• 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.

Xbox Series X’s BCPack Texture Compression Technique 'might be' better than the PS5’s Kraken

muteZX

Banned
what's killzone has to do with nanite? :messenger_grinning_smiling:


Nanite is a barren spell. The console transfers data from the SSD to RAM. It doesn't matter what you call the code on the side of the game engine that processes the data. Nanite, Granite, No-name. It doesn't matter. Efficient streaming aims to save bus throughput and RAM space. That is all.
 

Darius87

Member
Nanite is a barren spell. The console transfers data from the SSD to RAM. It doesn't matter what you call the code on the side of the game engine that processes the data. Nanite, Granite, No-name. It doesn't matter. Efficient streaming aims to save bus throughput and RAM space. That is all.
if it doesn't matter then why it says Nanite memory?
8wl1rua.png
 

Stooky

Member
man gaming communities are living of rumours and misinformation lately, what the fuck is going on
Not enough next gen games are out to compare. The games that are out doesn’t show any of the magic sauce Xbox fanboys have been wet dreaming about since the console launch. At least Sony fanboys have a few games that show what’s possible.
 
Last edited:

Darius87

Member
.. sure .. /as I have already spokenth !! .. .. "streaming pool is a streaming pool .. just a little bigger /+ textures, + audio/"./
no they not talking about streaming pool as a global streaming pool for every asset it's pool only for geometry.
textures takes up most of RAM out of 16gb geometry takes just 768mb speaking about UE5 demo hope that makes things clear.
 

muteZX

Banned
no they not talking about streaming pool as a global streaming pool for every asset it's pool only for geometry.
textures takes up most of RAM out of 16gb geometry takes just 768mb speaking about UE5 demo hope that makes things clear.

Definitive language barrier. I dont give a flying fuck what Nanite does with geometry in an ordinary demo. In the game UE5 will stream anything as needed / texture, geometry, audio /. It doesn't matter how much space the texture will take up from RAM. The aim is to work with as little data as possible, directly in the screen space. If the GTA6 has 500 gigabytes of data and 380 gigabytes of textures, it doesn't matter how much RAM you have.
 

Three

Member
Traditional mip streaming in today's AAA games market almost certainly means some form of virtualized texture paging system aka PRT. What would be the point in Microsoft building their console around a technique that has absolutely zero real benefit over what was already being used largely across Xbox consoles?

Microsoft built their system around technology and features designed to greatly outperform what was being used on xbox one generation titles. Enter Sampler Feedback Streaming. I highly doubt a Microsoft engineer working on it would be calling it a brand new feature or capability. I
traditional mip streaming means no PRT. Virtualised textures is not synonymous with PRT+ but is how PRT is done. PRT is very specific.Traditional mip streaming does not mean with PRT, he mentions this specifically in the demo where he says the entire floor is loaded. What is partially resident about that if the whole thing needs to be loaded?

Nobody is saying there is no benefit to SFS or PRT+. It's what you think this offers that's the issue. What is your baseline?

Are you talking about xbox one games? HDD vs SSD?
Possibility to save 2-3x the memory bandwidth/usage averaged MAXIMUM.

Are you talking about engines specifically written for gen 9 using SFS/PRT+ when compared to having no checks for distance, occlusion etc yet still loading the max quality texture. 2.5x+

Are you talking about PS5 vs Xbox SX? Very little to none because even if you assume that some hardware is missing (patents by M. Cerny, talks by AMD, UE5 demo and Sonys investment in a fast SSD suggest it isn't) you still have engines with software based PRT+.

Nobody is saying SFS is pointless or useless but you need to determine what you think it offers compared to what?
 

