cache on the die.
ну короче есть разница 20-30 фпс в 2к пользу x3d
Really? You just do a 1440p comparison and call the other ones (that compare all resolutions) flawed?
I wouldn't sum it like that, because it's not really specifically about 1440p, more about what using that resolution and the other graphics settings achieves, but if you insist on phrasing it like that, then yes. The other comparisons I've seen and referred to were flawed because they were GPU limited and/or using impractical/unreasonable graphics settings, which resulted in not showing the potential this specific CPU upgrade actually has.
If you set your games so you're always GPU limited, the CPU makes no difference.
@@Case_ 100% correct. So by that logic why only test 1440p, where GPU becomes the bottleneck and CPU matters less?
@@vincenthughes5795 I'm not sure what you mean by "GPU becomes the bottleneck". Banishers and Dragon's Dogma are almost never GPU limited in my video, even with the 5700X3D (and even less so with the 5600X), and while Cyberpunk certainly spends more time being GPU limited here, especially with the 5700X3D, there's still plenty of it not being GPU limited.
As to why test these specific settings (which happen to be at 1440p for all three games on this setup), that's simple - I wasn't trying to do a "pure" CPU test (in which case you'd run the games with lowest detail at something like 720p), but more of a "reasonable settings" test, meaning using settings that are not pointlessly cranked up to the max, but still have some headroom and correspond to what someone with that setup might use (specifically, I, because they're more or less my normal settings for those games, settings that balance performance and visuals while leaving some headroom for some more intense scenes).
@@Case_ mate, in 2 of 3 tests for 5600x fps is less than 60, to me it doesn't necessarily indicate "reasonable settings" and "someone with that setup might use". I'd say that's a 1080p setup for 144Hz monitor. In any case, to avoid all those questions it's better to just do at least two resolutions. IMHO.
@@vincenthughes5795 Dragon's Dogma 2 was notoriously CPU limited (no idea if they improved it by now), especially in this town, that was the entire reason it was included in the video, and lowering resolution did not really improve framerate. So dropping to 1080p would not really help anything except making the game look horrible. I only really drop resolution below 1440p in games as a last resort when everything else fails and when I really can't deal with the framerate at all, because 1080p looks horrible, especially with DLSS (which is often pretty much mandatory with modern games)
And while Cyberpunk has drops below 60 in Dogtown area (which I included in the video quite intentionally to point out that Dogtown runs worse than the rest of the game), the performance shown seems quite reasonable to me for normal playing. With the 5600X, dropping resolution further would again result in no real performance improvement and would just make the game more of a blurry mess than it already is with these settings. And while some performance could be gained by turning off RT, that's also something I don't think people should do in this case (unlike many other games). So some drops to 50s are IMO a perfectly acceptable compromise, especially given that most actual missions run a fair bit better than that anyway
You are obviously free to disagree or prefer something else, but those settings are not random and there's logic behind them. And, more importantly, they serve the purpose of the CPU comparison fairly well and actually show what could be expected of those CPUs. They could certainly be improved, but you know what they say about hindsight. I don't have the luxury of swapping the CPUs back and forth and optimizing the settings to perfection, it all had to be done as best as possible in one go and with quite a few games to go through, too (there's part 2).
that cpu bottleneck in ddm2 is pain to see
It is one of the worst cases of CPU bottleneck recently, yeah. I should check to see if they actually improved it since (even though this video was uploaded two months ago, the recording reflects the game's state fairly shortly after release).
i dont see big difference at all if you have 5600 there is no reason to buy 5700x3d just for 5 fps!
Try 10-20% instead of absolute figures (because you can see especially in Banishers that it really isn't 5 fps). And even if those 10-20 percent only give you 5(ish) fps, it might make a noticeable difference in playability.
But sure, if you don't care about higher performance, then don't get a better CPU, that's only logical. Keep what you have and enjoy.
Really? In cyberpunk where 5600x falls into 48fps 5700x3d gives 67. Wtf are you talking about lol?
@TheMrSatyricon just in one game what if you don't play cyberpunk
@@antimafiotbg for me this one of the reasons why I switched from 5600x to 5700x3d. Now I don't have any deeps into 50fps, I got stable 75+ in the same scenarios. And this is just one example, you can easily find more, for example Dragons Dogma 2 etc.
the 5700x3d has way more stability especially for those 1% lows