• 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

jroc74

Phone reception is more important to me than human rights
i know where you going...Pana and i gave you at least 3 or 4 practical example (that you ignored) of how file size is the limit and not the SSD / io speed .......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
...stop focusing on the best case scenario....
First of all his example was about Spider-man and general open world assets streaming (if you cannot tie the speed of movement to the speed data is fetched from the disk at well 🤷‍♂️), second of all you made several straw man arguments and I already said why I disagreed with them and especially how you framed them.
Should we start again from 22 GB/s being an edge case or how we are not talking about streaming completely unique data all of the time vs reuse/dedicating more RAM to what is on screen, reducing buffers depth, and thus streaming the same blocks of data over and over or how it is not how many GB you move per second but how many ms you need to move a specific piece of data that is important?

You are repeating them over and over regardless of what people reply back and using the fact you are repeating it over and over as evidence you were right to begin with…. If people stop replying == they admitted defeat eh ;)?
Exactly.

Once again...this is why I asked whats the average and best case scenarios for Series consoles.

We have this for PS5, I dont think we have this for Series consoles.
 
Last edited:

Lethal01

Member
The TL;DR of all this is that the 3rd party games, where you can actually make 1:1 comparisons between consoles, will be limited by whatever PC does anyway, so all of this kinda doesn't matter?

Disappointed Let Down GIF by SWR3

Only if you care more about comparing console that games being better than they would have been.
 

jose4gg

Member
I want to clarify something for some people, MS has been very specific by saying that often, and here is the word used OFTEN, GPUs only uses 1/3 of the textures. Here is the meaning of the word often:

c3QiO0K.png


Often doesn't mean, "the majority of the time", it does not mean "50% of the time", it does not mean "75% of the time", if that would be the case MS would be saying "THE MAJORITY OF THE TIME GPUS ONLY USE... 1/3.... TEXTURE", That is not what they are saying.


You cannot simply compare the speed of the PS5 I/O that should be maintaining its speed 90+% of the time, with something that happens "in many instances", and you shouldn't be saying that Xbox Series Consoles real speed is "x", if that "x" is only obtained "OFTEN",
 

Three

Member
On the contrary, the multiplier isn't compared to Xbox One Games at all, it's compared to a 9th gen console optimized Xbox Series X streaming system with an SSD, but one that isn't using SFS. In the below pic see the bar all the way to the top? That's the Xbox One X bar. The bar SFS is being compared to is a 9th gen console that uses a fast SSD, aka that's Xbox Series X if it weren't using SFS.

jIZrqgS.jpg
]


They even point out that the middle bar is unrealistically optimistic and would actually be using more memory due to over-streaming, meaning the SFS advantage is even bigger. If SFS on Series X were being compared to the Xbox One X or Xbox One era, the multiplier would actually be much higher. It would be 4.4x. But compared to the next gen equivalent to Series X, SFS' advantage is between 2.5x-3x.

Below is the part where they tell you that only the top bar is Xbox One X, not the middle bar, which is what the SFS bar is being compared to.





Xbox One X games can't use Xbox Velocity Architecture. Notice what the Sampler Feedback Streaming bar is being compared to? It's being compared to Xbox Velocity Architecture without Sampler Feedback Streaming.

uJ7t4Gm.jpg


Here is the source of the quoted 2-3x multiplier:
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. "So if a game never had to load pages that are ultimately never actually used, that means a 2-3x multiplier on the effective amount of physical memory, and a 2-3x multiplier on our effective IO performance."
From here

They monitored xbox one x games and found the 2-3x multiplier. It is compared to xbox one games.

What you have linked isn't a typical game. it is a demo of SFS and hence it would not be designed in a way to optimise memory usage based on view to begin with. After all it's designed to showcase the streaming. That's why the initial values would be much higher and exaggerated.

XVA without SFS would be more like basing what you load on distance or around corridors, an amount you load ahead would depend on what the player can do and drive speed whereas with PRT+/SFS you also include things like occlusion culling based on view.
 

jroc74

Phone reception is more important to me than human rights
Let's clarify something real quick.

Comparing BCPack to Kraken alone is not entirely correct, though it's mostly fair.


Emphasis mine


So, let's break down the main differences, starting by another excellent post, this time from this very own thread:



XSX/S
  • Zlib for general game data
  • BCPack for texture data
PS5
  • Oodle Kraken for all game data
  • Oodle Texture for texture data

Here's the deal - While most game data is comprised of Textures, the PS5 already has an advantage on other game data due to Kraken's advantage over Zlib. The compression ratio is already higher as a baseline, but the decompression stage is pretty much a no context. In theory, game devs can use higher compression on Kraken due to the time it takes to decompress said data (more compression = more time to decompress). So while a developer may settle on a 2:1 (made up numbers) with Zlib, it may go as far as 3:1 with Kraken, due to the time it takes to decompress being lower.

That said, the real advantage of BCPack comes with it being an all in one tool for both lossy and lossless compression. We don't know the exact numbers (as far as I'm aware, please correct me), but they could very well be better than Oodle Texture. The key difference comes to cost (2 hardware decompressors, with a tiny CPU impact vs single piece of hardware with no CPU cost). There's also nothing that prevents MS from using Kraken, but again, there would be added CPU cost (and wouldn't financially make sense)

Still, Microsoft's all in one solution is yet to yield better results on Next-Gen patched games (meaning games that are not gen aware, but actually patched to exploit next gen). This includes both decompression (Playstation loading times are more often than not ahead) but also games sizes (as evidenced by multiple games and patches).

On the topic of Compression and more in line with my fields of expertise, VFXVeteran VFXVeteran do you have any idea if games using lossless audio files still run with WAV, or have they moved to Flac or other lossless compressed file formats?
This is whats wild in all this.

For the SSD's in general the PS5 is starting from a higher baseline anyway.
 
Here is the source of the quoted 2-3x multiplier:

From here

They monitored xbox one x games and found the 2-3x multiplier. It is compared to xbox one games.

