• Hey, guest user. Hope you're enjoying NeoGAF! Have you considered registering for an account? Come join us and add your take to the daily discourse.

The PS5 worked fine with the slowest compatible SSD we could find

I think people are forgetting that R&C loads assets in&out on the fly, as you rotate the camera, which is an extremely demanding task since you have literally single milliseconds to do it, if not less. So if the slowest available drive on the market wouldn't be up to the task it would've been immediately visible on the screen with missing textures or entire objects, no benchmarks are needed for that.
When developers are presented with a new capability, it takes time to fully utilize it. First they have to wrap their heads around what they can do with it, then it takes time to implement it and iterate it and eventually push it to the limits.

With R&C we are seeing Insomniacs very early use of the PS5 hardware. For this reason it may indeed run "fine" on a 3500MB/s SSD. However anyone buying such an SSD will experience issues as we progress further into the generation. Guaranteed. Unless you have money to burn, buy a 7GB/s SSD. Or wait for the beta to conclude and then Sony will detail how they will handle the "too slow" problem. A too slow SSD may simply be locked out, or it may be restricted to PS4 games only. Just because they are allowing people to test slow SSDs in beta does not mean this will carry over once beta concludes.

Some developer quotes:
Behind the scenes, there’s so much to peel back about the SSD and the I/O around it. We’re just scratching the surface of it.
This title is the first one where we made the content knowing it would only ever be running on the PlayStation 5. And so our artists would say “What kind of mesh density can I have?” And I’d be like “... I don’t know.” Because we didn’t have the hardware. And we didn’t know, as the engine evolved, how the trade-offs would manifest themselves. Even once you have the hardware, it still takes you months or a year for your engine to evolve into it where you know how you want to spend your frame budget, what you do on the GPU versus the CPU, all that kind of stuff.
 

ZywyPL

Banned
Guaranteed.

Is it? Where does the assumption that the SSD will constantly run at its peak (sequential) 5.5GBps and not a single MB less comes from? Because again, random read speeds fall into hundreds MB/s performance (if even that given certain scenarios), where the difference between 3, 5, 6, and 7GBps drives is close to none, we're talking about mere couple-teens MB/s literally, and the extra I/O complex won't change than. So if anything, I'd say what's guaranteed is that PS5 internal drive will never be fully utilized, just like the CPUs and GPUs never are, unless there's an unlocked FPS mode, there's always a headroom left.
 
But some comparisons in here shows that the internal ssd ain't that fast as cerny and many others in here have stated.

But they aren't showing the drive being slower than the slowest SSDs if I'm not wrong. What I say with Digital Foundry is that a faster drive outperformed it in R@C I/O stress points. I haven't seen any evidence of slower drives (like the 3.9GB/s) outperforming it in the same tests.

Maybe there's been an update in trsting that I'm not aware of?
 

phil_t98

#SonyToo
I think from what I been reading in here is that slower SSD Will result in slower load times that’s about it. In terms of things like texture streaming and what not so far there doesn’t seem to be a difference using the slower drive. That could change of bee time and obviously it’s better to have the recommended speed
 
I think from what I been reading in here is that slower SSD Will result in slower load times that’s about it. In terms of things like texture streaming and what not so far there doesn’t seem to be a difference using the slower drive. That could change of bee time and obviously it’s better to have the recommended speed

There's a limit to how slow you can go. Like if you switched to a HDD you would have to load the entire level instead of streaming it in chunks. Which what Insomniac said they did with Ratchet on PS4.
 

Md Ray

Member
I think the PS5 reveal would've been vastly different if MS didn't show its cards first, but it is what it is, Sony had to switch the narrative to highlight their strong points like the SSD and Tempest audio, but the very moment Cerny said "it's dangerous to rely on TFlops and CU count should be avoided as well" you knew he's doing the damage control,
And yet we have many games on PS5 that are on par with XSX's 12TF w/ 52 CUs. Something that shouldn't have been possible with a 44% difference in CU count between PS5 and XSX.
 