Darius87

Member
Definitive language barrier. I dont give a flying fuck what Nanite does with geometry in an ordinary demo. In the game UE5 will stream anything as needed / texture, geometry, audio /. It doesn't matter how much space the texture will take up from RAM. The aim is to work with as little data as possible, directly in the screen space. If the GTA6 has 500 gigabytes of data and 380 gigabytes of textures, it doesn't matter how much RAM you have.
my argument to you was that streaming in UE5 demo 768mb is just for geometry not how engine streams anything it needs of course it does i'm not arguing about that. you seem ignorant to fact that textures takes way more then 768mb to stream one asset in UE5 takes more then that so how is? like you said textures + audio takes just little bit over that? that ' doesn't makes sense.
remember that we're talking about streaming SSD to RAM it doesn't work per frame basis this is not RAM to GPU caching.
 

muteZX

Banned
my argument to you was that streaming in UE5 demo 768mb is just for geometry not how engine streams anything it needs of course it does i'm not arguing about that. you seem ignorant to fact that textures takes way more then 768mb to stream one asset in UE5 takes more then that so how is? like you said textures + audio takes just little bit over that? that ' doesn't makes sense.
remember that we're talking about streaming SSD to RAM it doesn't work per frame basis this is not RAM to GPU caching.

elsewhere ..

And back to streaming. That's why I wrote that the final streaming pool will of course be a bit bigger, but that's an "academic question". No RAM is large enough to hold all the game assets, which will be hundreds of gigabytes for AAA+ games.
 
Microsoft might have an ace up its sleeve to trump the PS5′ Kraken tech.

Of the many things Mark Cerny explained in detail during Sony’s recent deep dive into the PS5’s tech, one thing he mentioned was the console’s new texture decompression tech, called Kraken. Though it’s too specific to be anything but technical jargon to the layman, it should prove to be quite useful to developers, owing to how much more efficiently it operates than the current gen consoles.

Texture decompression isn’t something that Microsoft have talked about for the Xbox Series X too much yet, but according to industry professionals, their solution might be better even than Sony’s Kraken. Richard Geldreich, who formerly worked at Valve and Ensemble Studios, took to Twitter to say that Microoft’s texture decompression, BCPack, is their “dark horse” and might be a stronger option than Kraken.

He posed that question to James Stanard, who works on graphics optimization R&D and engine architecture on Microsoft. Though Stanard was reluctant to divulge many details about it, he went on to explain it briefly, it is specifically designed for texture decompression, before Geldreich added that being dedicated tech for this purpose will allow it to function more efficiently than Kraken, which is more of a general purpose system.

One thing that’s becoming clearer in recent days is that the PS5 and the Xbox Series both have significant advantages over the other in different areas– but both of them are incredible pieces of hardware that developers are quite excited to work on. Here’s hoping both of them reveal more details about their next-gen consoles in the near future.


Reading through this article again it seems like it's a bit outdated since Oodle Textures was added to the dev kits.

before Geldreich added that being dedicated tech for this purpose will allow it to function more efficiently than Kraken, which is more of a general purpose system.

Now that Oodle Textures was added to the kit and available to every developer I think it's worth studying how it felt nostra to Microsofts texture decompression tech.

 
Last edited:

Darius87

Member
elsewhere ..

And back to streaming. That's why I wrote that the final streaming pool will of course be a bit bigger, but that's an "academic question". No RAM is large enough to hold all the game assets, which will be hundreds of gigabytes for AAA+ games.
does that link have to prove something?
do you think assets just accumulate in size inside RAM until it fills up when streaming? you know what's flushing is? you can have 4GB of RAM if you have very fast SSD it will work just fine even if game is 500GB in size.
 

muteZX

Banned
does that link have to prove something?
do you think assets just accumulate in size inside RAM until it fills up when streaming? you know what's flushing is? you can have 4GB of RAM if you have very fast SSD it will work just fine even if game is 500GB in size.

Total language barrier. You answer me in my own words and you ask something that no one asks, because it is completely clear and repeatedly answered. I have no idea what you're asking now. If I had the SSD with a RAM speed, it wouldn't be called the SSD but it's another RAM and it doesnt matter how big the main RAM is, because I have another RAM stuck to it. Go back to the beginning, go over it again.
 
