• 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.

PS5 focus is high performance, not high power. Are people underestimating the value of the approach?

VFXVeteran

Banned
Not if the amount of pixels are the same. Or do you mean post-processing?

Maybe I didn't make myself clear.

At the end of the day, we all want to see beautiful graphics at great framerates. This incorporates the entire hardware subsystem and it's components as well as good code that allows them to communicate with each other with the utmost of efficiency. The hard limit is the theoretical bandwidth of the hardware system which is 10.2TFLOPS for the PS5.

I'm saying no matter what kind of hardware you have or how fast any of the sub-components may be, your limit is that bandwidth number. So if you make something that tries to push enormous data (i.e. a full on ray-traced hair model using curved primitives for example) that the GPU can't process at a reasoanable framerate @ a certain resolution, then it won't matter how fast the SSD is.
 

CJY

Banned
You *do* care about the pixels because that's what's being shown to the gamer.

Why would 2 code paths need to be used simply because of an SSD while everything else is pretty much par the course (i.e. CPU, GPU, memory pool, etc..)? Give an example of what problem you are running into.
Pixels are important if you've got too few, but the importance of pixels next gen is greatly diminished the higher up the resolution ladder we go.

Sure, as a VFX artist, it makes you're placing undue importance on pixels, but if you were an engine developer or a level designer and therefore looking at things through a different lens, I think you wouldn't be downplaying the impact of having an SSD in every system and developing a game specifically targeting that spec.

You keep trying to spread the narrative of PS exclusive games coming to PC, and nobody is disputing that as we've already seen some, with more on the way and execs have come out and said already that it is something they are exploring further, but it makes no sense for a 1st party dev to hamper themselves by tying a next-gen game to the IO throughput of a HDD, which is what they would have to target on PC. Not every single dev will take advantage of the SSD fully, and the games from these devs have the potential of being ported to PC, but there is also no way some other devs with handicap themselves when SSD penetration in PC is so low.
 

VFXVeteran

Banned
Pixels are important if you've got too few, but the importance of pixels next gen is greatly diminished the higher up the resolution ladder we go.

Sure, as a VFX artist, it makes you're placing undue importance on pixels, but if you were an engine developer or a level designer and therefore looking at things through a different lens, I think you wouldn't be downplaying the impact of having an SSD in every system and developing a game specifically targeting that spec.

You keep trying to spread the narrative of PS exclusive games coming to PC, and nobody is disputing that as we've already seen some, with more on the way and execs have come out and said already that it is something they are exploring further, but it makes no sense for a 1st party dev to hamper themselves by tying a next-gen game to the IO throughput of a HDD, which is what they would have to target on PC. Not every single dev will take advantage of the SSD fully, and the games from these devs have the potential of being ported to PC, but there is also no way some other devs with handicap themselves when SSD penetration in PC is so low.

I am a graphics programmer and not an artist.
 

Croatoan

They/Them A-10 Warthog
Another thread?
OP was one of the main offenders on the speculation thread. Dude was delusional from the very beginning.

Out of curiosity how do you get performance without power? I guess they can run their games 1080p upscaled and claim "performance", but if you want 4k 60fps "performance" you actually need more than 13 floppy flops (power). And that is with a lot of current gen games. The 2080ti cannot even run everything maxed 4k 60fps for example.

If you think running games at 30fps = performance then...

HAHAHAHAHAHAHAHAHAHAHAH

BTW, OP, how is that crow?
 
Last edited:

MHubert

Member
Maybe I didn't make myself clear.

At the end of the day, we all want to see beautiful graphics at great framerates. This incorporates the entire hardware subsystem and it's components as well as good code that allows them to communicate with each other with the utmost of efficiency. The hard limit is the theoretical bandwidth of the hardware system which is 10.2TFLOPS for the PS5.

I'm saying no matter what kind of hardware you have or how fast any of the sub-components may be, your limit is that bandwidth number. So if you make something that tries to push enormous data (i.e. a full on ray-traced hair model using curved primitives for example) that the GPU can't process at a reasoanable framerate @ a certain resolution, then it won't matter how fast the SSD is.
Would it matter if the speed of the SSD (and its custom solutions) freed up CPU and GPU resources?
 

