DF x IGN closest GPU to PS5 pro is an RTX 4070

Gaiff

SBI’s Resident Gaslighter
It's more than double of the PS5 Pro if you get a GPU of that performance with AI (so, minimum 4070 Super, no AMD card in the market is equivalent), equivalent memory and SSD, plus controller, OS etc.

Over $1600 /1600€.
The 4070S that's 100% faster than the PS5's GPU is required to match the 45% faster than the PS5 GPU of the Pro?

Something isn't adding up.
 

yurinka

Member
The 4070S that's 100% faster than the PS5's GPU is required to match the 45% faster than the PS5 GPU of the Pro?

Something isn't adding up.
4070 is the closest one to the "45% extra performance" mentioned in the PS5 Pro reveal in benchmarks versus the most similar PC equivalent of the PS5 GPU, this is why people say it's the equivalent.

At that level there's also the 3080 but it has older tech and some AMD ones but don't feature AI.

To be more accurate, the PC GPU equivalent would be a potential future RDNA4 4070 equivalent from AMD featuring AI stuff but still doesn't exist. So, if people has to choose the most similar thing in the market, they chose 4070.
 
Last edited:

Gaiff

SBI’s Resident Gaslighter
4070S provides aprox. the 45% extra fps in benchmarks versus the most similar PC equivalent of the PS5 GPU, this is why people say it's the equivalent.
What kinda crack are you smoking?
At that level there's also the 3080 but it has older tech and some AMD ones but don't feature AI.
The 3080 is much faster than 45%. It's 80% faster and the 4070S is 100% faster. 45% is 3070 Ti tier.

Yk6vX4Y.png

XU3ZsKE.png


There's also Frontiers of Pandora where the 4070S is over twice the performance. The sole exception is TLOU Part I where the 4070S is a meager 36% faster.
 
Last edited:

yurinka

Member
What kinda crack are you smoking?

The 3080 is much faster than 45%. It's 80% faster and the 4070S is 100% faster. 45% is 3070 Ti tier.
II'm just talking about objective related benchmarks (not if it's ""faster"") and why people like Digital Foundry choose the 4070 as the equivalent.

If something you're the one smoking things and believing in fantasies.
 
Last edited:

Gaiff

SBI’s Resident Gaslighter
II'm just talking about objective benchmarks (not if it's ""faster"") and why people like Digital Foundry choose the 4070 as the equivalent.
Yes, and objective benchmark has the 4070S 100% faster than the PS5's GPU. They literally tested it there.


If something you're the one smoking things and believing in fantasies.
Maybe you should do your research before spouting nonsense? They used a 3070 Ti in Wukong and said it'd be broadly equivalent. The 4070 is used because the 3070 Ti has a paltry 8GB of VRAM. Cards like the 7700 XT don't have good ray tracing, so you're only left with the 4070 which is quite a bit faster, but still the closest one. 4070S is complete lunacy and Rich even said the 4070 should still be faster.
 

DoomD

Neo Member
It's more than double of the PS5 Pro if you get a GPU of that performance with AI (so, minimum 4070 Super, no AMD card in the market is equivalent), equivalent memory and SSD, plus controller, OS etc.

Over $1600 / 1600€ without counting keyboard, mouse and monitor.


Apparently Horizon gets native 4K in PS5 Pro (at 30fps), which provides better IQ than DLSS 4K in PC (at 60fps).
Its about $1000 to build a PC that is faster than the PS5 with a RTX 4070 non super and that's not even going as cheap as you could. You could probably get something for about $900 if you really skimped or found good deals. Then we have AMDs new cards coming out soon, most likely January from the most current leaks and from the sounds of things they are going to target $400 for the 8700XT which will basically be the same GPU as in the pro but with a higher clock and more cache. So in just a few months time you'll be able to build a PC that is faster than the PS5 Pro for about $800.....food for thought.
 

ChiefDada

Member
Apparently Horizon gets native 4K in PS5 Pro (at 30fps), which provides better IQ than DLSS 4K in PC (at 60fps).

Unfortunately, we don't know how exactly it compares to the base PS5 fidelity mode targeting native 4k because DF didn't do that specific comparison for reasons unknown.
 

yurinka

Member
Yes, and objective benchmark has the 4070S 100% faster than the PS5's GPU. They literally tested it there.