Last edited:

Darius87

Member
Total language barrier. You answer me in my own words and you ask something that no one asks, because it is completely clear and repeatedly answered. I have no idea what you're asking now.
does that link have to prove something?
.. sure .. /as I have already spokenth !! .. .. "streaming pool is a streaming pool .. just a little bigger /+ textures, + audio/"./
i'm telling you that's streaming pool can't be so small for every asset if that would be the case we won't need 16GB of ram because 15GB would be unused. we're in 2021 modern game assets take at least 8GB.
 

Fafalada

Fafracer forever
you can't have games long 10 hours with a size of 47gb and stream 22gb/s....cmon now ...lol the stream will be even lower 500 mb/s
Games are not video-playback (well, most games) so no - that's not how those numbers work.

Also speed of operations matters, not just peak number.
Ie - maybe I only need to read 16MB every frame (so at 60fps, 1GB/s throughput).
But if I/O can perform the operation in 0.8ms (@22GB/s) - that's a lot more useful than doing it in 17ms (@1GB/s), as the rest of the processing won't have to wait/work around the read-delays nearly as much.

It is as disingenuous (well unless taken and explained with a pinch of salt) as discussing about compression ratios in a way.
Yes and both are bad really.
It's a useful short-hand to simplify the explanation, but see above as to how it just confuses people to use-shorthands.
Compression multipliers are particularly dumb as there's no guarantees with lossless codecs - that 2x is just a statistical probability on the types of data we expect to encounter.
Efficiency multiplier is even worse as it's not actually comparing against any concrete baseline. It'd be far more relevant to talk about peak-memory use @ target resolution/sample count(decoupling fidelity from memory-usage), but of course that doesn't make as compelling sales-pitch to broader audiences.
 
Microsoft was not comparing to games w/ PRT on Xbox One.

You can tell that by looking at the slide right before the demo of SFS.. where they outright state PRT was rare last generation.

Which I believe it was.. the idea that most games were doing PRT last gen seems to be a bit of a straw man. The tech was developed and in the chips for last gen, and in PC GPUs, but the i/o limitations held it back.. iD was one of the only companies to do anything decent with it, as they had experience with that tech even on the PS3 gen, but the same I/O limitations existed between PS3/PS4, and Rage 2 still had pop-in issues fairly apparent.

Look at the slide that appears RIGHT before the SFS demo:

fg1Nf5M.png


The point I continue to make is Microsoft was comparing to all Xbox One generation games period, with PRT, without, doesn't matter. And they came away from all that extensive monitoring and analysis with this one conclusion.


Through specialized hardware added to the Xbox One X, we were able to analyze texture memory usage by the GPU and we discovered that the GPU often accesses less than 1/3 of the texture data required to be loaded in memory. A single scene often includes thousands of different textures resulting in a significant loss in effective memory and I/O bandwidth utilization due to inefficient usage.



As textures have ballooned in size to match 4K displays, efficiency in memory utilisation has got progressively worse - something Microsoft was able to confirm by building in special monitoring hardware into Xbox One X's Scorpio Engine SoC. "From this, we found a game typically accessed at best only one-half to one-third of their allocated pages over long windows of time," says Goossen.




What did they do with that information?

With this insight, we were able to create and add new capabilities to the Xbox Series X GPU which enables it to only load the sub portions of a mip level into memory, on demand, just in time for when the GPU requires the data. This innovation results in approximately 2.5x the effective I/O throughput and memory usage above and beyond the raw hardware capabilities on average. SFS provides an effective multiplier on available system memory and I/O bandwidth, resulting in significantly more memory and I/O throughput available to make your game richer and more immersive.

Finally, what people aren't appreciating is that there are 3 bars in that real-time tech demo. Xbox One X with a HDD, a 9th gen console running XVA with an SSD and a well optimized streaming system, but without SFS (Basically Series X), and finally the 3rd bar is showcasing Series X running XVA with SFS enabled this time. Microsoft is saying we are getting 2.5x effective multiplier effect on our system memory and I/O bandwidth compared to Xbox Series X optimized games. The multiplier isn't coming from comparing the Xbox One X to Series X with SFS, but from comparing Xbox Series X XVA without SFS to Xbox Series X XVA with SFS.