MHubert

Member
It doesn't work that way. The CPU and GPU need to be working all the time at full capacity. So those two will always wait on the SSD (5.5GB/s vs. 448GB/s).
Okay, then I must have misunderstood something, because my takeaway from the presentation was that one of the major benefits of Sony's solution was to minimize the workload required to transfer and/or decompress data from the ssd to ram at high speeds in realtime.
So I guess I could rephrase my question a bit: Will XBSX be able to reach its theoretical throughput of 12.1tf while maintaining a full speed transfer from ssd to ram? Honest questions, you seem to know way more about this than I do.
 
Last edited:
I'm not sure if you're trolling right now. What in the two words "custom silicon" is not clear enough?
PS5 has custom silicon pretty much everywhere, except CPU and RAM. In every other component.
Even GPU has new coherency engines the "scrubbers". They put it there just for fun, I suppose?
The PS5 is not the only one with custom silicon everywhere, no?...


"The form factor is cute, the 2.4GB/s of guaranteed throughput is impressive, but it's the software APIs and custom hardware built into the SoC that deliver what Microsoft believes to be a revolution - a new way of using storage to augment memory (an area where no platform holder will be able to deliver a more traditional generational leap). The idea, in basic terms at least, is pretty straightforward - the game package that sits on storage essentially becomes extended memory, allowing 100GB of game assets stored on the SSD to be instantly accessible by the developer. It's a system that Microsoft calls the Velocity Architecture and the SSD itself is just one part of the system.

"Our second component is a high-speed hardware decompression block that can deliver over 6GB/s," reveals Andrew Goossen. "This is a dedicated silicon block that offloads decompression work from the CPU and is matched to the SSD so that decompression is never a bottleneck. The decompression hardware supports Zlib for general data and a new compression [system] called BCPack that is tailored to the GPU textures that typically comprise the vast majority of a game's package size."

If I remember correctly, so did the x and pro in places other than the cpu and gpu... As is normal for recent consoles is it not? I don't know much about the classics.

Somehow (Through forward thinking coding) MS has found a way to make their software's code work on both their consoles and PC hardware in a way that's easy to port and that scales amazingly well... Are you trying to say that Sony doesn't have the same chops?

The reality is that if Sony are indeed planning to bring their software to ANY hardware outside of their new box at ANY point in time, then they would plan ahead with their code to make it as smooth and scalable a process as possible. Less they want to run into a cell situation again with all the pains that brought them, which I doubt.

Also, it would benefit both Sony and third party devs greatly the easier it is to port code between their "custom silicon" console hardware and PC(and xbox hardware by proxy). Third party devs, since they are going to want their code to run on as much hardware as possible with as little dev time to port as possible. And Sony, because they want third party games to look and perform best on their console. But that only happens if it's easy and cost effective to port the code to other hardware. Else, third party devs probably won't waist their time...
 

VFXVeteran

Banned
Okay, then I must have misunderstood something, because my takeaway from the presentation was that one of the major benefits of Sony's solution was to minimize the workload required to transfer and/or decompress data from the ssd to ram at high speeds in realtime.
So I guess I could rephrase my question a bit: Will XBSX be able to reach its theoretical throughput of 12.1tf while maintaining a full speed transfer from ssd to ram? Honest questions, you seem to know way more about this than I do.

Yes. This is not new. The GTX 2080Ti can push higher bandwidth and most PCs had regular HDDs. The SSDs still sped up loading.

The advantages that Cerny was talking about was comparing PS5 to last gen's really slow I/O PS4. In that case, yes, it will greatly speed up the pipeline. Now that you have a very fast SSD (5.5G/s) you can build larger worlds, etc.. but that doesn't mean the graphics pipeline will speed up. The graphics pipeline needs as much bandwidth as it can swallow. 5.5G/s is way too slow compared to 448G/s memory bandwidth. The idea is to get as much data into resident RAM so that the GPU/CPU can do their thing. PS4 had a bottleneck at the I/O level because of it's slow HDD. But it is still limited to a bandwidth number as far as the graphics pipe is concerned. The PS5, XSX, and PC are no different.
 
Last edited:

NickFire

Member
I think Sony needs to actually and tangibly show us its approach in action before its fair to opine on whether someone underestimates its value. Right now what do we really know? In my opinion all we really know is:

1) Less capable in terms of raw power;
2) Faster SSD; and
3) Heavy audio emphasis.