Maybe you should do your research before spouting nonsense? They used a 3070 Ti in Wukong and said it'd be broadly equivalent. The 4070 is used because the 3070 Ti has a paltry 8GB of VRAM. Cards like the 7700 XT don't have good ray tracing, so you're only left with the 4070 which is quite a bit faster, but still the closest one. 4070S is complete lunacy and Rich even said the 4070 should still be faster.

RTX 4070 non super

Edit: yes, I was saying 4070S/Super but I meant 4070, my bad. I thought they were the same card (not the Ti variant) but obviously are different.

Example, assuming 6700 as PS5 equivalent (that's a separate debate), adding a 45% faster rendering/fps the most similar one you get is the 4070 when removing the other ones due to the reasons mentioned (no DL in AMD, too low VRAM on 3080 etc). Here you have a random case:
https://technical.city/en/video/Radeon-RX-6700-vs-GeForce-RTX-4070

There were many previous benchmark made/shown when the specs got leaked to try to find a PC equivalent and when removing the AMD ones the 4070/3080 were also inside that range.

But well, obviously the case in any benchmark will vary depending on the CPU, GPU, memory, SSD etc used in each case.

That's the raw fps performance, on top of that there's also the differences between using DLSS, PSSR, FSR etc. obviously varies the performance on each device.

Obviously there will be also differences between RDNA3/4 vs Nvidia etc. And even between the PS5 Pro one and the future potential PC equivalent (integrated memory with higher bandwidth, different clock boost, etc).

Unfortunately, we don't know how exactly it compares to the base PS5 fidelity mode targeting native 4k because DF didn't do that specific comparison for reasons unknown.
Maybe they assumed that by default native 4K looks better than DLSS 4K, so no need to compare it.
 
Last edited:

Gaiff

SBI’s Resident Gaslighter
Edit: yes, I was saying 4070S but I meant 4070, my bad. I thought they were the same card (not the Ti variant) but are different.

Example, assuming 6700 as PS5 equivalent (that's a separate debate), adding a 45% faster rendering/fps the most similar one you get is the 4070 when removing the other ones due to the reasons mentioned (no DL in AMD, too low VRAM on 3080 etc). Here you have a random case:
https://technical.city/en/video/Radeon-RX-6700-vs-GeForce-RTX-4070

But well, obviously the case in any benchmark will vary depending on the CPU, GPU, memory, SSD etc used in each case.

That's the raw fps performance, on top of that there's also the differences between using DLSS, PSSR, FSR etc. obviously varies the performance on each device.

Obviously there will be also differences between RDNA3/4 vs Nvidia etc. And even between the PS5 Pro one and the future potential PC equivalent (integrated memory with higher bandwidth, different clock boost, etc).
Still incorrect. 45% performance increase over the 6700 gets you a 7700 XT. Here in Elden Ring, Rich uses a 6800 vs a PS5 and the 6800 is locked to 60fps. In intense moments, it's up to 55% faster than the regular PS5.



4070 is 70% faster (3080 is actually 80%, I keep forgetting they aren't equal)
3080 is 80% faster
4070S is around 100% faster

Here is the Techpowerup chart that just includes 1080p.



4070 is 60% faster than the 6700. If you include 1440p and 4K, the difference balloons up to 70%. It's not just 45% faster. And the technicalcity link is garbage. They just gather generic data. They don't do thorough testing like Techpowerup or HU.
 
Last edited:

yurinka

Member
Still incorrect. 45% performance increase over the 6700 gets you a 7700 XT. Here in Elden Ring, Rich uses a 6800 vs a PS5 and the 6800 is locked to 60fps. In intense moments, it's up to 55% faster than the regular PS5.
The 6800 is can't be an equivalent for then using PSSR because it doesn't have DL enhacements, this is why he did choose the 4070.

He did choose 6800 only for the case of games without PS5 Pro patch, to see the equivalent in the PS5 Pro Boost improvement. For that case yes, 6800 makes sense because it gets that somewhat 45% improvement depending on the areas. The exact % obviously will vary depending on the game and area.

