Support NeoGAF

[funking giblet]

Kind of funny how quickly the thread died, after someone claimed that no one is looking into transferring data directly from storage to GPU, and we have a link of nVidia explaining that exact thing....

And even more of a coincidence, nVidia talks about GPUDirect Storage for this feature, and somehow, the API that is considered to be a gamechanger for the XSX (and for the future in Windows) is called DirectStorage. That is quite the coincidence.

Yes I'm aware the consoles use AMD and not nVidia. But most likely this will become a spec for DX12U, and RDNA2 is DX12U compliant, so they will have it too.

GPU VRAM you mean. Which is what consoles already do as they use unified RAM.
GPUDirect allows the GPU to skip loading into normal system RAM first, which is a PC concern, for things machine learning, which use massive datasets to train on using matrix math and does so over networks too.

The game changer here isn't that you are loading into VRAM, it's that you are going through specialized hardware to do so, with low latency (The above is great for parallel processing in giant datacenters, not low latency environemnetrs), that benefits GAMES.

 

[Ascend]

The game changer here isn't that you are loading into VRAM, it's that you are going through specialized hardware to do so, with low latency (The above is great for parallel processing in giant datacenters, not low latency environemnetrs), that benefits GAMES.

And what would be the console equivalent of that?

 

[funking giblet]

And what would be the console equivalent of that?

The game changer was referring to console tech, sorry i wasn’t clear.

GPUDirect Storage isn’t useful for game consoles which can use Load and Store to hit VRAM, although the new consoles have made this more efficient.

GPUDirect isn’t really used for games.

 

[Degree]

Kind of funny how quickly the thread died, after someone claimed that no one is looking into transferring data directly from storage to GPU, and we have a link of nVidia explaining that exact thing....

And even more of a coincidence, nVidia talks about GPUDirect Storage for this feature, and somehow, the API that is considered to be a gamechanger for the XSX (and for the future in Windows) is called DirectStorage. That is quite the coincidence.

Yes I'm aware the consoles use AMD and not nVidia. But most likely this will become a spec for DX12U, and RDNA2 is DX12U compliant, so they will have it too.

well, it doesn’t fit the narrative that PS5 is YEARS ahead of PC/XSX lol

oh boy, I can’t wait for the games! Can’t wait for the DF threads, so that stupid narrative can die.

 

[thicc_girls_are_teh_best]

The game changer was referring to console tech, sorry i wasn’t clear.

GPUDirect Storage isn’t useful for game consoles which can use Load and Store to hit VRAM, although the new consoles have made this more efficient.

GPUDirect isn’t really used for games.

Nonetheless it still serves its use for data centers and seeing that MS want to use XSX hardware for Azure servers, an equivalent feature being present on the system to aid in that type of environment should be expected.

Still though I'm curious to know if there are any ways it could be leveraged for games that maybe we aren't able to consider at this time. Hopefully more official data is divulged on the platform in the next month or two.

 

[Ascend]

The game changer was referring to console tech, sorry i wasn’t clear.

GPUDirect Storage isn’t useful for game consoles which can use Load and Store to hit VRAM, although the new consoles have made this more efficient.

GPUDirect isn’t really used for games.

Oh. Well. From that perspective it sounds like you're saying the I/O on the consoles, mainly the decompression block, is the main (if not only) 'gamechanger' here. Is that correct or incorrect?

 

[funking giblet]

Oh. Well. From that perspective it sounds like you're saying the I/O on the consoles, mainly the decompression block, is the main (if not only) 'gamechanger' here. Is that correct or incorrect?

Yes, that’s part of it. I guess people are waiting to see what else XVA offers as PS5 has already detailed other improvements around DMA, memory mapping and coherency of shared data and invalidation.
There can be software solutions for these too.

 

[Tripolygon]

GPU VRAM you mean. Which is what consoles already do as they use unified RAM.
GPUDirect allows the GPU to skip loading into normal system RAM first, which is a PC concern, for things machine learning, which use massive datasets to train on using matrix math and does so over networks too.

The game changer here isn't that you are loading into VRAM, it's that you are going through specialized hardware to do so, with low latency (The above is great for parallel processing in giant datacenters, not low latency environemnetrs), that benefits GAMES.

Yup. Normally when GPU needs data, it gets copied to system ram into a bounce buffer before it is copied to the final destination which is the VRAM. GPUDirect makes it possible to bypass that step and DMA directly to VRAM skipping the System Ram. This does not apply to modern consoles because they use unified memory where CPU and GPU share from the same pool of memory.

 

[Tripolygon]

Kind of funny how quickly the thread died, after someone claimed that no one is looking into transferring data directly from storage to GPU, and we have a link of nVidia explaining that exact thing....