Also - please have an optical port Sony. All this talk about audio will be completely muted to me and many others if there is no optical out port. And that is a huge strike against SeX for me. Will be impossible for me to get 5/1 on SeX unless I buy a new TV with 5.1 optical out pass through (mine only does 2 channel out) or an entirely new receiver.
 

psorcerer

Banned
Are you trying to say that Sony doesn't have the same chops?

Nope. I'm trying to say that the sentiment "PS5 does clever things in hw only because they don't have the powah, mwahaha!" is a pretty bad idea.
And that's what I see here.

The reality is that if Sony are indeed planning to bring their software to ANY hardware outside of their new box at ANY point in time, then they would plan ahead with their code to make it as smooth and scalable a process as possible. Less they want to run into a cell situation again with all the pains that brought them, which I doubt.

That was not the case that was discussed.
The case was: PS5 will go PC-first and then "port" to their custom arch.
Which is exactly throwing out all of the careful engineering down the drain.
 

psorcerer

Banned
5.5G/s is way too slow compared to 448G/s memory bandwidth.

It's kind of apples to oranges though.
Your 448GB/sec is spent in the "working set", i.e. various render targets, deferred buffers, shadows, etc. etc.
All the data that is essentially re-created from scratch each and every frame.
Most of the bandwidth and compute is "wasted" there.
5.5GB/sec is the pipe from the artistic vision into your frame.
Essentially 5.5 is the "input bandwidth" and 448 is the "output bandwidth".
Last gen the "input" was so small, that the most of the games opted for "input arenas": you load your input into "arena" once and then use it all the time, until you load another chunk of input.
If we want to port PS5 (and probably XSeX) game to PC that's the only approach that will work. You reserve a large amount of RAM (arena) and then you render from that amount.
If there is a teleport/fast travel, you wait N sec on SSD, or 5xN sec on HDD for it to load. And continue.
The PC approach and "console" approach are not compatible here: PC can have 24GB of RAM for that "input arena" (~50sec on a typical PC SSD, 5 min on a typical HDD) , but PS5 can pump that 24GB in <5 sec, i.e. it can have no "arena" at all, just pump it continuously to the pages that need it.
There is nothing common here.
But, obviously you can make that common, using the slowest part. Make "arenas" for PS5 too, just smaller, XSeX slightly bigger, and PC even bigger.
But that's not playing to the console strength. It's the "most common denominator" approach.
 
>You say 'substantially more powerful' but most tech experts have clearly said the differences will likely be a slightly higher frame-rate and higher resolution. You are trying to make a mole-hill into a mountain.

A) I don't particularly understand why I have to listen to sources. We know the specs of both consoles, and at the end of the day they are computers, computers optimized for gaming, but computers. The difference between the two consoles is the difference between being outdated by 1-2 years.

B) Slightly is a flat out lie. We are looking at a PS4 and a half power level difference. And frankly if either console fails to get 4k at a solid framerate for most games, that console is a failure. The difference between the one x and the ps4 pro is 1.8 tfs, and we saw how large the gap could be. The difference here is bigger, FAR bigger if you consider the PS5 is only hitting double digits because of overclocking.

>If Sony wanted to, they could have made the exact same console as MS for the exact same price. Why not discuss why each company made the decisions they did, rather than constantly shouting 'damage control', 'butt-hurt', etc.

Sony made their decision for profit, MS made their decision to push tech forward and have the most powerful console on the market. We have, on multiple occasions, seen Sony get cocky after a period of success only to make decisions that baffle consumers. MS is taking a massive plunge forward, while Sony is doing a marginal upgrade.

>In the end, Sony is trying to make the best console for a specific price (MS did as well) and they both had to make some trade-offs. The power difference only becomes interesting if they are same price and I'd be willing to bet dollars to donuts that they aren't.

This is such a cop out. We get one new main console every 6+ years. A 50-100 dollar price difference shouldn't really be a barrier for anyone, that's a month or two of extra saving. I mean shit, if it's a hundred dollar difference, get the Series X since it will play all xbox one games anyway, and there is your price difference. When the PS4 had the ever slightest, and I mean SLIGHTEST advantage over the one, people couldn't shut up about how massive a difference 900 to 1080p was. Now a much bigger power difference is irrelevant?