 

Gaiff

SBI’s Resident Gaslighter
The 6800 is can't be an equivalent for then using PSSR because it doesn't have DL enhacements, this is why he did choose the 4070.
I'm not saying the 6800 is equivalent. I'm saying the 6800 is already faster than the quoted 45% and the 4070 is faster than the 6800. Clearly, it's not a merely 45% faster than the regular PS5. The 4070 is closer to being 70% faster. So you telling us we'll need a 4070S to match the Pro doesn't add up.
He did choose 6800 only for the case of games without PS5 Pro patch, to see the equivalent in the PS5 Pro Boost improvement. For that case yes, 6800 makes sense because it gets that somewhat 45% improvement depending on the areas. The exact % obviously will vary depending on the game and area.
Yes, and again, this is to counter your claim that you'll need a 4070S to match the PS5 Pro's GPU. You won't. A 4070 will be more than enough and might actually end up a notch faster in most cases. Whatever the case, the experience they provide shouldn't be too far apart. The 4070 can be had for $520 at the moment. A system in the Pro's ballpark can be built right now for around $1000, probably even less. Nowhere near "over twice the price" like you claimed. You can get a system with 4070 Ti Super at that price point.

Not that I really care about building a PC that's comparable or better than a PS5. No one builds a PC with the explicit goal of beating consoles, so those discussions are always dumb, but there's no need to claim falsehoods like you need to shell out $1400 on a PC to get a PS5 Pro experience. Not even close.
 

Zathalus

Member
*reads the last 7 to 8 pages*

PS5 Pro is making and exposing actual retardation. Idiocracy levels.

Thank you Zathalus Zathalus for having an actual thinking cap on.
I’m only upset about the price. Not because I can’t afford one, it just sucks that the days of subsidised hardware are behind us. I’m still going to end up buying one because I enjoy PS5 games. Bought all the exclusives so far this gen bar one. No guess needed as to which one I skipped.
 

Tqaulity

Member
I've already said that this talk about the PS5 PRO GPU being near a RX 6800 or RX7700XT is a bit silly and shortsighted for various reasons. Mostly coming from people who just see "60CUs" in the 6800 or try to calculate +45% from a ~RX 6700 and land at a 7700XT which is fairly naïve thinking. Well here is some more discussion from MLID in an interview with an actual developer who both agreed the association with an RX 6800 is selling it a bit short. I'm not getting into the weeds with you guys, just some food for thought for those that know what how to interpret this data correctly (timestamped below):

 

Bojji

Member
Base PS5 runs this game fixed 1440p 75-100fps. Likely high settings with max textures.

x600bLX.png

There were quite big differences between console settings and pc settings in GOW2018, I doubt PS5 version in performance mode is close to max settings.
 

rsouzadk

Member
There were quite big differences between console settings and pc settings in GOW2018, I doubt PS5 version in performance mode is close to max settings.

You are right. Settings are different. The game even has the console settings if you want to in graphics options.
 
Last edited:

ChiefDada

Member
You are right. Settings are different. The game even has the console settings if you want to in graphics options.

What do you mean? There is no unified preset shared between the quality and performance mode on PS5. In the new DF interview, the devs confirmed PS5 fidelity mode is using ultra settings, and there's 120hz available that locks fidelity mode to 40 fps. We don't know settings yet for HFR mode on PS5. Based on everything in seeing, PS5 Pro will easily scale past 4070 in this game.

CTVQT5n.jpeg
 

rsouzadk

Member
What do you mean? There is no unified preset shared between the quality and performance mode on PS5. In the new DF interview, the devs confirmed PS5 fidelity mode is using ultra settings, and there's 120hz available that locks fidelity mode to 40 fps. We don't know settings yet for HFR mode on PS5. Based on everything in seeing, PS5 Pro will easily scale past 4070 in this game.

CTVQT5n.jpeg
Dunno what youre talking. I was referring to ps4.
 

SpokkX

Member
It literally almost doubled the framerates on a "supposed CPU bound game" with just a 10% clock boost. While having a much cleaner image.

That tells me the GPU does most of the heavy lifting. 🤡

yeah dig your hand in the sand and deny that this will be an issue going forward

This is not an attack on your favorite company - just stating what probably will be an issue with a product. You do not need to take offense

It will be left behind by 3-4 year old PC:s in CPU bound titles already. Imagine in 2 years..
 