What you have linked isn't a typical game. it is a demo of SFS and hence it would not be designed in a way to optimise memory usage based on view to begin with. After all it's designed to showcase the streaming. That's why the initial values would be much higher and exaggerated.

XVA without SFS would be more like basing what you load on distance or around corridors, an amount you load ahead would depend on what the player can do and drive speed whereas with PRT+/SFS you also include things like occlusion culling based on view.


But we have a real-time demo comparing it to a 9th gen console aka Xbox Series X using XVA but without SFS.

inSiAnF.jpg


Below it's being compared against an optimized gen 9 streaming engine. You see XVA? That means it's Xbox Series X with an SSD, and it's purposely conservative on the memory use. They state it would be more. In fact, in the picture above the Xbox One X is the top bar. The advantage for SFS is 5.2x. Microsoft throughout the entire demo is quite literally showing you the advantages over Xbox One, but are only having the multiplier reflect the advantage over Series X WITHOUT SFS.




Also, you don't seem to have watched the demo at all where Microsoft actually states that they intentionally made the things they were comparing SFS to BETTER than would be in real world use by having them utilize less memory.

"I should also mention the absolute numbers you see in the memory bars here. They aren't all that high, especially compared to the 10 or 16GB of memory in our consoles. Our content in this tech demo is fairly simple A real AAA title would likely have significantly more complex materials and more objects visible. Crucially however, the comparison between sfs and traditional mip streaming holds true regardless of material complexity. The numbers will scale with content and the multiplier will still ring true."

They are confirming in the demo that the multiplier will STILL be true even in AAA games with significantly more complex materials and more objects visible.

 
Last edited:

Three

Member
But we have a real-time demo comparing it to a 9th gen console aka Xbox Series X using XVA but without SFS.

inSiAnF.jpg


Below it's being compared against an optimized gen 9 streaming engine. You see XVA? That means it's Xbox Series X with an SSD, and it's purposely conservative on the memory use. They state it would be more. In fact, in the picture above the Xbox One X is the top bar. The advantage for SFS is 5.2x. Microsoft throughout the entire demo is quite literally showing you the advantages over Xbox One, but are only having the multiplier reflect the advantage over Series X WITHOUT SFS.


As i mentioned there is a difference between a game and a demo. You've only linked these two things because 2.5x is near the 2-3x that MS saw in xbox one games.

A game would usually build in optimisations knowing they are memory limited to begin with. A faster drive would mean they would need to load less into memory in advance depending on player speed/interaction/level design to make sure assets are resident when needed.
A faster drive would need less data loaded in RAM at any given time since they are loaded in faster (XVA without SFS) on top of that you have PRT+/SFS that does complete occlusion culling.


Look at how data is handled in the Spiderman presentation, specifically the part about imposters. This would not be present in this demo. In this demo there is no initial optimisation hence you would be looking at massive memory usage to begin with and it showing a greater multiplier than those seen in typical xbox one games.
 
Last edited:

mrmeh

Member
This is a weird take. I would perhaps have believed it if Control was the only game, but that's not the case anymore.

And unless the person can publicly say (and prove) exactly what it is, we should Ockham Razor this issue.

I mentioned Control specifically in response to that post. I'll explain the logic...

The PS5 SSD is much faster? so if its loading assets at the same speed as Series X then it is constrained by something else i.e. CPU

Both consoles have a very similar CPU hence when the Load times are very close it is logical that the similarity is caused by the CPU decompressing the textures instead of some fancy HW decoder.

Stay with me... Now.. if it is using Oodle/Kraken and its taking the same time to load a game what's the fucking point in the Uber SSD...

Can't have it both ways..

I believe that the large difference in overall game size we have seen in some cross format titles is down to that fact that Xbox shares a SDK with PC and PS5 requires a specific new build with a new asset pipeline. Note that the supposed compression % don't add up to the difference in game sizes.

Will PS5 have smaller game sizes through this gen than Series probably yes... will the difference be as big as it is with some titles now ...probably not.

I guess the other probability could be its using Kraken but not Oodle but then like explained before the compression ratios are off.
 

Connxtion

Member
R5ca0d0b39e236bd924109dba8681e8b6

Am going to say the 1 second to 8 seconds is a marketing deal thing because Sony has the marketing deal with RE8, & the only thing they could improve over the XBox without looking like they deliberately done it was the load time. As there is zero reason the XSX/S can’t load the game in 2 to 3 seconds like every other game where the PS5 has loaded quicker.

How does Subnotica and it’s sequel compare in loading? I suspect a marginal difference.
(I can time the XSX versions when I get back home tomorrow)
Update to the last part.

Subnautica from pressing A on Survival to Press any button screen. 19 seconds (16 seconds to black screen finished loading)

Subnautica BZ same as above, 22 seconds (20 seconds to black screen finished loading)

both can go a couple seconds higher randomly 🤷‍♂️ But from multiple tests the above values are what I got most times.
 

VFXVeteran

Banned
On the topic of Compression and more in line with my fields of expertise, VFXVeteran VFXVeteran do you have any idea if games using lossless audio files still run with WAV, or have they moved to Flac or other lossless compressed file formats?
I'm sorry to say but that is not my field of expertise. I wouldn't have any info on audio files for games. Sorry man.
 

Heisenberg007

Gold Journalism
I mentioned Control specifically in response to that post. I'll explain the logic...

The PS5 SSD is much faster? so if its loading assets at the same speed as Series X then it is constrained by something else i.e. CPU

Both consoles have a very similar CPU hence when the Load times are very close it is logical that the similarity is caused by the CPU decompressing the textures instead of some fancy HW decoder.

Stay with me... Now.. if it is using Oodle/Kraken and its taking the same time to load a game what's the fucking point in the Uber SSD...

Can't have it both ways..

I believe that the large difference in overall game size we have seen in some cross format titles is down to that fact that Xbox shares a SDK with PC and PS5 requires a specific new build with a new asset pipeline. Note that the supposed compression % don't add up to the difference in game sizes.

Will PS5 have smaller game sizes through this gen than Series probably yes... will the difference be as big as it is with some titles now ...probably not.