Last edited:

Eliciel

Member
I believe we need a new teardown from CERNY and while he is at it maybe he delivers all kind of teardowns to explain to us why we are on the best plattform there is in this solar system
 

Bojanglez

The Amiga Brotherhood
I know DF briefly eluded to perhaps the native storage having to account for some background tasks, but never went any further on that. So has anyone seen any tests yet that take into account the following conditions on the I/O:
  • Speed comparisons of native and extended storage when a game is downloading in the background (to native storage)
  • Speed comparisons of native and extended storage when a game is downloading in the background (to extended storage)
  • Speed comparisons of native and extended storage when the console has native game recording switched on
  • Speed comparisons of native and extended storage when the console has native game recording switched off
and permutations of the above also taking into account some different external drive speeds? Basically, does writing to a drive simultaneously impact the speeds and if so, to what extent. Would be interested to know.
 

Md Ray

Member
Games built for a 18 CU console, tbh.
No, it's because CU count and TF are only one part of the equation. A GPU is comprised of many other components and sub-components, and these components all run at a much higher frequency on PS5 than XSX.

Why do you think Xbox One X, with only 11% more CU than PS4 Pro, had around 40% advantage in resolution quite often? Because the GPU frequency alone was 30% higher than Pro's GPU, and a massive 50% bandwidth advantage further elevated real-world GPU perf.
 

ZywyPL

Banned
And yet we have many games on PS5 that are on par with XSX's 12TF w/ 52 CUs. Something that shouldn't have been possible with a 44% difference in CU count between PS5 and XSX.

Since this year more and more games run at higher resolution on XSX thanks to beefier GPU and RAM bandwidth, I actually cannot recall when was the last time PS5 had a clear advantage in any comparison, Dirt 5? The Tools were a funny meme for some time, until it wasn't a meme anymore.
 
No, it's because CU count and TF are only one part of the equation. A GPU is comprised of many other components and sub-components, and these components all run at a much higher frequency on PS5 than XSX.

Why do you think Xbox One X, with only 11% more CU than PS4 Pro, had around 40% advantage in resolution quite often? Because the GPU frequency alone was 30% higher than Pro's GPU, and a massive 50% bandwidth advantage further elevated real-world GPU perf.
Performance is basically CU times frequency, assuming no bottlenecks. There is no secret sauce. Cross gen games still prefer narrow architectures because the base system is 18 CU. That's why the PS5 can keep up in some games.
 

Md Ray

Member
Performance is basically CU times frequency, assuming no bottlenecks. There is no secret sauce. Cross gen games still prefer narrow architectures because the base system is 18 CU. That's why the PS5 can keep up in some games.
And it'll still keep up in next-gen games, and in some, you'll see around 15-18% difference in titles with dynamic resolution engaged as it was expected since before launch. You won't see a 44% difference or anywhere close to it.
 

Riky

$MSFT
This means if you unlock the frame rate, you'll consistently see lower frame rate on XSX compared to PS5.

But it holds 60 FPS for like 95% of the game, then you have VRR anyway.
The Devs obviously decided what was best for both machines, 4k 60fps for Series X and 1800p 60fps for PS5, if they thought PS5 could do 4k 60fps they would have tried, they obviously decided the framerate would be a lot worse so settled on 1800p.
We've seen s constant trend of higher resolutions in Series X versions after the launch games.
 

Md Ray

Member
Since this year more and more games run at higher resolution on XSX thanks to beefier GPU and RAM bandwidth, I actually cannot recall when was the last time PS5 had a clear advantage in any comparison, Dirt 5? The Tools were a funny meme for some time, until it wasn't a meme anymore.
PS5's resolution turning up to be on par XSX in the last couple of games comparison.
 

Snake29

RSI Employee of the Year
Since this year more and more games run at higher resolution on XSX thanks to beefier GPU and RAM bandwidth, I actually cannot recall when was the last time PS5 had a clear advantage in any comparison, Dirt 5? The Tools were a funny meme for some time, until it wasn't a meme anymore.

Maybe 1 or 2 games, but even in those games it was small and PS5 had better fps without stutters or frame pacing issues.

XSX isn’t showing what MS and fans were hoping for before launch. If this is what you were expecting from the XSX then it looks disappointing compared to what the PS5 already can do. Tools look still like a meme since none of the games showing any big differences with PS5 games, heck the PS5 is showing mostly better textures, higher res or sharper graphics settings in some games. Again if only a small res difference will be the only big difference this whole gen…….lol.

All that hyperbole from Xbox and their fans.
 
Last edited:

Concern

Member
This means if you unlock the frame rate, you'll consistently see lower frame rate on XSX compared to PS5.