And even more of a coincidence, nVidia talks about GPUDirect Storage for this feature, and somehow, the API that is considered to be a gamechanger for the XSX (and for the future in Windows) is called DirectStorage. That is quite the coincidence.

Yes I'm aware the consoles use AMD and not nVidia. But most likely this will become a spec for DX12U, and RDNA2 is DX12U compliant, so they will have it too.

The thread died because the person keeping it alive requested a self ban. What Nvidia talks about has nothing to do with this. Its a case of someone reading something and misunderstanding what it is in order to support a preconceived idea. RAM is always involved in each step they are not going to be loading data directly from SSD to GPU cache. GPU Direct is bypassing system RAM and going to VRAM. And also when Someone tries to explain you and various others choose not to understand so there is really no point in explaining further is there?

 

[Major_Key]

Yup. Normally when GPU needs data, it gets copied to system ram into a bounce buffer before it is copied to the final destination which is the VRAM. GPUDirect makes it possible to bypass that step and DMA directly to VRAM skipping the System Ram. This does not apply to modern consoles because they use unified memory where CPU and GPU share from the same pool of memory.

So SX could have better latency of transfer ?

 

[Ascend]

The thread died because the person keeping it alive requested a self ban. What Nvidia talks about has nothing to do with this. Its a case of someone reading something and misunderstanding what it is in order to support a preconceived idea. RAM is always involved in each step they are not going to be loading data directly from SSD to GPU cache. GPU Direct is bypassing system RAM and going to VRAM. And also when Someone tries to explain you and various others choose not to understand so there is really no point in explaining further is there?

Then what is the function of DirectStorage on the XSX, considering there is no differentiation between system RAM and GPU RAM, since they are the same? If you're still loading from SSD to RAM why would it need a specific name? You're still doing the same thing.

 

[Tripolygon]

So SX could have better latency of transfer ?

I don't understand the question. Lower transfer latency compared to what?

 

[funking giblet]

Then what is the function of DirectStorage on the XSX, considering there is no differentiation between system RAM and GPU RAM, since they are the same? If you're still loading from SSD to RAM why would it need a specific name? You're still doing the same thing.

Xbox Series X: What do game developers say about next-gen? We asked several for their thoughts.

The Xbox Series X is upon us. But what do game developers think of this next-gen beast? We spoke to several to find out their thoughts.

www.windowscentral.com

This became the foundation for the Xbox Velocity Architecture, which comprises our custom-designed NVME SSD, a custom dedicated hardware decompression block, our new DirectStorage API which provides developers with direct low-level access to the NVME controller, and our innovative Sampler Feedback Streaming solution which acts as a multiplier of the physical memory in the console.

Looks like low level controller access, by passing some of the slower abstractions, and might let you play with IO queues directly, and maybe priorities. This would be software driven though.


Microsoft already work with NVMeDirect, might be something to look into there.

 

[Major_Key]

Xbox Series X: What do game developers say about next-gen? We asked several for their thoughts.

The Xbox Series X is upon us. But what do game developers think of this next-gen beast? We spoke to several to find out their thoughts.

www.windowscentral.com

Looks like low level controller access, by passing some of the slower abstractions, and might let you play with IO queues directly, and maybe priorities. This would be software driven though.

Yeah, it will be nice to see FP studios, DirectStorage,Hardware Decompression with BCPack & SFS how they will use it.

The Xbox Velocity Architecture was designed to be the ultimate solution for game asset streaming, delivering a level of performance well beyond the raw specifications of the hardware, allowing developers to virtually eliminate loading times and enabling them to deliver even larger and more immersive, dynamic, living worlds for gamers to explore and enjoy."

With the Xbox Series X, out of the gate, we reduced our load-times by more than 4x without any code changes. With the new DirectStorage APIs and new hardware decompression, we can further improve I/O performance and reduce CPU overhead, both of which are essential to achieve fast loading. As we look to the future, the Xbox Series X's Sampler Feedback for Streaming (SFS) is a game-changer for how we think about world streaming and visual level of detail. We will be exploring how we can use it in future titles to both increase the texture detail in our game beyond what we can fit into memory, as well as reduce load times further by increasing on-demand loading to just before we need it, instead of pre-loading everything up-front as we would use a more traditional 'level loading' approach.

The technical director of Remedy said that next gen games can have the same loading time as we currently have because games will get bigger.

With SFS, DirectStorage (MS let 100GB accessible with decompression hardware), 560Gb/s bandwitch, for future ambitious games it is better to increase the load on demand so that the player knows very little loading time than try to display all the data on the screen and have loading time.