>In the end, the system with the best games always wins.

I mean MS just last year went on a massive buying spree and bought a load of massively talented developers. If only half of them produce worthwhile games, we are still looking at a much closer exclusive game comparison. Not to mention MS studios like Rare that seem to be on the right track, or Halo Infinite which has the biggest budget in gaming ever, to date. Coalition seems to be getting much better at gears as well. I just don't buy it. PS5 is under powered hardware and the two developers under Sony I actually care about seem to be blowing their load this year anyway, with TLOU2 and Sucker Punches new IP. Only game I'd buy a PS5 day one for is Bloodborne 2, which I doubt.
LOL, I can't tell if you are being serious; you either are trolling or being intentionally obtuse....or maybe both. Instead of addressing all of the inaccuracies in your post, what do you think will be the difference in Frame-Rate and Resolution from this 'Substantial' power difference?

To simplify things, it will help you be more realistic to think about power differences in percentages vs raw flops. The XSX will have about 10x TF of XOne; do you expect the XSX to be able to push 10X the pixels?
 

VFXVeteran

Banned
It's kind of apples to oranges though.
Your 448GB/sec is spent in the "working set", i.e. various render targets, deferred buffers, shadows, etc. etc.
All the data that is essentially re-created from scratch each and every frame.
Most of the bandwidth and compute is "wasted" there.
5.5GB/sec is the pipe from the artistic vision into your frame.
Essentially 5.5 is the "input bandwidth" and 448 is the "output bandwidth".
Last gen the "input" was so small, that the most of the games opted for "input arenas": you load your input into "arena" once and then use it all the time, until you load another chunk of input.
If we want to port PS5 (and probably XSeX) game to PC that's the only approach that will work. You reserve a large amount of RAM (arena) and then you render from that amount.
If there is a teleport/fast travel, you wait N sec on SSD, or 5xN sec on HDD for it to load. And continue.
The PC approach and "console" approach are not compatible here: PC can have 24GB of RAM for that "input arena" (~50sec on a typical PC SSD, 5 min on a typical HDD) , but PS5 can pump that 24GB in <5 sec, i.e. it can have no "arena" at all, just pump it continuously to the pages that need it.
There is nothing common here.
But, obviously you can make that common, using the slowest part. Make "arenas" for PS5 too, just smaller, XSeX slightly bigger, and PC even bigger.
But that's not playing to the console strength. It's the "most common denominator" approach.

Which will happen 90% of the time.

For exclusives that don't show on the PC, sure, it can load up "arenas" faster, but it better have the output to sustain at least 30FPS and it better be at a true 4k.

Bottlenecks are a moving target. If you free up one side of the pipe, then the other side becomes the bottleneck. This is nothing new.

To say that the PS5s SSD will somehow be working as a graphics pipeline secret sauce is absurd. I can already predict people equating beautiful artistic technology exclusive only on the PS5 to it's SSD advantages just to feel good about their platform of choice. It's ridiculous. If the shoe was on the other foot and PS5 had 13TFLOPS of performance, we wouldn't get anywhere near the attention on an SSD drive as we are now.
 
giphy.gif
 

bitbydeath

Member
And you know that how?

Because of that Spider-Man tech demo from last year? You are seriously comparing tech-demo to an actual unoptimized build of State of Decay 2?

Two different "products"?

It was reiterated that there would be no load screens and content loading at a turn in the road to PS5.
 

bitbydeath

Member
You can agree to disagree with him, but it doesn’t change the fact that you are just wrong in what you’ve been saying. Your OP is all about the ssd, whether you are aware of it or not.

The point is that they could just stick an ssd in there that is capable of 100 times faster speeds, but without additional customisations to the pipeline it would not be able to give anything like that sort of increase to loading and streaming, as the cpu in particular would bottleneck it.

The opening quote sets the focus of the rest of the content.

For PlayStation 5 our goal was not just that the SSD itself be a hundred times faster. It was that game loads and streaming would be a hundred times faster

Game loads and streaming are seperated from the SSD because as it later goes on there is custom hardware to assist the operations.

This is why the same drive on PC doesn’t result in 0 load times.