lZ4uksx.jpg



Microsoft planned for, created and implemented a solution they feel is integral to the architecture of Xbox Series X, and that will make all Xbox games better unless a developer out there can come up with something superior. Maybe PRT saw limited adoption because this extensive degree of work was never done before to ensure its adoption and make it more accessible and useful, and Microsoft saw a way to make that happen with their custom Xbox implementation of Sampler Feedback. Some are stuck on comparing to PS5. I know this thread is titled in a way that gets attention, but the larger point here has nothing to do with Playstation 5. This is exciting for the future of Xbox games period. If the majority of them were not using PRT, and would have likely never started using PRT, then Sampler Feedback Streaming by Microsoft is a way to make that happen once Xbox and PC games start making use of it.
 

IntentionalPun

Ask me about my wife's perfect butthole
The point I continue to make is Microsoft was comparing to all Xbox One generation games period, with PRT, without, doesn't matter. And they came away from all that extensive monitoring and analysis with this one conclusion.

edit: Your point about the 3 bars is a good one though, I do appreciate that.

edit 2: Wait, no it's not... XVA w/o SFS is XVA w/o PRT.
 
Last edited:
In that demo if they were comparing to SFS running on Series X to Xbox One X, the multiplier would be 4.9x or higher in some cases.

Microsoft is saying that games that use Sampler Feedback Streaming on Series X will have a major advantage over games that don't use Sampler Feedback Streaming on Series X where effective system memory and effective I/O bandwidth are concerned.
 

IntentionalPun

Ask me about my wife's perfect butthole
In that demo if they were comparing to SFS running on Series X to Xbox One X, the multiplier would be 4.9x or higher in some cases.

Microsoft is saying that games that use Sampler Feedback Streaming on Series X will have a major advantage over games that don't use Sampler Feedback Streaming on Series X where effective system memory and effective I/O bandwidth are concerned.
I edited my post.. your point about the 3 bars is a good one.

edit: you know what? Actually it's a terrible point lol

They are clearly comparing against non-PRT solutions since SFS IS their next-gen PRT solution.

SFS does not make PRT 3 times more efficient.. that's.. nutty.
 
Last edited:

IntentionalPun

Ask me about my wife's perfect butthole
Sorry for the edits.. you are making no sense bro.

SFS is PRT+ w/ includes PRT.. which is why MS's explanations of it include an explanation of what PRT does w/o SFS.

You don't load entire mip levels with PRT... so their explanation of how games used 3 times the memory they needed for textures, they clearly aren't referring to PRT.

MIcrosoft's demo is about an advancement over last gen.. which is A-OK.. what's bizarre is this obsession w/ acting like it's some multiplier over what PS5 could do.. PS5 almost certainly supports PRT at the very least.
 
Last edited:
The conclusion that "X, we were able to analyze texture memory usage by the GPU and we discovered that the GPU often accesses less than 1/3 of the texture data." Often.. meaning, for the majority of games?

Like.. how they said the majority of games don't use PRT last gen?

Why in the world would you read that statement and conclude that it was based on some average over "all games"?

Even if it was.. MS themselves stated that PRT had "limited adoption of this feature".. right before doing the SFS demo.

So if it wasn't used very much.. even if your weird logic is true, and they were comparing based on some average, that average would lean heavily towards how much data non PRT solutions use.

You're getting hung up on the wrong stuff my friend. The point is xbox games get to benefit NOW. They didn't before? They will now. Microsoft would have analyzed the texture data of all the biggest AAA titles and the best engines that ran on Xbox One X, enough to know exactly what they're dealing with across a massive dataset.

Microsoft literally installed special monitoring hardware into Xbox One X that would literally gift wrapped this data for them. Know what that means? Every big Frostbite release, they have the data, every big COD release, they have the data, any major release from their own first parties, they have the data, every assassins creed game, they have the data, every major unreal title, they have the data. Metro Exodus? A killer engine, right? They have the texture usage data. Bethesda games like Fallout or Doom or Elder Scrolls? They have the data. Openworld, corridor shooter, racer, fighter, they have all that data, sports games, they have the texture usage data for all of it. They might have a similar such tool inside Series X|S for all we know.