 

[Tripolygon]

Then what is the function of DirectStorage on the XSX, considering there is no differentiation between system RAM and GPU RAM, since they are the same?

This is all we know.

DirectStorage – DirectStorage is an all new I/O system designed specifically for gaming to unleash the full performance of the SSD and hardware decompression. It is one of the components that comprise the Xbox Velocity Architecture. Modern games perform asset streaming in the background to continuously load the next parts of the world while you play, and DirectStorage can reduce the CPU overhead for these I/O operations from multiple cores to taking just a small fraction of a single core; thereby freeing considerable CPU power for the game to spend on areas like better physics or more NPCs in a scene. This newest member of the DirectX family is being introduced with Xbox Series X and we plan to bring it to Windows as well.

If you're still loading from SSD to RAM why would it need a specific name? You're still doing the same thing.

Because, it is a new API that Microsoft says reduces CPU overhead and they want to differentiate it? I mean they named their IO Velocity Architecture. They named the DMA in Xbox One data move engines. It is a new storage api and follows previous Microsoft naming conventions when it comes to DirectX APIs akin to DirectML, Direct3D, Direct2D, DirectXAudio etc, and now DirectStorage.

 

[martino]

Xbox Series X: What do game developers say about next-gen? We asked several for their thoughts.

The Xbox Series X is upon us. But what do game developers think of this next-gen beast? We spoke to several to find out their thoughts.

www.windowscentral.com

Microsoft already work with NVMeDirect, might be something to look into there.

interesting read
crunch codec and NVMeDirect like solutions could be the 2 missing part the equation.
but where or how is done decompression and streaming of super-compressed textures in case of pc (what does it means if cpu is used ? memory copy seems unavoidable. does it mean gpu will have to do this work ? it seems taxing lot of resources in both case )

 

[Tripolygon]

Xbox Series X: What do game developers say about next-gen? We asked several for their thoughts.

The Xbox Series X is upon us. But what do game developers think of this next-gen beast? We spoke to several to find out their thoughts.

www.windowscentral.com

Looks like low level controller access, by passing some of the slower abstractions, and might let you play with IO queues directly, and maybe priorities. This would be software driven though.


Microsoft already work with NVMeDirect, might be something to look into there.

Thanks for finding this. This is exactly what I was talking about.

With the Xbox Series X, out of the gate, we reduced our load-times by more than 4x without any code changes. With the new DirectStorage APIs and new hardware decompression, we can further improve I/O performance and reduce CPU overhead, both of which are essential to achieve fast loading. As we look to the future, the Xbox Series X's Sampler Feedback for Streaming (SFS) is a game-changer for how we think about world streaming and visual level of detail. We will be exploring how we can use it in future titles to both increase the texture detail in our game beyond what we can fit into memory, as well as reduce load times further by increasing on-demand loading to just before we need it, instead of pre-loading everything up-front as we would use a more traditional 'level loading' approach.

We are going from caching data in RAM to pulling data from SSD to RAM just in time.

With SFS, DirectStorage (MS let 100GB accessible with decompression hardware), 560Gb/s bandwitch, for future ambitious games it is better to increase the load on demand so that the player knows very little loading time than try to display all the data on the screen and have loading time.

Thats not what that means. It just means loading data just before you need it rather than caching data that is not needed for immediate use. Games already cull what is not in view and being displayed on the screen although the data is still in RAM. With just in time loading, they can now afford to load what is needed to be displayed in the next 1 to 2 seconds rather than keeping data for what is needed for the next 10 to 30 seconds.

 

[John254]

well, it doesn’t fit the narrative that PS5 is YEARS ahead of PC/XSX lol

oh boy, I can’t wait for the games! Can’t wait for the DF threads, so that stupid narrative can die.

Oh, don't worry. They'll just move the goalposts to something like "devs don't know how to utilize ps5's strengths"

 

[Dnice1]

Oh, don't worry. They'll just move the goalposts to something like "devs don't know how to utilize ps5's strengths"

Bookmarked!

 

[thicc_girls_are_teh_best]

Tripolygon So it's basically a different methodology for addressing a current issue regarding RAM wastage of unneeded assets, essentially attacking the same problem Sony is but through a different vector and implemetation.

One approach is relying on modest I/O hardware and smart/proprietary software algorithms to attack the problem, the other is relying on beefier I/O hardware and great (yet non-proprietary) software algorithms to attack the same problem. Arguably the key difference being the former still uses a fraction of a CPU core for some of this and the other offloads 100% of the important work to a dedicated processor in the I/O block.