And streaming (displaying the image on screen) is the real interesting component.
 

bitbydeath

Member
OP was one of the main offenders on the speculation thread. Dude was delusional from the very beginning.

Out of curiosity how do you get performance without power? I guess they can run their games 1080p upscaled and claim "performance", but if you want 4k 60fps "performance" you actually need more than 13 floppy flops (power). And that is with a lot of current gen games. The 2080ti cannot even run everything maxed 4k 60fps for example.

If you think running games at 30fps = performance then...

HAHAHAHAHAHAHAHAHAHAHAH

BTW, OP, how is that crow?

I can eat crow with the best of em.
I was following what I thought was the logical outcome given the information we had at the time and I was wrong.

I have no problems admitting it.

This thread is about delving into the real interesting information that Mark spoke about which wasn’t being discussed here at the time.

Of course many go back to focus on the SSD
 
Last edited:

Jonsoncao

Banned
I can already predict people equating beautiful artistic technology exclusive only on the PS5 to it's SSD advantages just to feel good about their platform of choice. It's ridiculous. If the shoe was on the other foot and PS5 had 13TFLOPS of performance, we wouldn't get anywhere near the attention on an SSD drive as we are now.

I doubt there are people dumb enough to claim this, but I will mark your word.
 
Last edited:
Maybe I didn't make myself clear.

At the end of the day, we all want to see beautiful graphics at great framerates. This incorporates the entire hardware subsystem and it's components as well as good code that allows them to communicate with each other with the utmost of efficiency. The hard limit is the theoretical bandwidth of the hardware system which is 10.2TFLOPS for the PS5.

I'm saying no matter what kind of hardware you have or how fast any of the sub-components may be, your limit is that bandwidth number. So if you make something that tries to push enormous data (i.e. a full on ray-traced hair model using curved primitives for example) that the GPU can't process at a reasoanable framerate @ a certain resolution, then it won't matter how fast the SSD is.
That was very insightful. Thank you. Warning, you're probably going to be labeled as biased by many who don't like what you have to say.
 

Three

Member
They did it probably to save 30$ per console. They gave away 20% of their performance for 7% of their BOM. Not very smart in my opinion.

Imagine all of this optimization + 3d audio + superfast SSD in a 52 CU console. That would have been killer.

According to Sony's own calculations, launch year buyers spend on average 1,600$ during the lifetime of the console. They would have made those 30$ back in the blink of an eye.
Where do you stop though? Why not another $30 on top of that. They concentrated on nailing the architecture rather than just bigger/more CUs. They can always do bigger later in a pro.
 

psorcerer

Banned
we wouldn't get anywhere near the attention on an SSD drive as we are now

What's disingenuous.
Why? Because RDNA2 GPU has nothing that wasn't already done by NV or AMD earlier.
There's no new technology there whatsoever.
We have seen everything that AMD APU can offer in 2018.
SSDs on other hand in both XSeX and PS5 are a new technology.
There's nothing like that on PC, and it will not be there until end of the year, at least.
That's why it has a lot of attention and it should have it.
 
What's disingenuous.
Why? Because RDNA2 GPU has nothing that wasn't already done by NV or AMD earlier.
There's no new technology there whatsoever.
We have seen everything that AMD APU can offer in 2018.
SSDs on other hand in both XSeX and PS5 are a new technology.
There's nothing like that on PC, and it will not be there until end of the year, at least.
That's why it has a lot of attention and it should have it.
You clearly don't understand the point he is making. 1. The PS5s faster SSD is not going to improve graphics etc due to limited bandwidth. There is not enough bandwidth to feed the fast SSD speeds while targeting 4k resolution and a good framerate. People are taking what Mark Cerny said out of context

2. He is simply stating that the reason many sony fans on the interwebs hyping the PS5s SSD so much is because you know that its hardware specs are inferior to the XSX. The SSD is the scapegoat.
 
Last edited:

psorcerer

Banned
1. The PS5s faster SSD is not going to improve graphics etc due to limited bandwidth. There is not enough bandwidth to feed the fast SSD at 4k resolution and a good framerate.

I'm not sure how is it related? "framerate" and "ssd bandwidth"?