I guess the other probability could be its using Kraken but not Oodle but then like explained before the compression ratios are off.
1. Compression ratios will never always scale linearly. For example, in Crash Bandicoot, the difference is just 10%. In Yakuza, there is no difference. In Control, the difference is around 60%. In Subanutica, the difference is around 125%. They will never scale linearly.

2. Why do both PS5 and XSX load Control in the same time? It's not because of CPU, and PS5 isn't constrained by anything in this scenario. You're ignoring one major fact. On XSX, the game is decompressing the files from a relatively uncompressed 42GB game file and then loading the game, in say, 5 seconds. On the PS5, the game is decompressing the files from a heavily compressed 25GB file and the loading the game in the same time, i.e., 5 seconds. So, in essence, the PS5 is doing more in less time.

3. The same example can be more easily observed in Resident Evil Village. On PS5, the game loads in 1.5 seconds. On the XSX, the game is 30% bigger and also loads ~400% slower (around 8 seconds). What does this mean? It means that XSX loading time could theoretically be improved if the game size was even bigger, so there were fewer files to decompress and then load.
 

mrmeh

Member
1. Compression ratios will never always scale linearly. For example, in Crash Bandicoot, the difference is just 10%. In Yakuza, there is no difference. In Control, the difference is around 60%. In Subanutica, the difference is around 125%. They will never scale linearly.

2. Why do both PS5 and XSX load Control in the same time? It's not because of CPU, and PS5 isn't constrained by anything in this scenario. You're ignoring one major fact. On XSX, the game is decompressing the files from a relatively uncompressed 42GB game file and then loading the game, in say, 5 seconds. On the PS5, the game is decompressing the files from a heavily compressed 25GB file and the loading the game in the same time, i.e., 5 seconds. So, in essence, the PS5 is doing more in less time.

3. The same example can be more easily observed in Resident Evil Village. On PS5, the game loads in 1.5 seconds. On the XSX, the game is 30% bigger and also loads ~400% slower (around 8 seconds). What does this mean? It means that XSX loading time could theoretically be improved if the game size was even bigger, so there were fewer files to decompress and then load.

2. You are wrong. You're fan boying instead of using logic. Back compat games are the same, they take the same time to load because they are constrained by the CPU. Your argument here is essentially that the HW decompressor and fast SSD primary use is to make the game smaller. Fucking nuts. They're loading the same amount of data, the Xbox version just is not packaged efficiently and both systems are using the CPU to decode textures.

3. That is one of the titles where I suspect it maybe using Kraken/Oodle. It's loading much quicker. Having said that The PC version is the same size as the Xbox version and as the PC version does not specify an SSD I'd expect it has duplicate assets and as it the same size as the X I'd assume that does as well so again probably packaged inefficiently for Series. If those numbers are accurate the load times also fit in reasonably with the Xbox 4.8gb/s vs 22gb/s bandwith bonus PS5 can theoretically get with all its toys turned on.

Look you're just a waste of my time, you shit post PS5 fan boy garbage all the time, I love tech discussions but your just here for the war, I'm out.
 
Maybe Sony PS plus deal, as noted the PS5 build had major code changes. They put in extra time re packaging assets, Xbox series version not so much.



I think there are bits here and there but I don't think direct storage has been fully released into the latest version of DX, maybe new first party games later this year will be able to show off some of the more interesting features. Same with the PS5 though, what games have been confirmed using Kraken/oodle? If the game isn't loading significantly faster on PS5 than X then it isn't. Still lots of cool stuff to come.
The thing is that Xbox and its fanboys are often screaming out of the top of their lungs of how great Xbox is. Xbox talks a lot but rarely walks the walk. Don't get me wrong, Playstation sometimes misses the mark too in delivering the things they promised.

The difference is often that Playstation gives out the info silently and often let the games do the confirming of their tech, business plans and practices. Whereas Xbox often aims for 'epic headlines', such as 'Xbox is Working on New IP That Will “Blow Your Mind'.

Xbox reminds me a lot of Square Enix, where they keep promising you the moon, keep saying things are going great and please be excited about an announcement of an announcement. Their whole PR side seems extremely insecure, chaotic, incompetent, childish and disingenuous.
A few examples of this; Xbox always opening with 'World premier' and 'Xbox exclusive' at nauseum with their presentations. Also their fake crowd cheering and applauding. It's rather disrespectful to treat your customers like that and think their fanbase is dumb enough to buy all that.
Besides this, they also always talk about their quantity of games, because they know damn well that they simply can't compete with Playstation when it comes to quality. The TLOU2 Xbox review is the perfect example of this. Xbox knows the level of their quality and that is why they often have to scream for attention with headlines such as 'will blow your mind'.

Xbox recently has definitely made some good moves, but those good moves are all still theories. Most of it has yet to be proven by their games.

You said 'If the game isn't loading significantly faster on PS5 than X then it isn't'. But we both know that games that are most likely to utilize these features, are Playstation exclusives. So we probably won't often get direct comparisons. We have seen time and time again of just how great the Playstation exclusives are. Both creatively as well as technically. Even third party games more often than not run better on the Playstation 5. Granted, the difference are often small. But the narrative that Xbox and its fanboys have been screaming that the Xbox is far superior than the Playstation 5 is proven to be wrong many times.

I would love to make a detailed thread about this with all the details of why I think what I think. But unfortunately I can't make threads as I'm still new and don't have a membership.

Sorry for being off topic btw, but damn I just wish Xbox would provide real competition by letting their games do the talking. Stay silent, play the catch up game, focus on what you need to do and THEN blow our minds and show us why Xbox is better than Playstation.
 

Heisenberg007

Gold Journalism
2. You are wrong. You're fan boying instead of using logic. Back compat games are the same, they take the same time to load because they are constrained by the CPU. Your argument here is essentially that the HW decompressor and fast SSD primary use is to make the game smaller. Fucking nuts. They're loading the same amount of data, the Xbox version just is not packaged efficiently and both systems are using the CPU to decode textures.