Microsoft did exactly this kind of extensive analysis when they built Xbox One X.




They analyzed all kinds of Xbox One games in prep for Xbox One X, game engines, all the top selling Xbox One games, to see how they did things. I'm pretty sure they know exactly what to expect from games this generation, and planned for such in their planning for Series X. They know better than anyone what the memory usage was on Xbox One era titles. They also know better how those same practices will scale on XVA and a faster SSD without SFS.

This is why Sampler Feedback Streaming was created and implemented into the Xbox API, why the GPU was specifically customized and made to work really well with the feature. Limited adoption of PRT is irrelevant to the larger point. For Xbox games, since they largely mostly all used this more inferior and inefficient method of texture streaming, SFS represents largely new untapped potential to make games better. Xbox One generation games did not benefit from powerful features like Sampler Feedback Streaming, but now Xbox Series X will.

I hope you see you're kind of making my point of why what Microsoft says about SFS will represent a step up for Xbox games development and Xbox games in general.
 
I edited my post.. your point about the 3 bars is a good one.

edit: you know what? Actually it's a terrible point lol

They are clearly comparing against non-PRT solutions since SFS IS their next-gen PRT solution.

SFS does not make PRT 3 times more efficient.. that's.. nutty.

If a lionshare of xbox games weren't using PRT, since you know it had "limited adoption" 2.5x-3x more efficient holds when devs start using SFS.

Coalition, makers of some of the best looking first party xbox games, said they weren't using anything like it, and cosign the big jumps in efficiency.

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.”

And the Metro Exodus devs weren't using Sampler Feedback level capability either. Two top tier, very technically talented dev studios weren't using it, so the efficiencies will hold. Microsoft literally said in the tech demo that the multiplier will ring true in AAA titles this gen that use SFS. The information from Microsoft is pretty much against you here. You want to believe the 2.5x multiplier effect won't be happening, but it almost has to if the large majority of Xbox One generation games and Xbox Series X games now aren't using PRT.

Oh, and here's more evidence that PRT is not as effective without Sampler Feedback and leads to wastage that doesn't occur with SFS. This very tweet describes SOME efficiency boost, and likely a big one, over standard PRT without Sampler Feedback.

 

Godfavor

Member
Sorry for the edits.. you are making no sense bro.

SFS is PRT+ w/ includes PRT.. which is why MS's explanations of it include an explanation of what PRT does w/o SFS.

You don't load entire mip levels with PRT... so their explanation of how games used 3 times the memory they needed for textures, they clearly aren't referring to PRT.

MIcrosoft's demo is about an advancement over last gen.. which is A-OK.. what's bizarre is this obsession w/ acting like it's some multiplier over what PS5 could do.. PS5 almost certainly supports PRT at the very least.
While this is true. By using PRT without SF, the game would need a lot of mips stored into RAM just to have them on standby in case they need to be used later (still better memory savings than having the whole texture stored into RAM). As HDD (which game engines were built in mind) is not fast enough to stream mips in real time, and there is no information from SF of what would load in the next frame.
 
Last edited:

IntentionalPun

Ask me about my wife's perfect butthole
You want to believe the 2.5x multiplier effect won't be happening, but it almost has to if the large majority of Xbox One generation games and Xbox Series X games now aren't using PRT.

I never said this and it wasn't my point.

The multiplier is real; it's just not some multiplier fully exclusive to Xbox Series consoles, as we don't know how much more efficient it is than PRT w/o it, as MS is clearly using numbers comparing to last-gen games that don't use PRT.

Their R&D lead says it's more efficient than PRT, and I believe them.. but that's not what the 3x multiplier is based on.
 
Last edited:
While this is true. By using PRT without SF, the game would need a lot of mips stored into RAM just to have them on standby in case they need to be used later (still better memory saving than storing the whole texture). As HDD is not fast enough to stream them in real time.