Last edited:

DeepEnigma

Gold Member
yeah dig your hand in the sand and deny that this will be an issue going forward

This is not an attack on your favorite company - just stating what probably will be an issue with a product. You do not need to take offense

It will be left behind by 3-4 year old PC:s in CPU bound titles already. Imagine in 2 years..
Why are you gaslighting and projecting all in one post?

Offended? Surely not. Calling out people that claim one thing, but the results show another, sound.

It's doing what it was set out to do.
 

hinch7

Member
As expected:



The system is unbalanced. Weak CPU and powerful GPU.

Seems like such an easy miss - why not upgrade the CPU 45% also?

No shit. If a 3080 was bottlenecked by Zen 2's back in 2020. Consoles ones are even more cut down and clocked lower.

3080 > CPU needed review
gSYF5CbsMWtvkA6Jr5b5S9-1200-80.png.webp

Zen 5 nearly offers 2x gaming performance over Zen 2 desktop.
 
Last edited:

winjer

Member
As expected:



The system is unbalanced. Weak CPU and powerful GPU.

Seems like such an easy miss - why not upgrade the CPU 45% also?


That is only true for very poorly optimized games. And Dragon's Dogma is one of the worst offenders.
Even a 7800X3D only does 86 fps on average.
For games that are decent to good optimization, the Zen2 CPU on the PS5 is more than enough for 60 fps.

 

hinch7

Member
That is only true for very poorly optimized games. And Dragon's Dogma is one of the worst offenders.
Even a 7800X3D only does 86 fps on average.
For games that are decent to good optimization, the Zen2 CPU on the PS5 is more than enough for 60 fps.


Something tells me that RE Engine just isn't well suited for open world games. Performance recommendations for MW Wilds is daft, needing frame generation to run it.
 

winjer

Member
Something tells me that RE Engine just isn't well suited for open world games. Performance recommendations for MW Wilds is daft, needing frame generation to run it.

And that is another game that is seems to be very poorly optimized. But you can be sure that when it releases, some people will use that as an example to say the Zen2 CPU is not good enough.
 

Lysandros

Member
I always found completely dismissing CPU contribution to overall performance readings when comparing PC hardware to consoles simply disingenuous, especially (but not limited to) around 1080P resolution. Just as a relevant example, DF's literally XSX CPU equipped (even when allowed to boost higher in all likelihood) PC offers noticably worse performance when the same GPU is coupled with a CPU like R5 3600 (which is still undeniably faster than the CPUs found in consoles based on hardware).
 
Last edited:

hinch7

Member
I always found completely dismissing CPU contribution to overall performance readings when comparing PC hardware to consoles simply disingenuous, especially (but not limited to) around 1080P resolution. Just as a relevant example, DF's literally XSX CPU equipped (even when allowed to boost faster in all likelihood) PC offers noticably worse performance when the same GPU is coupled with a CPU like R5 3600 (which is still undeniably faster than those CPUs found in consoles based on hardware).
It does make sense though since most games are running some sort of DRS and/or using some sort of upscaling; like PSSR going forward. Would've also opened to the doors to the posibility of more games running at 120fps at reasonable image quality. Which would've been really cool to see on console.
 
Last edited:

winjer

Member
I always found completely dismissing CPU contribution to overall performance readings when comparing PC hardware to consoles simply disingenuous, especially (but not limited to) around 1080P resolution. Just as a relevant example, DF's literally XSX CPU equipped (even when allowed to boost faster in all likelihood) PC offers noticably worse performance when the same GPU is coupled with a CPU like R5 3600 (which is still undeniably faster than those CPUs found in consoles based on hardware).

The comparisons that DF made with those CPUs are terrible.
The 4800S has only 4 lanes of PCIe bandwidth, which will become a bottleneck frequently, as the CPU as to send data to the GPU and vice versa.
Their 3600 system is so poorly configured, that it has a memory latency of 90ns.