I'll admit these recent discussions do have me reconsidering speculation on the XSX's GPU using the data streamed from storage to load into the caches directly to work with from there; if I were maybe 80% feeling that before I'd say it's a lot lower than that now. Could it still be doing something like that? Maybe. But it's a low probability seems like.

But at least one thing this has all helped highlight is that, indeed, both companies have considered these I/O bottlenecks and have gone about addressing them in ways that play to their own strengths and market necessities. And I still expect Sony's to have the edge in that regard, but I expect MS's solution to punch above its weight as it seems a very specifically tailored approach to the problem with just enough hardware behind it to let it do it's thing.

The sooner we can see these I/O systems (and really the entire hardware) in action with real next-gen games running on real hardware, the better.

 

[Ascend]

Tripolygon So it's basically a different methodology for addressing a current issue regarding RAM wastage of unneeded assets, essentially attacking the same problem Sony is but through a different vector and implemetation.

One approach is relying on modest I/O hardware and smart/proprietary software algorithms to attack the problem, the other is relying on beefier I/O hardware and great (yet non-proprietary) software algorithms to attack the same problem. Arguably the key difference being the former still uses a fraction of a CPU core for some of this and the other offloads 100% of the important work to a dedicated processor in the I/O block.

I'll admit these recent discussions do have me reconsidering speculation on the XSX's GPU using the data streamed from storage to load into the caches directly to work with from there; if I were maybe 80% feeling that before I'd say it's a lot lower than that now. Could it still be doing something like that? Maybe. But it's a low probability seems like.

But at least one thing this has all helped highlight is that, indeed, both companies have considered these I/O bottlenecks and have gone about addressing them in ways that play to their own strengths and market necessities. And I still expect Sony's to have the edge in that regard, but I expect MS's solution to punch above its weight as it seems a very specifically tailored approach to the problem with just enough hardware behind it to let it do it's thing.

The sooner we can see these I/O systems (and really the entire hardware) in action with real next-gen games running on real hardware, the better.

Yeah I agree. I've been convinced that we're not likely going to see direct SSD to GPU data transfer without RAM. Slowly but surely the idea of transferring directly to the GPU from SSD seems to be looking less plausible. But all this brings us directly to square one, which was trying to clarify what the instant 100GB access means. I'm not convinced it's virtual memory in the traditional sense. But we'll see.

I still like the idea of having the RAM and the 100GB of SSD work as extended levels of cache. Maybe I should try and design my own console lol xD

 

[thicc_girls_are_teh_best]

Yeah I agree. I've been convinced that we're not likely going to see direct SSD to GPU data transfer without RAM. Slowly but surely the idea of transferring directly to the GPU from SSD seems to be looking less plausible. But all this brings us directly to square one, which was trying to clarify what the instant 100GB access means. I'm not convinced it's virtual memory in the traditional sense. But we'll see.

I still like the idea of having the RAM and the 100GB of SSD work as extended levels of cache. Maybe I should try and design my own console lol xD

x3 I know what you mean dude; part of the reason I started learning about this stuff is because I like trying to come up with my own hardware ideas and wanted to base them in actual technical & market realities. But there's a reason they have massive teams for that type of stuff!

Ironically this has also helped clarify for me how AMD's SSG cards work; if they're using an equivalent of GPUDirect so the GPU fetches data from the NAND to store in the HBM2 memory, that explains a lot. I still think conceptually, for very specific types of workloads, the idea of having the GPU stream in data from storage NAND to the caches wouldn't be impossible, as long as it was for large sets of smaller-scale data only needing a few dozen MBs per frame, and maybe not requiring so many read accesses to NAND.

But theoretically it would almost always be better to just have that stuff streamed to RAM "just in time" since you have byte-level addressability and true random access in that case, plus much lower latencies and MUCH faster access speed by the GPU for both read and write operations. I just don't think NAND will ever have the speeds, bandwidth or latency to be overall useful as a solution for pseudo RAM-like access...

...thankfully that's where technologies like 3D Xpoint, ReRAM, MRAM, FRAM etc. can come into the picture. But particularly 3D Xpoint if we're talking about something mature enough and at mass quantities that can go into a game console, if the price-per-GB can come down further and capacities/speeds increase. Is it too much to ask for potential 4+ GB/s 3D Xpoint chips at decent capacities 3-4 years from now? I don't think so, fingers crossed

*EDIT: I have an idea that the "instantly accessible" could be in reference to lack of needing to convert the files in that space o something to that effect. oldergamer mentioned the idea to me when bringing up DXTC textures if Sony were trying to do something similar to SFS. Apparently the DXTC textures would be standard compressed textures. They base this off the fact that Sony haven't mentioned anything similar to SFS's PRTs, so if it wanted to do something like that there could be an additional step involved on their end.