Yep, the fact that PRT is waster of memory due to its lack of accuracy even if it does have some memory saving elements is being lost in the mix. It's not as effective without Sampler Feedback, as explained by the man who works on this stuff more than any of us in this thread.



Do I believe SFS is roughly 2.5x more memory efficient than PRT on last gen consoles and 2.5x more memory efficient than PRT on next gen consoles? Absolutely! Because no matter how much faster the SSD gets for next gen consoles, that fundamental lack of accuracy of what is missing and what to stream next is not as powerful without Sampler Feedback.


Read the below and it's clear. PRT might save some memory, but it doesn't truly come into its own without the help of Sampler Feedback. Hence, Sampler Feedback Streaming is CLEARLY more efficient than plain PRT, I'm betting the exact 2.5x-3x rough figure Microsoft has given.



oJSt1tQ.jpg

pRrmuoi.jpg
 
I never said this and it wasn't my point.

The multiplier is real; it's just not some multiplier fully exclusive to Xbox Series consoles, as we don't know how much more efficient it is than PRT w/o it, as MS is clearly using numbers comparing to last-gen games that don't use PRT.

Their R&D lead says it's more efficient than PRT, and I believe them.. but that's not what the 3x multiplier is based on.

I never said it was... I'm only talking about how this can make xbox series games better over games that don't use it. I'm not comparing it to anything else. But clearly this is a major focus of Microsoft's entire Xbox Velocity Architecture for next gen. SFS is the heart of it all.
 

Godfavor

Member
SF doesn't actually tell you anything about the next/future frame - that remains a software problem until we invent hardware that can predict the future.
From MS:

CheckAccessFullyMapped or other residency determination mechanisms, to detect and load non-resident tiles on demand has room for improvement. That’s where sampler feedback comes in. Sampler feedback streamlines the process of writing out “ideal mip levels”, since the detection of the “ideal mip level” and writing-it-out can be done all in one step, allowing the driver to optimize this process.

Edit: https://microsoft.github.io/DirectX-Specs/d3d/SamplerFeedback.html

Edit2: Along with each tiled texture, create a small “MinMip map” texture and small “feedback map” texture.
The MinMip map represents per-region mip level clamping values for the tiled texture; it represents what is actually loaded.
The feedback map represents and per-region desired mip level for the tiled texture; it represents what needs to be loaded.
Update the mip streaming engine to stream individual tiles instead of mips, using the feedback map contents to drive streaming decisions.
 
Last edited:

IntentionalPun

Ask me about my wife's perfect butthole
I never said it was... I'm only talking about how this can make xbox series games better over games that don't use it. I'm not comparing it to anything else. But clearly this is a major focus of Microsoft's entire Xbox Velocity Architecture for next gen. SFS is the heart of it all.
For sure.. others have implied it's a multiplier over what other systems could do, sorry if I misunderstood you.

And I do agree with you, the multiplier is real, it's logical.

People have been repeating the myth that every game used PRT last generation to justify it not being a good measure... but that wasn't true. PRT existed, the tech was invented many years ago.. it was used here or there.. but 5400RPM drives weren't really going to ever be able to do PRT for the most part.

Many games did some really insane things with data streaming as far back as Xbox 360 days... the fact that ANYONE pulled off anything like PRT on last gen, and iD the gen before it.. is insane...and the best of the best streaming engines likely outdid what MS is using as a baseline by quite a bit. So it's not likely to be a "3 times multiplier over the absolute best streaming data implementation from last gen".. but it definitely would still be an advancement over even those engines.

Hopefully it's also as easy to use as MS has said, and hopefully it actually DOES get used lol

Either way, fast I/O will enable people doing similar things w/ engines that don't actually leverage SFS on Xbox as well. It's sort of that compounding factor of having many times the speed of last gen as well as that much more processing power on the CPU side. The speed enables techniques that make memory usage more EFFICIENT as well as just... well.. giving more speed.
 
Last edited:
For sure.. others have implied it's a multiplier over what other systems could do, sorry if I misunderstood you.

And I do agree with you, the multiplier is real, it's logical.