But then there are the differences between consoles and PCs.
The PS5 CPU does not have to decompress data, unlike the CPUs on the PC.
The PS5 also has a lower level API than DX12_2
And the OS on that console is not the broken crap that is Windows 11. It doesn't have all the bloatware, the spyware and all gigantic amount of bugs that Windows has.
FFS, some games in Windows 11 23H3 were losing up to 30% performance because of a bug with the Windows scheduler.
 

solidus12

Member
That build Colteastwood posted is terrible. As are some of the builds here. The absolute best you can do in the US for a build that would roughly match/best the performance of the Pro (aka 4070) would be around $1000. Assuming you are not using bottom of the barrel PSU and including the price of Keyboard and Mouse and Windows OS.
Yeah well it’s Colt. He said that the Pro won’t be able to outperform the Series X because devs have yet to utilize its full RDNA 2 feature set.
 

SABRE220

Member
What kinda crack are you smoking?

The 3080 is much faster than 45%. It's 80% faster and the 4070S is 100% faster. 45% is 3070 Ti tier.

Yk6vX4Y.png

XU3ZsKE.png


There's also Frontiers of Pandora where the 4070S is over twice the performance. The sole exception is TLOU Part I where the 4070S is a meager 36% faster.
Isnt that with a significantly better cpu? Also, the games seem to be performing significantly better than the 45% as even in dragons dogma a cpu heavy game it is going from low 30s to mid 50s with apparently better settings and resolution.
 
As expected:



The system is unbalanced. Weak CPU and powerful GPU.

Seems like such an easy miss - why not upgrade the CPU 45% also?


Even DF was impressed but you're trying to present it like it's shit. Did you even watch the video? It's a big boost over the base PS5. Even in a CPU limited game I'm still get 10+ more FPS in a game and you're trying to present that to me like it's bad.
 
Last edited:

hinch7

Member
Isnt that with a significantly better cpu? Also, the games seem to be performing significantly better than the 45% as even in dragons dogma a cpu heavy game it is going from low 30s to mid 50s with apparently better settings and resolution.
PSSR is a way better quality upscaler than anything AMD has put out. Scaling would of course increase as the GPU is more performant and PSSR delivers better image quality verses FSR 1/2/3 on the regular PS5.
 

Three

Member
No shit. If a 3080 was bottlenecked by Zen 2's back in 2020. Consoles ones are even more cut down and clocked lower.

3080 > CPU needed review
gSYF5CbsMWtvkA6Jr5b5S9-1200-80.png.webp

Zen 5 nearly offers 2x gaming performance over Zen 2 desktop.
On a 3080 at 1080p medium? Of course better CPU will boost your 87fps higher to 160fps. This isn't the same as being CPU bound at 60fps and higher res.
 
Last edited:

Gaiff

SBI’s Resident Gaslighter
Isnt that with a significantly better cpu? Also, the games seem to be performing significantly better than the 45% as even in dragons dogma a cpu heavy game it is going from low 30s to mid 50s with apparently better settings and resolution.
The CPU wouldn’t change anything at those frame rates.
 

Skifi28

Member
As expected:



The system is unbalanced. Weak CPU and powerful GPU.

Seems like such an easy miss - why not upgrade the CPU 45% also?

As expected? I have a much better CPU than the PS5/pro and I can't get anywhere near 50 FPS in the city. Sounds to me like the CPU in the console is fine and punching way above its weight during a heavy CPU stress-test.

Also, If the game is in the 50s in the city then that means locked 60 everywhere else. That's pretty huge.
 
Last edited:

hinch7

Member
On a 3080 at 1080p medium? Of course better CPU will boost your 87fps higher to 160fps. This isn't the same as being CPU bound at 60fps and higher res.
There are games that are heavy on CPU though. Space Marine 2 being one of them. I played that at 100-120fps on my PC with my 5800X3D+4070Ti maxed out. The game would struggle to even get 60 and regularly drops to 30 when in large battles on base PS5; when there's a lot of things going on the screen at one time. Granted the Series X manages 10-15 higher fps for those sections of the game, I'm guessing it would be a hard task to get that to a locked 60 with current hardware in the Pro.

Games like GTA VI I can imagine it still being locked to 30 or perhaps and maybe an unlocked performance mode, if we're lucky. It was 1440P 30 as per trailer and DF analysis. Seeing as how insane the density of assets, models, fidelity and physics going on in that game I can't imagine them getting to 60fps. At least without some massive compromises.