3. That is one of the titles where I suspect it maybe using Kraken/Oodle. It's loading much quicker. Having said that The PC version is the same size as the Xbox version and as the PC version does not specify an SSD I'd expect it has duplicate assets and as it the same size as the X I'd assume that does as well so again probably packaged inefficiently for Series. If those numbers are accurate the load times also fit in reasonably with the Xbox 4.8gb/s vs 22gb/s bandwith bonus PS5 can theoretically get with all its toys turned on.

Look you're just a waste of my time, you shit post PS5 fan boy garbage all the time, I love tech discussions but your just here for the war, I'm out.
You're comparing BC games? Why? Why compare PS4/Xbox One games on PS5 and XSX? Of course, they will off-load to CPU because PS4 and Xbox One didn't have any HW decompressors to right code for. I was only talking about this-gen games, not last-gen games.

You're using extremely weird examples and logics and calling me a shitposter.

And (for PS5/XSX games), they are loading the same amount of data. I didn't contradict that; I mentioned the point that you missed. PS5 is first decompressing more stuff before loading it. XSX is loading less stuff before decompressing it. It's evident in the file size when the textures in the game are the same for both consoles.

Plus, you are making an assumption that all developers are making the same mistake, using the same PC version for the XSX versions and therefore creating the same inefficiency. What's your evidence for that?
 
Last edited:

Lethal01

Member
But we have a real-time demo comparing it to a 9th gen console aka Xbox Series X using XVA but without SFS.

inSiAnF.jpg


Below it's being compared against an optimized gen 9 streaming engine. You see XVA? That means it's Xbox Series X with an SSD, and it's purposely conservative on the memory use. They state it would be more. In fact, in the picture above the Xbox One X is the top bar. The advantage for SFS is 5.2x. Microsoft throughout the entire demo is quite literally showing you the advantages over Xbox One, but are only having the multiplier reflect the advantage over Series X WITHOUT SFS.




Also, you don't seem to have watched the demo at all where Microsoft actually states that they intentionally made the things they were comparing SFS to BETTER than would be in real world use by having them utilize less memory.

"I should also mention the absolute numbers you see in the memory bars here. They aren't all that high, especially compared to the 10 or 16GB of memory in our consoles. Our content in this tech demo is fairly simple A real AAA title would likely have significantly more complex materials and more objects visible. Crucially however, the comparison between sfs and traditional mip streaming holds true regardless of material complexity. The numbers will scale with content and the multiplier will still ring true."

They are confirming in the demo that the multiplier will STILL be true even in AAA games with significantly more complex materials and more objects visible.



"Traditional Mip Streaming" and "without SFS" in this case would not include PRT correct?

so the 2.5x~3x in this case is not over what we would expect from a game using regular PRT/PRT+ right?
 

mrmeh

Member
The thing is that Xbox and its fanboys are often screaming out of the top of their lungs of how great Xbox is. Xbox talks a lot but rarely walks the walk. Don't get me wrong, Playstation sometimes misses the mark too in delivering the things they promised.

The difference is often that Playstation gives out the info silently and often let the games do the confirming of their tech, business plans and practices. Whereas Xbox often aims for 'epic headlines', such as 'Xbox is Working on New IP That Will “Blow Your Mind'.

Xbox reminds me a lot of Square Enix, where they keep promising you the moon, keep saying things are going great and please be excited about an announcement of an announcement. Their whole PR side seems extremely insecure, chaotic, incompetent, childish and disingenuous.
A few examples of this; Xbox always opening with 'World premier' and 'Xbox exclusive' at nauseum with their presentations. Also their fake crowd cheering and applauding. It's rather disrespectful to treat your customers like that and think their fanbase is dumb enough to buy all that.
Besides this, they also always talk about their quantity of games, because they know damn well that they simply can't compete with Playstation when it comes to quality. The TLOU2 Xbox review is the perfect example of this. Xbox knows the level of their quality and that is why they often have to scream for attention with headlines such as 'will blow your mind'.

Xbox recently has definitely made some good moves, but those good moves are all still theories. Most of it has yet to be proven by their games.

You said 'If the game isn't loading significantly faster on PS5 than X then it isn't'. But we both know that games that are most likely to utilize these features, are Playstation exclusives. So we probably won't often get direct comparisons. We have seen time and time again of just how great the Playstation exclusives are. Both creatively as well as technically. Even third party games more often than not run better on the Playstation 5. Granted, the difference are often small. But the narrative that Xbox and its fanboys have been screaming that the Xbox is far superior than the Playstation 5 is proven to be wrong many times.

I would love to make a detailed thread about this with all the details of why I think what I think. But unfortunately I can't make threads as I'm still new and don't have a membership.

Sorry for being off topic btw, but damn I just wish Xbox would provide real competition by letting their games do the talking. Stay silent, play the catch up game, focus on what you need to do and THEN blow our minds and show us why Xbox is better than Playstation.

I think there's allot of bullshit made up by fanboys that just fills in the void left by either Sony staying quiet and MS releasing all the specs..

I don't want to be to harsh on MS as its a great machine, but Sony has done a better job of getting its initial games out. Really looking forward to seeing how X handles flight sim.

Yep the exclusives are where we will see the tech first.. Ratchet will be a great tech showcase for the SSD and compression.

There's also an ignore button if the fanboys get too annoying.. :)
 
I think there's allot of bullshit made up by fanboys that just fills in the void left by either Sony staying quiet and MS releasing all the specs..

I don't want to be to harsh on MS as its a great machine, but Sony has done a better job of getting its initial games out. Really looking forward to seeing how X handles flight sim.

Yep the exclusives are where we will see the tech first.. Ratchet will be a great tech showcase for the SSD and compression.

There's also an ignore button if the fanboys get too annoying.. :)

Yes but why would you press the ignore button when you can see the desperation and those salty tears coming out? It's the ultimate form of entertainment, maybe even better than actually playing the next-gen games we got now vs the eternal promises of uncle Phil. I actually have to thank all those warriors for their pure fantasy stories. I'm thinking each day what story I would tell to my little boy at night and this forum keeps giving me the most amazing fairytales.