Barely a difference especially considering Ps5 is usually doing 1440p - 1800p if not lower when it a frame advantage. But spin away :messenger_winking:
 

Md Ray

Member
But it holds 60 FPS for like 95% of the game, then you have VRR anyway.
The Devs obviously decided what was best for both machines, 4k 60fps for Series X and 1800p 60fps for PS5, if they thought PS5 could do 4k 60fps they would have tried, they obviously decided the framerate would be a lot worse so settled on 1800p.
We've seen s constant trend of higher resolutions in Series X versions after the launch games.
They were simply being conservative with the PS5 version. The frame rate is clearly at least 46% higher on the PS5 in Mendoza level. So I'd say it'd be a tie even in HM3 if the resolution were to be equal.
 
“bUt xBoX wAs sUpPoSeD tO bE a rEaL 4k cOnSoLE”

None of that is close then true.
4K console vs 1080P console

That definitely wasn't the case.

With that said I don't think it's wrong for devs to use dynamic resolution. But I do know that in those case the XSX will tend to have a higher resolution than the PS5. But not so much more that it makes the PS5 look weak by any means.
 

Topher

Gold Member
Doom Eternal it wasn't. I'm sure PS5 dipped below 1080p at points.

It did? According to VG Tech
"PS5 in the 120fps Mode uses a dynamic resolution with the highest resolution found being 2816x1584 and the lowest resolution found being approximately 1992x1120.

Xbox Series X in the 120fps Mode uses a dynamic resolution with the highest resolution found being 3200x1800 and the lowest resolution found being approximately 2266x1275. "

 

Concern

Member
Does Hitman 3 have dynamic res?

Starting at the top, Xbox Series X runs the game at native 4K resolution (incidentally doubling frame-rate and increasing pixel count over Xbox One X's 1440p). Just beneath this sits PlayStation 5, running at 1800p. Meanwhile, Xbox Series S aims for the same 60fps target, but this time the GPU is tasked with running the game at 1080p resolution. There's no dynamic resolution scaling from what we can see, but there is a smart pass of TAA anti-aliasing on all systems.
 

Md Ray

Member
Doom Eternal it wasn't. I'm sure PS5 dipped below 1080p at points.
The res difference comes out around 15-18% between them, and that's with VRS engaged on SX. I imagine the difference would have been even less than that without it. About what we expected from these machines before launch, as I said. I still see no 44% difference though.
 
Maybe 1 or 2 games, but even in those games it was small and PS5 had better fps without stutters or frame pacing issues.

XSX isn’t showing what MS and fans were hoping for before launch. If this is what you were expecting from the XSX then it looks disappointing compared to what the PS5 already can do. Tools look still like a meme since none of the games showing any big differences with PS5 games, heck the PS5 is showing mostly better textures, higher res or sharper graphics settings in some games. Again if only a small res difference will be the only big difference this whole gen…….lol.

All that hyperbole from Xbox and their fans.
?

It is quite common now the same game always pushing higher resolution on the XSX than on the PS5. And the res difference in absolute numbers is significant.
 

Topher

Gold Member
?

It is quite common now the same game always pushing higher resolution on the XSX than on the PS5. And the res difference in absolute numbers is significant.

How is it significant when we depend on folks like DF or VG Tech to tell us what the difference is? Even DF has questioned the need for all this pixel counting when we can't see the difference ourselves.
 
Last edited:
Is it? Where does the assumption that the SSD will constantly run at its peak (sequential) 5.5GBps and not a single MB less comes from? Because again, random read speeds fall into hundreds MB/s performance (if even that given certain scenarios)
The SSD will not "constantly run at it's peak" because that's not a realistic scenario.

However it delivers repeatable performance. This is not a HDD with wildly varying performance. It is also not the SSD in your computer where environmental variables can affect benchmarks. It delivers deterministic performance and games will eventually come to rely on this performance, pulling assets just-in-time. If the assets are late then you're going to get hitching or placeholder textures, or worse.

where the difference between 3, 5, 6, and 7GBps drives is close to none
This is a strange assertion, if a 3GB/s drive delivers the same performance as a 7GB/s drive then one of them is falsely marketed.

So if anything, I'd say what's guaranteed is that PS5 internal drive will never be fully utilized, just like the CPUs and GPUs never are, unless there's an unlocked FPS mode, there's always a headroom left.
Highly optimized games on consoles approach full utilization of the CPU and GPU. That does not mean the CPU and GPU are pegged 100% of the time. It means you can not (hypothetically) substitute a 40% slower CPU into your PS5 and expect games to run as normal.