It's an interesting idea but I'm thinking the "instantly accessible" could be in reference to that.

 

[Dnice1]

I think a big part of what people are missing is, once you have the data in RAM, the next part is important, and it's that part where the XSX will shine. (Rendering).

It seem folks have really forgot about the processing part after the 13TF dream died.

 

[Greatest Of All Time]

Regardless of the technical jargon it seems clear to me third party devs are going to cater to XSX SSD speeds which negates PS5s only advantage. And that means the raw power advantage of the XSX will then make those 3rd party games look better overall. So 3rd party wise the winner is XSX.

There's also the fact that Sony sucks with cooling. The PS4 was a mess with that jet engine sound from the fans. Seriously, fuck that noise. Literally. I dont trust them to make a quiet console. They don't seem to trust themselves to make a quiet, well cooled console either because of this variable speed and power and only max strength at low loads bullshit. I dont trust PS5 yet.

 

[Shambala]

This thread is sad as fuck lmao. Y’all need help

 

[Dodkrake]

Regardless of the technical jargon it seems clear to me third party devs are going to cater to XSX SSD speeds which negates PS5s only advantage. And that means the raw power advantage of the XSX will then make those 3rd party games look better overall. So 3rd party wise the winner is XSX.

There's also the fact that Sony sucks with cooling. The PS4 was a mess with that jet engine sound from the fans. Seriously, fuck that noise. Literally. I dont trust them to make a quiet console. They don't seem to trust themselves to make a quiet, well cooled console either because of this variable speed and power and only max strength at low loads bullshit. I dont trust PS5 yet.

Your comment is so ignorant I won't even bother addressing most of it, but wanted to add something:

Sony can't make any decent cooling, but Xbox produced a console that basically killed itself because of heat. 50+% fail rates by 2009.

ign.com/articles/2009/08/17/report-xbox-360-failure-rate-reaches-54

 

[Allandor]

It seem folks have really forgot about the processing part after the 13TF dream died.

But at all, the visual difference should be really, really small.
Even on current gen consoles, we need to zoom in to see differences (most of the time).
It gets more difficult to see the differences, the higher the base-resolution and the assets quality is.
The Unreal Demo just showed that. Polygons the size of a Pixel is just a waste of power. Yes, you can do it with the nvme SSDs but I would bet, that if you increase the polygon size (just double it) you won't see a huge difference at all, because the details are already smaller than individual pixels.
The only thing I see is RT where a difference my be possible. But just only because you need a "base" resolution for RT up to a point where it no longer makes big jumps in image quality. We will see if both consoles have that minimal RT power to deliver that.

 

[TBiddy]

Your comment is so ignorant I won't even bother addressing most of it, but wanted to add something:

Sony can't make any decent cooling, but Xbox produced a console that basically killed itself because of heat. 50+% fail rates by 2009.

ign.com/articles/2009/08/17/report-xbox-360-failure-rate-reaches-54

I love how you counter the "PS4 fan noise is loud" with the "B-B-BUT MICROSOFT IN 2009!".

 

[Deleted member 775630]

I love how you counter the "PS4 fan noise is loud" with the "B-B-BUT MICROSOFT IN 2009!".

I liked how he refused to address the 3rd party issue and just puts it on the "ignorant" shelf. Because it actually makes a lot of sense that every 3rd party game will look and play better on the XSX. They also have to take into account PC with slower SSD's and even HDD. So the SSD can only make a difference in loading time.

 

[TBiddy]

I liked how he refused to address the 3rd party issue and just puts it on the "ignorant" shelf. Because it actually makes a lot of sense that every 3rd party game will look and play better on the XSX. They also have to take into account PC with slower SSD's and even HDD. So the SSD can only make a difference in loading time.

That would make sense yes, unless they increase the LOD slightly on the PS5, just like they will increase the resolution or the RT slightly on the XSX.

 

[MadViking]

I think people are focusing too much on the 100 GB. My understanding is that they meant it just an example, a typical size for next gen games asset size, at least the critical ones that need to be accessed by gpu. So there would be no any artificial limit or dedicated space on the drive.

 

[Ascend]

I think people are focusing too much on the 100 GB. My understanding is that they meant it just an example, a typical size for next gen games asset size, at least the critical ones that need to be accessed by gpu. So there would be no any artificial limit or dedicated space on the drive.

I don't know. It doesn't sound like an example;

Enter Xbox Velocity Architecture, which features tight integration between hardware and software and is a revolutionary new architecture optimized for streaming of in game assets. This will unlock new capabilities that have never been seen before in console development, allowing 100 GB of game assets to be instantly accessible by the developer. The components of the Xbox Velocity Architecture all combine to create an effective multiplier on physical memory that is, quite literally, a game changer.