Rick And Morty GIF
 
Last edited:
"Traditional Mip Streaming" and "without SFS" in this case would not include PRT correct?

so the 2.5x~3x in this case is not over what we would expect from a game using regular PRT/PRT+ right?

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 equally doubt the Metro Exodus technical director would be saying they plan to use it in future titles if they were already using exactly that for years.

And let's look at what the coalition has to say. Clearly this isn't something that was already in use on Xbox.


If the below statement by Coalition doesn't make crystal clear that this is new functionality for Xbox game developers then nothing will.

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.

So coalition thinks SFS is a game changer. Says future titles will use it. In other words, Gears 5 and any other title they've ever made makes absolutely no use of anything remotely like it. They believe it will allow them to increase texture detail beyond what they can fit into memory, and change loading to just before they need it instead of the more traditional level loading approach.
 
Last edited:

Thief1987

Member
On XSX, the game is decompressing the files from a relatively uncompressed 42GB game file and then loading the game, in say, 5 seconds. On the PS5, the game is decompressing the files from a heavily compressed 25GB file and the loading the game in the same time, i.e., 5 seconds. So, in essence, the PS5 is doing more in less time.
There is no compression at all afaik, atleast on pc, but going by the size of xbox version there is also no compression.
 
Last edited:

Heisenberg007

Gold Journalism
There is no compression at all afaik, atleast in pc version, but going by the size of xbox version there is also no compression.
I believe there was. The reason is that the Xbox One version was roughly around 42.3 GB, and the XSX version was 42.1 GB. The XSX version had ray-tracing + 4K textures that the Xbone version didn't. So there was at least some compression or else that XSX version would have been much bigger in size.

When someone asked the Control devs on Twitter if there is any chance that the XSX version could be compressed more, he responded that "no, sorry, that was the best we could do".

So I think there is enough to suggest that there was some compression.
 

sinnergy

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 equally doubt the Metro Exodus technical director would be saying they plan to use it in future titles if they were already using exactly that for years.

And let's look at what the coalition has to say. Clearly this isn't something that was already in use on Xbox.


If the below statement by Coalition doesn't make crystal clear that this is new functionality for Xbox game developers then nothing will.
I bet nothing will .. but that’s a hunch .
 
I think there's allot of bullshit made up by fanboys that just fills in the void left by either Sony staying quiet and MS releasing all the specs..

I don't want to be to harsh on MS as its a great machine, but Sony has done a better job of getting its initial games out. Really looking forward to seeing how X handles flight sim.

Yep the exclusives are where we will see the tech first.. Ratchet will be a great tech showcase for the SSD and compression.

There's also an ignore button if the fanboys get too annoying.. :)
I absolutely agree with you. Xbox is a great machine and it's unfortunate that it gets clouded with all these things instead of showcasing its prowess with its games.

It's rather hard to ignore for me personally to be frank. As it's everywhere on the internet. Even from those incompetent journalists.

On the other hand 'console wars', can also turn out into a fun and insightful discussion. Where both parties can learn from as well as better each other.
 
Last edited:

93xfan

Banned
Yes but why would you press the ignore button when you can see the desperation and those salty tears coming out? It's the ultimate form of entertainment, maybe even better than actually playing the next-gen games we got now vs the eternal promises of uncle Phil. I actually have to thank all those warriors for their pure fantasy stories. I'm thinking each day what story I would tell to my little boy at night and this forum keeps giving me the most amazing fairytales.

Rick And Morty GIF

At the end of the day, I don’t think either side needs to be desperate. Either way you go, you’re getting a truly great gaming experience.
 
At the end of the day, I don’t think either side needs to be desperate. Either way you go, you’re getting a truly great gaming experience.

Exactly. It’s amazing how overly emotional a great part of the gamer community is. And with that I don’t mean Xbox fans specifically, but all of them in general, with some exceptions.

Basketball Wives Shes Fifty Shades Of Cray GIF by VH1
 
At the end of the day, I don’t think either side needs to be desperate. Either way you go, you’re getting a truly great gaming experience.

That's true, but it's interesting discussion.

But I think where I'm at right now is I will stand on comments from the Metro games Technical Director and the Technical Director for Coalition.

Both say SFS is a very big deal, and plan to make use of it in their FUTURE titles. So two studios who have made some of the best looking last gen games on Xbox confirm they weren't using anything like it.
 

DForce

NaughtyDog Defense Force
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 equally doubt the Metro Exodus technical director would be saying they plan to use it in future titles if they were already using exactly that for years.

And let's look at what the coalition has to say. Clearly this isn't something that was already in use on Xbox.


If the below statement by Coalition doesn't make crystal clear that this is new functionality for Xbox game developers then nothing will.
You're talking about something that would easier to adapt from Triple A and Non-Triple A developers.

Unity supports Virtual texture streaming, which only loads a portion of the texture to save room on virtual memory.


fOD36AR.png


QXmlA7n.png


A lot of Xbox fans are treating this as a new and unique feature from Microsoft, but we've seeing similar things being done in Unity and Unreal engine.

If we're seeing similar techniques being done on other engines, then I wouldn't label it a brand new feature.
 

Panajev2001a

GAF's Pleasant Genius
I bet nothing will .. but that’s a hunch .
Cannot do much when arguing a straw man: nobody is arguing that a developer on XSX|S will get a LOT better results than on XOX thanks to XVA (DirectStorage + BCPack + SFS [+ SSD]).
Nobody is arguing against it so posting statements about how a MS developer loves XSX and the impact of SFS over what they had on XOX is beautiful, believable (for sure a much slower and much higher latency storage solution such as a mechanical HDD and much less efficient File I/O API’s hampered real-time assets streaming solutions… or “using the SSD as if it were an extension of RAM” as MS put it too), but beside the point.
Nobody is arguing against that.

Traditional mip streaming in today's AAA games market almost certainly means some form of virtualized texture paging system aka PRT
Says you (traditional mip streaming, by definition, upload the entire mip level… think of the texture made up of several super thin slices each smaller than the one before it)… the person you quote already put into context SFS, PRT, regular full mip level texture loading, and thus the multiplier.
DC9lDs1.jpg
lyRWHgi.jpg

People tried to explain it using different sources, including Xbox dedicated gaming forum threads, but since you used to quote him a lot before… I keep quoting Stanard, but you stopped acknowledging his words for some reason ;).