This is akin to what you are suggesting. The performance of the I/O complex is (or will be) as fundamental to the system as memory bandwidth or compute power. To think that you can substitute a 40% slower SSD and everything will be fine is to completely misunderstand the hardware.

It is your money; buy what you like. Anyone who buys a 3500MB/s SSD because The Verge said so will feel very silly when we get into this generation proper.
 
Last edited:

Thirty7ven

Banned
It’s quite incredible how the same cast of characters keep spreading the same fud and yet they keep getting feedback.
 

Riky

$MSFT
The res difference comes out around 15-18% between them, and that's with VRS engaged on SX. I imagine the difference would have been even less than that without it. About what we expected from these machines before launch, as I said. I still see no 44% difference though.

Hitman 3 is, DF saw less than 1080p and said PS5 was blurrier in Doom Eternal.
 
Last edited:
So people are still hoping that the XSX is massively more powerful?
No, I think just the expected 15ish%
They were simply being conservative with the PS5 version. The frame rate is clearly at least 46% higher on the PS5 in Mendoza level. So I'd say it'd be a tie even in HM3 if the resolution were to be equal.
In the flowers, not all the time. Xbox also hasn't had transparency issues lately so maybe it was a bug/tools issue.
 
No, I think just the expected 15ish%

Yeah around 15%-25% is what I'm seeing for the most part. Never denied that judt saying there isn't a massive difference between the two. And you really can't expect it given how similar the two are in specs.

Honestly I'm just tired of seeing Mr X theories constantly surfacing time and time again.
 
Last edited:

phil_t98

#SonyToo
There's a limit to how slow you can go. Like if you switched to a HDD you would have to load the entire level instead of streaming it in chunks. Which what Insomniac said they did with Ratchet on PS4.
We ain’t talking about HDD though we are talking what is capable of being plugged into the expansion slot on ps5. Why swap what I said to that? As I said it seems the slower SSD is working with games fine and no issue playing or texture streaming so far. Maybe a slightly slower load time that’s it
 

phil_t98

#SonyToo
So people are still hoping that the XSX is massively more powerful?

the same argument can be levelled at the SSD, as I said in a post earlier it’s to early to tell if the power difference will matter later in the gen between raw grunt of graphics or if the SSD will be better. So far we ain’t seeing much difference in both so far
 
Gen 3 Max speed is 3.9gbs, so the same speed that was tested. Nothing shown proves Ratchet isn't possible on a 2.4gbs drive. Obviously nothing proves it does either, but there is zero confirmation of anything really.
The thing is, that SSD isn't even running at constant 3.9gbs sustained speed . 3.9 is the MAX speed. Its running at maybe 1-3.0 sustained speed. While XS runs at constant 2.4 sustained speed.

Its quite obvious that people who were preaching that PS5's SSD has 12 billion channels and R&C is bringing the SSD to its knees and wont even run on XS and you need a 7GB/s to even match the performance because of all the priority channels were clueless. This isn't any different than the UE5 scenario.
 
Last edited:

Tiamat2san

Member
not-so-secret-special-sauce_edited-1.jpg
 

Md Ray

Member
Hitman 3 is, DF saw less than 1080p and said PS5 was blurrier in Doom Eternal.
But 46% higher framerate on PS5 in Hitman 3 though. In DOOM Eternal, PS5 can be blurrier in one mode but the other two modes are clearly sharper on PS5 due to the absence of VRS. So we're seeing the expected 15-18% difference, and both on par in other games, not 44%.
 
Last edited:
We ain’t talking about HDD though we are talking what is capable of being plugged into the expansion slot on ps5. Why swap what I said to that? As I said it seems the slower SSD is working with games fine and no issue playing or texture streaming so far. Maybe a slightly slower load time that’s it

Yeah but are we really comparing vastly slower drives than what the PS5 has?
 
the same argument can be levelled at the SSD, as I said in a post earlier it’s to early to tell if the power difference will matter later in the gen between raw grunt of graphics or if the SSD will be better. So far we ain’t seeing much difference in both so far
Well a 3090 is already vastly more powerful than a 3060.

You didn't have to wait several years to see the results.

If a system is vastly more powerful than another the results should be be visible at launch. Kind of like the XSS vs the XSX or the PS5 compared to the XSS.
 
Top Bottom