Xbox Series X: A Closer Look at the Technology Powering the Next Generation - Xbox Wire

A few months ago, we revealed Xbox Series X, our fastest, most powerful console ever, designed for a console generation that has you, the player, at its center. When it is released this holiday season, Xbox Series X will set a new bar for performance, speed and compatibility, all while allowing...

news.xbox.com

 

[Dodkrake]

I liked how he refused to address the 3rd party issue and just puts it on the "ignorant" shelf. Because it actually makes a lot of sense that every 3rd party game will look and play better on the XSX. They also have to take into account PC with slower SSD's and even HDD. So the SSD can only make a difference in loading time.

Look better? Debatable. What is your playback medium? Can you see the difference between native 4K or resolution over 15% lower resolution upscaled? Hell, even Digital Foundry's dictator could not count pixels on a 1440p Unreal 5 Demo.

Regarding FX, we're not talking 7th gen here, we're talking 2 consoles that have actually nicely optimized IO's, similar RAM configurations and CU fill rates, CPU, etc. It will not make any sense to downgrade FX quality when you can lower the amount of pixels in display.

Even Ray Tracing scales with Clock Speed, so the delta should still be roughly the same, so as long as you are pushing -15 to 20% pixels, everything else should feel and play the same.

So they may look better, depending on the medium and how eagle eyed you are, but the differences will likely be so minute that I'm actually expecting Digital Foundry to zoom in 500 times to find 1 pixel out of place.

As for play better, depends on your preferred controller. I'm not expecting an FPS variation between them, or at least a meaningful one. Case in point? Those pesky games that have better resolution in the Xbox One X but better FPS on the PS4 Pro. I think Call of Duty has one of those?

 

[Deleted member 775630]

Look better? Debatable. What is your playback medium? Can you see the difference between native 4K or resolution over 15% lower resolution upscaled? Hell, even Digital Foundry's dictator could not count pixels on a 1440p Unreal 5 Demo.

Regarding FX, we're not talking 7th gen here, we're talking 2 consoles that have actually nicely optimized IO's, similar RAM configurations and CU fill rates, CPU, etc. It will not make any sense to downgrade FX quality when you can lower the amount of pixels in display.

Even Ray Tracing scales with Clock Speed, so the delta should still be roughly the same, so as long as you are pushing -15 to 20% pixels, everything else should feel and play the same.

So they may look better, depending on the medium and how eagle eyed you are, but the differences will likely be so minute that I'm actually expecting Digital Foundry to zoom in 500 times to find 1 pixel out of place.

As for play better, depends on your preferred controller. I'm not expecting an FPS variation between them, or at least a meaningful one. Case in point? Those pesky games that have better resolution in the Xbox One X but better FPS on the PS4 Pro. I think Call of Duty has one of those?

I see you already came to the conclusion yourself that it will look better, and when I said play better I meant purely what you see on screen. You are correct that CoD has a more stable framerate. This is 1 game. That doesn't make not meaningful if all the other games have a more steady and stable framerate.

 

[Naddy]

I liked how he refused to address the 3rd party issue and just puts it on the "ignorant" shelf. Because it actually makes a lot of sense that every 3rd party game will look and play better on the XSX. They also have to take into account PC with slower SSD's and even HDD. So the SSD can only make a difference in loading time.

And to add to that: It's much easier to scale with CPU and GPU.

• Have a weaker CPU? Just reduce FPS, e.g. have 30fps for the platform instead of 60 FPS.
• Have a weaker GPU? Just reduce the resolution from native 4K to 1440p for example.

This is very easy for the developer and not much work is needed and thus it would not cost much resources for the developer.

On the other hand, scaling with SSD is not really feasible, you won't create all kinds of different assets just for:

1. Xbox One/PS4
2. PC HDD
3. PC SSD
4. XSX
5. PS5

This just isn't feasible.

I think, what devs will do is just:

• HDD assets (for Xbox one, ps4 and PC until 2021/2022)
• SSD assets based on PC SSD with slower IO than XSX/PS5

that way all those assets will work on basically ALL nextgen platforms easily without having to do a lot of work with the scaling. AAA development already is extremely expensive and that way, they could easily make use of all SSDs without using a lot of resources.

 

[JareBear: Remastered]

here is an article I just saw pop up regarding devs addressing the difference in storage bandwidth between the two consoles

Developers making Xbox Series X games will learn to address PS5's SSD advantage, says ex-Xbox lead

Developers working on multiplatform games will find a way to get around Xbox Series X's relatively slower SSD compared to PS5's.

www.vg247.com

Developers making Xbox Series X games will learn to address PS5’s SSD advantage, says ex-Xbox lead

Developers working on multiplatform games will find a way to get around Xbox Series X’s relatively slower SSD compared to PS5’s.