It is quite clear, but it is not an insult to XVA if we do not add another 2-3x multiplier out of thin air and the end result (in pure theoretical peak numbers, reality is what it is and what you can see in exclusives vs multi platform titles) is not as close to PS5’s SSD as you seem to think (especially considering most third parties will take a while to properly make use of either solution… aside from generous offers from the console maker for timed exclusives perhaps).
 

Corndog

Banned
2. You are wrong. You're fan boying instead of using logic. Back compat games are the same, they take the same time to load because they are constrained by the CPU. Your argument here is essentially that the HW decompressor and fast SSD primary use is to make the game smaller. Fucking nuts. They're loading the same amount of data, the Xbox version just is not packaged efficiently and both systems are using the CPU to decode textures.

3. That is one of the titles where I suspect it maybe using Kraken/Oodle. It's loading much quicker. Having said that The PC version is the same size as the Xbox version and as the PC version does not specify an SSD I'd expect it has duplicate asset
Cannot do much when arguing a straw man: nobody is arguing that a developer on XSX|S will get a LOT better results than on XOX thanks to XVA (DirectStorage + BCPack + SFS [+ SSD]).
Nobody is arguing against it so posting statements about how a MS developer loves XSX and the impact of SFS over what they had on XOX is beautiful, believable (for sure a much slower and much higher latency storage solution such as a mechanical HDD and much less efficient File I/O API’s hampered real-time assets streaming solutions… or “using the SSD as if it were an extension of RAM” as MS put it too), but beside the point.
Nobody is arguing against that.


Says you (traditional mip streaming, by definition, upload the entire mip level… think of the texture made up of several super thin slices each smaller than the one before it)… the person you quote already put into context SFS, PRT, regular full mip level texture loading, and thus the multiplier.
DC9lDs1.jpg
lyRWHgi.jpg

People tried to explain it using different sources, including Xbox dedicated gaming forum threads, but since you used to quote him a lot before… I keep quoting Stanard, but you stopped acknowledging his words for some reason ;).

It is quite clear, but it is not an insult to XVA if we do not add another 2-3x multiplier out of thin air and the end result (in pure theoretical peak numbers, reality is what it is and what you can see in exclusives vs multi platform titles) is not as close to PS5’s SSD as you seem to think (especially considering most third parties will take a while to properly make use of either solution… aside from generous offers from the console maker for timed exclusives perhaps).
What’s the argument here anyways? That ps5 io is faster? I assume most assume that is true. So is this just an argument of how much?
 

Dodkrake

Banned
I mentioned Control specifically in response to that post. I'll explain the logic...

The PS5 SSD is much faster? so if its loading assets at the same speed as Series X then it is constrained by something else i.e. CPU

Both consoles have a very similar CPU hence when the Load times are very close it is logical that the similarity is caused by the CPU decompressing the textures instead of some fancy HW decoder.

Stay with me... Now.. if it is using Oodle/Kraken and its taking the same time to load a game what's the fucking point in the Uber SSD...

Can't have it both ways..

I believe that the large difference in overall game size we have seen in some cross format titles is down to that fact that Xbox shares a SDK with PC and PS5 requires a specific new build with a new asset pipeline. Note that the supposed compression % don't add up to the difference in game sizes.

Will PS5 have smaller game sizes through this gen than Series probably yes... will the difference be as big as it is with some titles now ...probably not.

I guess the other probability could be its using Kraken but not Oodle but then like explained before the compression ratios are off.

Loading speeds are limited by the engine as well.

Oodle and kraken are not separate things. Oodle is the product family, kraken is the compressor.

Your post just shows a huge amount of ignorance towards this subject. Stop
 
...stop focusing on the best case scenario....

Exactly.

Once again...this is why I asked whats the average and best case scenarios for Series consoles.

We have this for PS5, I dont think we have this for Series consoles.

If SFS works the way Microsoft says it does in AAA releases, the best case for Series X is upwards of 22GB/s effective I/O performance and literally putting on screen the visual equivalent of 22GB worth of textures, but getting it to fit into an 8.8GB package.

But that sounds extreme, so lets go with 15GB worth of textures, which fits into a 6GB package. 15GB/s effective, 18GB/s effective, that's all possible for series X.

The only limiting factor is we know a game will never be 100% textures and nothing else, so we know 33.75GB/s effective ain't happening. But if SFS works as advertised, anywhere from 12GB/s effective well into 18-20GB/s territory is easily possible for Xbox devs using it.

Average for SFS if I'm guessing is 6GB/s-15GB/s is my personal assumption, but they can clearly go higher.
 
Last edited:

Lethal01

Member
What’s the argument here anyways? That ps5 io is faster? I assume most assume that is true. So is this just an argument of how much?
The question is what kind of texture streaming setup the 2.5x multiplier is being compared to, is it one that's already using prt? Because the way many speak of it, it sounds like "no sfs" means "no prt at all"
 
Last edited:

Heisenberg007

Gold Journalism
If SFS works the way Microsoft says it does in AAA releases, the best case for Series X is upwards of 22GB/s effective I/O performance and literally putting on screen the visual equivalent of 22GB worth of textures, but getting it to fit into an 8.8GB package.

But that sounds extreme, so lets go with 15GB worth of textures, which fits into a 6GB package. 15GB/s effective, 18GB/s effective, that's all possible for series X.

The only limiting factor is we know a game will never be 100% textures and nothing else, so we know 33.75GB/s effective ain't happening. But if SFS works as advertised, anywhere from 12GB/s effective well into 18-20GB/s territory is easily possible for Xbox devs using it.

Average for SFS if I'm guessing is 6GB/s-15GB/s is my personal assumption, but they can clearly go higher.
Is there any evidence for any of it? Of course we haven't seen anything in game. On top of that, even Microsoft didn't share these numbers, so I wonder where are these coming from.