2. He is simply stating that the reason many sony fans on the interwebs hyping the PS5s SSD so much is because you know that its hardware specs are inferior to the XSX. The SSD is the scapegoat.

Which is disingenuous. Bceause SSD implementation is really fast and should be discussed.
 

psorcerer

Banned
Yes, but not for making up the power deficit that the PS5 has vs XSX. The SSD will be more beneficial for loading content, not taking the place of Teraflops,gpu,cpu,ram,bandwidth etc...

The whole discussion here started because of a claim that future PS5 games will use PC as their primary target platform.
I'm not sure how it got to teraflops, actually...
 

VFXVeteran

Banned
The whole discussion here started because of a claim that future PS5 games will use PC as their primary target platform.
I'm not sure how it got to teraflops, actually...

Dude, come on man! You are being naive on purpose.

If you work in the industry on Sony products, you already know the rumors. If you don't, then why are you laying claim that what I've heard is false? There will be PS5 exclusives that will end up coming to the PC. We just have to wait and see. If those games don't take advantage of the SSD for the sake of making better graphics features that the PS5 can't possibly muster, so be it. PC owners won't be very envious this generation for sure. You can take that to the bank.
 
Last edited:

John254

Banned
It was reiterated that there would be no load screens and content loading at a turn in the road to PS5.
Yup. Because Cerny knows for sure how devs will utilize the platform.

Jesus just wake up. Every generation is the same. Platform holder will promise something which can be done with platform but devs are devs and they always find the way to break that promise.

And about that video you just send. In the description is is said that this is unoptimized straight from back compact version of State of Decay 2. So no optimalization for Velocity architecture. I'm sure that native XsX games will perform even better in terms of loading times.

But nice try...
 
The Xbox one had a faster CPU and also faster clocks on the GPU side and it was still under powered compared to the PS4.

Sony can't OC it's way to close the power gap
 
That's true. The question is: when and the second one is: how.
A prediction without a date is worthless. 😅
How? No different than how PS4 games have gone to PC. Or how next gen MLB The Show games will be on XSX and other platforms.

Your mistake is believeing that future PS5 games are going to be developed differently in some new alien/radical new way. You're setting yourself up for disappointment.
 
Last edited:
I am a graphics programmer and not an artist.
My issue is are gpus currently being underutilized? I've seen some extremely sharp textures and extremely detailed models running on current gen, but often used sparingly.

For example we know that a rtx2080 ti with say just 500MB of vram would have inferior graphics to one with 4GB. Yet the site I linked previously suggest the 4GB 580 gpu can run most games at ultra setting(far higher than consoles), the rtx2080ti can't run games at higher than ultra detail. That means essentially at 1080p and 1440p the rtx 2080ti is handicapped to no more graphical detail than what 580 4GB gpu can run in most current gen games(despite having 11GB of ram.).

Why are current gen games using less than 4GB of graphics? Probably because of the limits of the HDD.

Next gen consoles can use and fill 16GB quickly due to the ssds. We know that going from 500MB to 4GB there's plenty of graphical difference presumably. But going from current less than 4GB use to 8+GB won't there be a graphical difference? The xbox has sampler feedback for example which effectively multiplies available memory 2-3 fold, is that excessive and useless? Or can we expect exponentially more detail next gen due to that memory multiplier?
There is not enough bandwidth to feed the fast SSD speeds while targeting 4k resolution and a good framerate. People are taking what Mark Cerny said out of context
I assume many next gen games will use ai reconstruction 4k and run at less than 4k. So bandwidth is plenty for that, and won't result in perceptible differences.
 
Last edited:

PocoJoe

Banned
The PS5 is not the only one with custom silicon everywhere, no?...

This thread is about PS5.

What the heck is wrong with you people?

"I think it is cool that PS5....."

"Bbbbut xbox have this and that, and 12 tflops!!" -all insecure childish xbots

Why it is so difficult to talk about PS5 and it's tech without having 10 xbox fanatics crying all around?

And then they have the nerve to claim everything is just damage control.

I guess it is this old phenomenom:

Xbox fanboys are so used to do damage control, that they really cant understand that some people just want to talk about the tech and what it can do, not compare their dicks?

Just unbelievable how many xbots spam their low IQ console warriorism and disgard everything as damage control.