They do these tests at lower resolutions as to test the CPU. Otherwise it would become GPU limited. Its all scaling in the end. Plus even GPU intensive tasks like RT also uses a lot of CPU power.
 
Last edited:

SABRE220

Member
The CPU wouldn’t change anything at those frame rates.
Are you sure that the CPU would make absolutely no difference to the framerate? Also what about the second point the console seems to be performing signifigantly better than the stated 45% even in cpu heavy games like dragons dogma 2 with better settings/resolution.
 

Gaiff

SBI’s Resident Gaslighter
Are you sure that the CPU would make absolutely no difference to the framerate?
Yes. Or are we back to claiming that the games are held back by the CPU?
Also what about the second point the console seems to be performing signifigantly better than the stated 45% even in cpu heavy games like dragons dogma 2 with better settings/resolution.
No idea. Would need to know how the PS5 runs in those areas these days because there have been massive improvements to the CPU performance. I also think PSSR is involved? Realistically, they could keep the same resolution and upscale it to get a much clearer image and higher frame rates too. We would need to have a run in the city with matched content to see how they stack up to one another.
 
Last edited:

Nex240

Member
It's more than double of the PS5 Pro if you get a GPU of that performance with AI (so, minimum 4070, no AMD card in the market is equivalent), equivalent memory and SSD, plus controller, OS etc.

Over $1600 / 1600€ without counting keyboard, mouse and monitor.


Apparently Horizon gets native 4K in PS5 Pro (at 30fps), which provides better IQ than DLSS 4K in PC (at 60fps).
RTX 4070, not Super or Ti. Also 4070 just got a price cut to $399 Euro's. Where is this 1600 Euro's coming from.
 

yurinka

Member
RTX 4070, not Super or Ti. Also 4070 just got a price cut to $399 Euro's. Where is this 1600 Euro's coming from.
True. I got it wrong, I thought there were only the Ti and the non Ti one, so I thought the 4070 and the 4070 were the same but aren't.
 

SpokkX

Member
It will be entirely up to the developers to mask the anemic CPU in this machine, and I'm sure that some of them will do an amazing job at it, but others... not so much.
Yep

sony first party will surely do a good job. But we will se a lot of 30fps games later in this gen on pro (gta 6 comes to mind - too much simulation for that shit cpu)
 

twilo99

Member
Yep

sony first party will surely do a good job. But we will se a lot of 30fps games later in this gen on pro (gta 6 comes to mind - too much simulation for that shit cpu)

You will get a lot of "told you so" from both sides of the argument as the results swing depending on the dev team, but, minimizing the importance of the CPU is simply a "narrative" that a lot of people will fall for.

For me personally, a "best in class" gaming machine shouldn't rely on a CPU that is several generations behind already, especially if you consider that most people who buy these boxes will keep them for at least 4-5 years, at which point that Zen2 CPU will be 10 years old!
 

Topher

Identifies as young
yeah dig your hand in the sand and deny that this will be an issue going forward

This is not an attack on your favorite company - just stating what probably will be an issue with a product. You do not need to take offense

It will be left behind by 3-4 year old PC:s in CPU bound titles already. Imagine in 2 years..

Even without the GPU upgrades the CPU was likely going to be a problem for some games. Hell, my 5600x could not get to 60fps in Dragon's Dogma 2 in the cities so even if they had upgraded to Zen 3 it wouldn't have made much difference in games that really just need to be more optimized.

The upgrades with the GPU and PSSR are what people are buying. Had they upgraded the CPU then we would probably be looking at a $800 console instead of $700. Remember, this thing ain't being subsidized by games like base PS5,

Either way, you really should watch the video instead of depending on the guy tweeting it because the two don't match up.
 
Last edited:

Topher

Identifies as young
For me personally, a "best in class" gaming machine shouldn't rely on a CPU that is several generations behind already, especially if you consider that most people who buy these boxes will keep them for at least 4-5 years, at which point that Zen2 CPU will be 10 years old!

Nah.....the people who are buying PS5 Pro will more than likely be the first ones upgrading to PS6 in 2027/2028.
 
Top Bottom