To recall: it's 2.4 Gb/s (raw) and 4.8 Gb/s compressed. That's it. SFS does not increase the streaming speed beyond 4.8 Gb/s.
 

jroc74

Phone reception is more important to me than human rights
Yeah.... I read this:
What’s the argument here anyways? That ps5 io is faster? I assume most assume that is true. So is this just an argument of how much?

And then I read this:
If SFS works the way Microsoft says it does in AAA releases, the best case for Series X is upwards of 22GB/s effective I/O performance and literally putting on screen the visual equivalent of 22GB worth of textures, but getting it to fit into an 8.8GB package.

But that sounds extreme, so lets go with 15GB worth of textures, which fits into a 6GB package. 15GB/s effective, 18GB/s effective, that's all possible for series X.

The only limiting factor is we know a game will never be 100% textures and nothing else, so we know 33.75GB/s effective ain't happening. But if SFS works as advertised, anywhere from 12GB/s effective well into 18-20GB/s territory is easily possible for Xbox devs using it.

Average for SFS if I'm guessing is 6GB/s-15GB/s is my personal assumption, but they can clearly go higher.
And there's clearly a disconnect somewhere.

I get it, we are all guessing and assuming but still. I'm still working with 6GB/s-12GB/s. And not knowing which is the average and best case scenarios. And this is just based on what MS employees have said.
 
Last edited:
Is there any evidence for any of it? Of course we haven't seen anything in game. On top of that, even Microsoft didn't share these numbers, so I wonder where are these coming from.

To recall: it's 2.4 Gb/s (raw) and 4.8 Gb/s compressed. That's it. SFS does not increase the streaming speed beyond 4.8 Gb/s.

Hence the use of the word "effective"

If you request 12GB of texture data for your game, but SFS makes it so that to get the same visual result on screen, you only need 2.5x less than that 12GB, which leaves you with only 4.8GB of data to load due to efficiency of SFS, what would you call such speed?

The SSD speed isn't changing from 2.4GB/s and neither is the compressed speed changing from 4.8GB/s, but you know what has changed? You've gotten 2.5x more effective work done in way less time, it's on screen, and it only took Series X 1 second to do it with SFS.

Microsoft actually HAS said this, and has been saying this from the start, but nobody seemed to take them serious when they said it. They've said from the start Sampler Feedback Streaming gives us a multiplier effect on our SSD bandwidth and I/O.

This guy below works for Microsoft's Advanced Technology Group. He's Graphics R&D at Microsoft and is personally working on Xbox's Texture Compression stuff and other graphics technologies, such as Sampler Feedback Streaming. So he personally has said this.



Here are more examples from Microsoft. Jason Ronald said this last year July.


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.

The coalition technical director said this.

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.

Here is Jason Ronald saying it's a massive boost to I/O in the inside Xbox Series S video revealing Series S. What do we think he means when he says "massive leap in IO and memory efficiency"






Here again they said it as early as March of Last year. They've always been saying it, but people were so caught up in how impressive Sony's solution is and how much faster their SSD is compared to the Series X SSD that people laughed off Microsoft talking about Sampler Feedback Streaming as just marketing or PR, or "power of the cloud" talk. But these are the engineers and the technical people who have been saying this stuff now and shouting it from the roof tops, not just the PR folks.



GTxqvTM.jpg
 
Last edited:

Panajev2001a

GAF's Pleasant Genius
Hence the use of the word "effective"

If you request 12GB of texture data for your game, but SFS makes it so that to get the same visual result on screen, you only need 2.5x less than that 12GB, which leaves you with only 4.8GB of data to load due to efficiency of SFS, what would you call such speed?

The SSD speed isn't changing from 2.4GB/s and neither is the compressed speed changing from 4.8GB/s, but you know what has changed? You've gotten 2.5x more effective work done in way less time, it's on screen, and it only took Series X 1 second to do it with SFS.

Microsoft actually HAS said this, and has been saying this from the start, but nobody seemed to take them serious when they said it. They've said from the start Sampler Feedback Streaming gives us a multiplier effect on our SSD bandwidth and I/O.

This guy below works for Microsoft's Advanced Technology Group. He's Graphics R&D at Microsoft and is personally working on Xbox's Texture Compression stuff and other graphics technologies, such as Sampler Feedback Streaming. So he personally has said this.



Here are more examples from Microsoft. Jason Ronald said this last year July.




The coalition technical director said this.



Here is Jason Ronald saying it's a massive boost to I/O in the inside Xbox Series S video revealing Series S. What do we think he means when he says "massive leap in IO and memory efficiency"






Here again they said it as early as March of Last year. They've always been saying it, but people were so caught up in how impressive Sony's solution is and how much faster their SSD is compared to the Series X SSD that people laughed off Microsoft talking about Sampler Feedback Streaming as just marketing or PR, or "power of the cloud" talk. But these are the engineers and the technical people who have been saying this stuff now and shouting it from the roof tops, not just the PR folks.



GTxqvTM.jpg


I believe in SFS and XVA, they are good advancements and will allow XSX to run circles around the XOX if devs target it fully despite RAM going “only” from 12 GB to 16 GB. Which is a tiny increase… yet you will see much better use of memory.

You can keep posting quotes as if people were saying XVA and SFS are mere buzzwords, but that would be an argument nobody is making.
What people are saying and your quotes agree os that the 2.5-3x multiplier is SFS compared to regular full texture data upload and not compared to PRT (PRT = storing only portion of a texture and not the entire texture in RAM).
 
Last edited:
I believe in SFS and XVA, they are good advancements and will slow XSX to run circles around the XOX if devs target it fully despite RAM going “only” from 12 GB to 16 GB. Which is a tiny increase… yet you will see much better use of memory.

You can keep posting quotes as if people were saying XVA and SFS are mere buzzwords, but that would be an argument nobody is making.
What people are saying and your quotes agree os that the 2.5-3x multiplier is SFS compared to regular full texture data upload and not compared to PRT (PRT = storing only portion of a texture and not the entire texture in RAM).

Some actually have been dismissing it as a mere buzzword, though. See it all the time. Now, what's totally fair, and I agree with them all on this, is that Microsoft must prove it with the games.