Why there would even need to have damage control? It is clear that xbox have more raw power, rest is uncertain until games are out

Gen havent even started and there is no real damage, specs are specs and people dont just walk into store and ask "console with most tflops please"
 
Last edited:
The Xbox one had a faster CPU and also faster clocks on the GPU side and it was still under powered compared to the PS4.

Sony can't OC it's way to close the power gap
Yup
My issue is are gpus currently being underutilized? I've seen some extremely sharp textures and extremely detailed models running on current gen, but often used sparingly.

For example we know that a rtx2080 ti with say just 500MB of vram would have inferior graphics to one with 4GB. Yet the site I linked previously suggest the 4GB 580 gpu can run most games at ultra setting(far higher than consoles), the rtx2080ti can't run games at higher than ultra detail. That means essentially at 1080p and 1440p the rtx 2080ti is handicapped to no more graphical detail than what 580 4GB gpu can run in most current gen games(despite having 11GB of ram.).

Why are current gen games using less than 4GB of graphics? Probably because of the limits of the HDD.

Next gen consoles can use and fill 16GB quickly due to the ssds. We know that going from 500MB to 4GB there's plenty of graphical difference presumably. But going from current less than 4GB use to 8+GB won't there be a graphical difference? The xbox has sampler feedback for example which effectively multiplies available memory 2-3 fold, is that excessive and useless? Or can we expect exponentially more detail next gen due to that memory multiplier?

I assume many next gen games will use ai reconstruction 4k and run at less than 4k. So bandwidth is plenty for that, and won't result in perceptible differences.
I don't think so. Pretty sure Native 4k will be a common thing next gen.
 

bitbydeath

Member
Yup. Because Cerny knows for sure how devs will utilize the platform.

Jesus just wake up. Every generation is the same. Platform holder will promise something which can be done with platform but devs are devs and they always find the way to break that promise.

And about that video you just send. In the description is is said that this is unoptimized straight from back compact version of State of Decay 2. So no optimalization for Velocity architecture. I'm sure that native XsX games will perform even better in terms of loading times.

But nice try...

Why are you so defensive?
I’m presenting content for discussion, someone else asked if XSX could do it and I said they had different goals and proved why. Different goals doesn’t mean it’s bad, not everything has to be console wars.

Can’t we just have a technical discussion without stating who you think is better?
 

adamosmaki

Member
Really PS fanboys ? How many threads are you gonna make about SSD speed? Get over it Xbox has a decently more powerful GPU and that matters more { and is not as if it has a slow ssd}
Having said that who wins next gen will come down to pricing and to a degree first party games (which is where sony has an advantage)
 
ITT we are watching a bunch of people who may not even know what an affine transform is for triangulation lecturing a graphics programmer.
I'm just going by logic here. At sub4k resolutions cards with similar Tflops to the next gen consoles are running games at 100-120+fps(ignoring rdna2 might be even more efficient than current gen cards with similar tflops). There's the horsepower to run 3-4~x the amount of current gen detail easy at 30fps. But where is that detail going to come from? If it is not procedural, all that additional detail will take more ram. Yet you're limited in what you can have in ram due to avoiding inordinate loading times with hdds.

In order to fill ram for the detail next gen consoles can easily handle you need the ssds.
 

bitbydeath

Member
Really PS fanboys ? How many threads are you gonna make about SSD speed? Get over it Xbox has a decently more powerful GPU and that matters more { and is not as if it has a slow ssd}
Having said that who wins next gen will come down to pricing and to a degree first party games (which is where sony has an advantage)

Not an SSD thread.
 
It might be common, but the best looking games are unlikely to be native 4k.
I disagree. I think native 4K will be the standard next gen just like 1080p is the standard this gen or the bare minimum. We'll see. I can see the PS5 maybe having more games slightly below native 4K with the lower gpu.
 
Last edited:
I disagree. I think native 4K will be the standard next gen just like 1080p is the standard this gen or the bare minimum. We'll see. I can see the PS5 maybe having more games slightly below native 4K with the lower gpu.
It might even become the standard, but it still doesn't deny that there will be significant performance advantage for titles using 4k reconstruction over 4k native titles. In the rx 5700 xt going from 4k to 1440p in some titles essentially doubles available performance.
 
Last edited:
Top Bottom