Microsoft’s William Stillwell, who currently manages the company’s Mixed Reality program, spent years at Xbox working on backwards compatibility, xCloud, and Xbox platform services.

Stillwell was recently a guest on the Iron Lords Podcast, where he was asked about the PS5’s impressive SSD tech, how faster it is compared to Xbox Series X’s solution, and whether it’ll even matter.

“I’ll say two things with that. One, I understand that it’s a marketing point and a high point – and I think I said in the last podcast we did, I’m really impressed with what they’ve done in the architecture with their drive and I think they’ll be able to do really cool stuff with it,” said Stillwell.

“I also don’t think it’s as impressive over what you’ll be able to do [with Xbox Series X] and so I have high confidence in that team. I will also say – look I reserve the right to wake up smarter tomorrow and someone may prove me wrong and we’ll learn. The one thing I have learned is [to] never underestimate game developers’ [ability] to take advantage of something and work around it, but that works on both sides.

“So the ability of a game developer to look and say, ‘Ok, I am building this game and I wanna have seamless transition with no loading screens on [PS5], but I am also gonna wanna sell on Xbox, what am I gonna do to manage that kinda thing?’

“Like, that will be the tricks and the techniques, so – I am just making this up – maybe there are elevators in the Xbox Series X version and not on the [PS5] version. I don’t know yet how they’re gonna address it, but they’ll learn to address it and they will figure out ways to work around it.”

 

[Ascend]

I suspect that the hardware filters in the XSX are sharpening filters. When you load the lower quality assets, the texture is sharpened, awaiting the 'true' texture to be loaded a couple of frames later if required, smoothing the transition, avoiding pop-in. It could possibly negate the slower loading visually.

 

[Greatest Of All Time]

Your comment is so ignorant I won't even bother addressing most of it, but wanted to add something:

Sony can't make any decent cooling, but Xbox produced a console that basically killed itself because of heat. 50+% fail rates by 2009.

ign.com/articles/2009/08/17/report-xbox-360-failure-rate-reaches-54

Could you repeat that? I couldn't hear you over the sound of the PS4 fans whirring.

 

[Allandor]

On the other hand, scaling with SSD is not really feasible, you won't create all kinds of different assets just for:

1. Xbox One/PS4
2. PC HDD
3. PC SSD
4. XSX
5. PS5

This just isn't feasible.

Can you write why?
It is quite easy to scale assets. Alls assets have an (more or less) resolution.
- Polygon coordinates
- Texture resolution
- sound quality
...

All those can be reduced just like it is done all the time in the past. Developers/Designers make high-quality stuff in a separate program and than export those things in the needed quality the system can handle. All this can be part of a build process that reduces the resolution of textures, ... while building the project.
There is not much needed to do it. Yes, this is not as good as doing it manual but good enough.
e.g. Reducing the Texture resolution from 8k to 4k saves much, much memory, but in most cases doesn't make any visual difference. And that is also where texturecompression comes into play. The quality get reduced (just a little bit) but also the memory footprint. With todays high quality textures you really don't see a difference.

That's also why I write it and write it again. We won't see much difference between both consoles. The base-quality of both is so high, so there will only be minimal differences.
The only thing that can make a difference is RT (but well, we will see) and the CPU. But we don't know how far the CPU goes down with it's clock speed on PS5 to increase GPU speed. But well, in the past we've seen, that even with a bit weaker CPU it always seemed that Sony had the less CPU dependent API. Even though the xbox one had the stronger CPU it almost always struggled to get better framerates, even on lower resolution.

 

[Dodkrake]

Could you repeat that? I couldn't hear you over the sound of the PS4 fans whirring.

Hmmm. I prefer my fans to make some noise over my console melting itself due to its poor engineering. But that's me, you do you.

 

[Redlight]

Hmmm. I prefer my fans to make some noise over my console melting itself due to its poor engineering. But that's me, you do you.

The RROD article you linked to was from eleven years ago.

Might be time to let it go.

 

[Kenpachii]

Can you write why?
It is quite easy to scale assets. Alls assets have an (more or less) resolution.
- Polygon coordinates
- Texture resolution
- sound quality
...

All those can be reduced just like it is done all the time in the past. Developers/Designers make high-quality stuff in a separate program and than export those things in the needed quality the system can handle. All this can be part of a build process that reduces the resolution of textures, ... while building the project.
There is not much needed to do it. Yes, this is not as good as doing it manual but good enough.
e.g. Reducing the Texture resolution from 8k to 4k saves much, much memory, but in most cases doesn't make any visual difference. And that is also where texturecompression comes into play. The quality get reduced (just a little bit) but also the memory footprint. With todays high quality textures you really don't see a difference.