People have been repeating the myth that every game used PRT last generation to justify it not being a good measure... but that wasn't true. PRT existed, the tech was invented many years ago.. it was used here or there.. but 5400RPM drives weren't really going to ever be able to do PRT for the most part.

Many games did some really insane things with data streaming as far back as Xbox 360 days... the fact that ANYONE pulled off anything like PRT on last gen, and iD the gen before it.. is insane...and the best of the best streaming engines likely outdid what MS is using as a baseline by quite a bit. So it's not likely to be a "3 times multiplier over the absolute best streaming data implementation from last gen".. but it definitely would still be an advancement over even those engines.

Hopefully it's also as easy to use as MS has said, and hopefully it actually DOES get used lol

Either way, fast I/O will enable people doing similar things w/ engines that don't actually leverage SFS on Xbox as well. It's sort of that compounding factor of having many times the speed of last gen as well as that much more processing power on the CPU side. The speed enables techniques that make memory usage more EFFICIENT as well as just... well.. giving more speed.

It better fucking get used!!! :messenger_tears_of_joy:

Or else this is gonna be me coming back to the thread 7 years from now..

Jimmy Kimmel Oscars GIF by The Academy Awards
 
Last edited:

Thief1987

Member
Yep, the fact that PRT is waster of memory due to its lack of accuracy even if it does have some memory saving elements is being lost in the mix. It's not as effective without Sampler Feedback, as explained by the man who works on this stuff more than any of us in this thread.



Do I believe SFS is roughly 2.5x more memory efficient than PRT on last gen consoles and 2.5x more memory efficient than PRT on next gen consoles? Absolutely! Because no matter how much faster the SSD gets for next gen consoles, that fundamental lack of accuracy of what is missing and what to stream next is not as powerful without Sampler Feedback.


Read the below and it's clear. PRT might save some memory, but it doesn't truly come into its own without the help of Sampler Feedback. Hence, Sampler Feedback Streaming is CLEARLY more efficient than plain PRT, I'm betting the exact 2.5x-3x rough figure Microsoft has given.



oJSt1tQ.jpg

pRrmuoi.jpg

How many more times do you want to post these screenshots? You already done this atleast ten times in last 24 hours alone.
 

IntentionalPun

Ask me about my wife's perfect butthole
It better fucking get used!!! :messenger_tears_of_joy:

Or else this is gonna be me coming back to the thread 7 years from now..

Jimmy Kimmel Oscars GIF by The Academy Awards
lol

I'm just being cautious because everyone can find the articles from early last gen about what PRT was going to bring; and it was used scantly.

But that was for a very obvious reason; the massive i/o bottleneck. We also had shit CPUs and data operations that would take up more of that CPU if you tried to do things in the background.

So it should be getting used.

The combo of all of these things will make these next-gen systems an I/O playground. For the high end devs it's going to mean some crazy advancements, and for the low-end devs they can likely easily catch up to what high-end devs were doing last gen, and more.

Do I believe SFS is roughly 2.5x more memory efficient than PRT on last gen consoles and 2.5x more memory efficient than PRT on next gen consoles? Absolutely!

This is where we disagree though.. it's pretty clear to me MS isn't comparing to PRT, but to non-PRT solutions using XVA.
 
Last edited:

sinnergy

Member
How many more times do you want to post these screenshots? You already done this atleast ten times in last 24 hours alone.
And? It's the guy who practically wrote the code.. so he is right. Yet everyone keeps posting here .. they should close this thread and move on.
 
Last edited:

Thief1987

Member
And? It's the guy who practically wrote the code.. so he is right. Yet everyone keeps posting here .. they should close this thread and move on.
How is it relevant? Nobody said that content of tweets is not legit. Doesn't mean that we should post this every 15 minutes.
 
SF doesn't actually tell you anything about the next/future frame - that remains a software problem until we invent hardware that can predict the future.

Since you mentioned future frames I have a question about that.

In The Road to PS5 it was talked about how developers had to load assets for the next X seconds of gameplay.

How are they able to predict what those assets will be? It's not like for the next 30 seconds they will know what the player will see.
 
Top Bottom