That's also why I write it and write it again. We won't see much difference between both consoles. The base-quality of both is so high, so there will only be minimal differences.
The only thing that can make a difference is RT (but well, we will see) and the CPU. But we don't know how far the CPU goes down with it's clock speed on PS5 to increase GPU speed. But well, in the past we've seen, that even with a bit weaker CPU it always seemed that Sony had the less CPU dependent API. Even though the xbox one had the stronger CPU it almost always struggled to get better framerates, even on lower resolution.

But why create 8k textures to start with? why drop it into the PS5 which will eat disc space like no other and also memory? when u can just drop 4k version and be done with it. The dude is right, what devs will do is look at what people got and build there stuff around that. I think xbox is a better designed box all around over the PS5 because of this. It focuses on the stuff that matters.

 

[D.Final]

Well, good for the development

 

[MoreJRPG]

Hmmm. I prefer my fans to make some noise over my console melting itself due to its poor engineering. But that's me, you do you.

I’d check your bank statements to see if your credit infos been compromised. I heard there’s a hacking issue!

See, isn’t this fun? Bringing up stuff from over a decade ago that isn’t even relevant anymore. You look so petty trying to defend a piece of plastic. Let it go, these billion dollar corporations don’t care about you.

 

[Dodkrake]

I’d check your bank statements to see if your credit infos been compromised. I heard there’s a hacking issue!

See, isn’t this fun? Bringing up stuff from over a decade ago that isn’t even relevant anymore. You look so petty trying to defend a piece of plastic. Let it go, these billion dollar corporations don’t care about you.

Ahm... Hasn't this thread been 50 pages of people trying to violate the laws of physics to justify that MS has a slower IO overall in their machine? Because It sure looks like it.

And I don't give a rats ass about Sony, I like their Playstation product, none of the others.

 

[Dontero]

Can you write why?
It is quite easy to scale assets. Alls assets have an (more or less) resolution.
- Polygon coordinates
- Texture resolution
- sound quality

Because it is not just as easy as to devide everything by factor you want and process will be done automatically. Texture which look great at high res might look like complete garbage at low resolution, geometric detail which looks amazing at high polycount could look abysmal at low. The sound you produced with 7.1 in mind can be completely killed when you compress it to 128kb/s mono and so on.
Moreover none of those things are automatic. You have to effectively redo most of the work.

Trust me, if those things were easy then you would see any game on any platform. Do you think company like Activision doesn't want to release their call of duty on phones ?

 

[Allandor]

Because it is not just as easy as to devide everything by factor you want and process will be done automatically. Texture which look great at high res might look like complete garbage at low resolution, geometric detail which looks amazing at high polycount could look abysmal at low. The sound you produced with 7.1 in mind can be completely killed when you compress it to 128kb/s mono and so on.
Moreover none of those things are automatic. You have to effectively redo most of the work.

Trust me, if those things were easy then you would see any game on any platform. Do you think company like Activision doesn't want to release their call of duty on phones ?

That does not make any sense at all. We have 2 consoles which are on more or less the same power level. So porting it from one to another is quite easy. And yes you can scale things quite easy. And as I wrote, you can scale all those things quite easy without loosing much information but used space. A phone is a completely different platform (other CPU, GPU, IO, ...) where even the game logic can be enough for overwhelm the phones CPU. PS5 and xbox s x won't be that different.

 

[Dontero]

That does not make any sense at all. We have 2 consoles which are on more or less the same power level. So porting it from one to another is quite easy. And yes you can scale things quite easy. And as I wrote, you can scale all those things quite easy without loosing much information but used space. A phone is a completely different platform (other CPU, GPU, IO, ...) where even the game logic can be enough for overwhelm the phones CPU. PS5 and xbox s x won't be that different.

I don't know what you even talk about. There is no asset porting involved when you talk about similar power level hardware. You just run game in lower frame-rate or lower resolution.

 

[ToadMan]

Activision doesn't want to release their call of duty on phones ?

 

[MoreJRPG]

Ahm... Hasn't this thread been 50 pages of people trying to violate the laws of physics to justify that MS has a slower IO overall in their machine? Because It sure looks like it.

And I don't give a rats ass about Sony, I like their Playstation product, none of the others.

The people who say the Xbox IO is faster have no idea what they’re talking about and should be ignored. It’s as ludicrous as those saying the PS5 SSD will make PC gaming obsolete. The fact that you’re getting riled up by trolls says more about you than anything.