PRT does equal storing only portion of a texture and not the entire texture in RAM, but according to James Stanard, Graphics R&D at Microsoft, PRT doesn't sample missing texture regions. One of his best tweets on the subject would likely be this one here. Sampler Feedback Streaming helps make PRT a whole lot more precise. You have exactly what you need in memory, and exactly what you don't need not in memory. So even if you're storing parts of textures in memory with PRT, it still suffers from inaccuracies. Sampler Feedback makes the whole promise more accurate that you end up being smarter with what you use memory on.




And if you're wasting less memory on what you don't need, and using more on what you do need, your memory efficiency has improved.
 

Panajev2001a

GAF's Pleasant Genius
Some actually have been dismissing it as a mere buzzword, though. See it all the time. Now, what's totally fair, and I agree with them all on this, is that Microsoft must prove it with the games.

PRT does equal storing only portion of a texture and not the entire texture in RAM, but according to James Stanard, Graphics R&D at Microsoft, PRT doesn't sample missing texture regions. One of his best tweets on the subject would likely be this one here. Sampler Feedback Streaming helps make PRT a whole lot more precise. You have exactly what you need in memory, and exactly what you don't need not in memory. So even if you're storing parts of textures in memory with PRT, it still suffers from inaccuracies. Sampler Feedback makes the whole promise more accurate that you end up being smarter with what you use memory on.




And if you're wasting less memory on what you don't need, and using more on what you do need, your memory efficiency has improved.

That is purely dependent on how you as a developer implement the logic to keep track of what is visible and what is not and use it to inform your pre-fetching strategy: the HW is not keeping track of it for you and helping you to accurately prefetch what you need. “SF is getting feedback from the GPU that you tried to sample those regions”…

Some developers might stream a few more memory pages than needed and some may do a better job at it. That might explain how for some developers you may get some overhead and lower the effective multiplier PRT brings, but it is not 150-250% higher.

2t2PZ9s.jpg

9Ajga0Z.jpg
 
Last edited:
  • Like
Reactions: Rea

Corndog

Banned
I believe in SFS and XVA, they are good advancements and will allow XSX to run circles around the XOX if devs target it fully despite RAM going “only” from 12 GB to 16 GB. Which is a tiny increase… yet you will see much better use of memory.

You can keep posting quotes as if people were saying XVA and SFS are mere buzzwords, but that would be an argument nobody is making.
What people are saying and your quotes agree os that the 2.5-3x multiplier is SFS compared to regular full texture data upload and not compared to PRT (PRT = storing only portion of a texture and not the entire texture in RAM).
Maybe some can ask one of the developers what they meant instead of us guessing.
 

DJ12

Member
To recall: it's 2.4 Gb/s (raw) and 4.8 Gb/s compressed. That's it. SFS does not increase the streaming speed beyond 4.8 Gb/s.
Wow, with compression, it's almost as fast as PS5 raw throughput.

Why is this thread still going, some desperate folks around here trying to polish a turd.
 

Panajev2001a

GAF's Pleasant Genius
Maybe some can ask one of the developers what they meant instead of us guessing.
People did ask him (and devs here). Stanard states it in plain English I think: I have not seen anybody claiming that SFS grants you a 2-3x efficiency advantage on top of best in class PRT based virtual texturing solutions. Stanard’s comments fully qualifies the early 2-3x multiplier:
2t2PZ9s.jpg

9Ajga0Z.jpg

vkkodl1.jpg
 
Last edited:
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.

“Might be”. By this reasoning it “might not be”. Or it might be worse.
 
Wow, with compression, it's almost as fast as PS5 raw throughput.

Why is this thread still going, some desperate folks around here trying to polish a turd.

It's still going because of RAM efficiency. If you have a means of putting the same thing up on screen, but actually being able to do so while using 2.5x less RAM and you have the ability to replace textures no longer seen and replace them with what's going to be on screen next fast enough and accurately enough, you can save yourself a lot of system memory.

If you don't need to stream nearly as much texture data into RAM, but you can get the exact same results by only streaming in 2.5x less data, that effectively is a big boost on your I/O and capabilities. The SSD isn't the reason for it, because it's still 2.4GB/s raw 4.8GB/s compressed, SFS alone isn't the reason for it, because if series x has an HDD instead of an SSD it's way less effective, and not as efficient in doing its work due to the slower drive speed.

It's a collective of their DirectStorage API, Sampler Feedback Streaming, the Decompression Unit and BCPack, and the SSD. All of that is what comes together to produce this outcome. No one thing may seem extraordinary, but together they make one kick ass architecture.
 
Last edited:
All good thanks for your reply!
Wow, with compression, it's almost as fast as PS5 raw throughput.

Why is this thread still going, some desperate folks around here trying to polish a turd.
Yes and also some people completely forget that prt+ is a thing, which is prt and sf combined, which accidentally is exactly what sfs is, except anyone is free to use it. Difference is that it’s not built in the hardware. So if you use the sfs multiplier for the xsx with some mental gymnastics you should do the same to ps5. So lets see, 17GB/s times 3 equals 51GB/s. Or no lets take the 22GB/s and times 3 equals 66gb/s. Oh my, look at that. The gap only widens in ps5’s favor.

Magic GIF by The Paley Center for Media
 
Last edited:
That is purely dependent on how you as a developer implement the logic to keep track of what is visible and what is not and use it to inform your pre-fetching strategy: the HW is not keeping track of it for you and helping you to accurately prefetch what you need. “SF is getting feedback from the GPU that you tried to sample those regions”…

Some developers might stream a few more memory pages than needed and some may do a better job at it. That might explain how for some developers you may get some overhead and lower the effective multiplier PRT brings, but it is not 150-250% higher.

This is definitely not accurate. The hardware almost certainly helps better inform the streaming system in the game. That's the whole point of Sampler Feedback. To say the hardware doesn't actually help you accurately prefetch what you need goes against everything Microsoft has been saying about the feature. What it's different from what you expected? Because it does this. They even show it in the real-time demo telling the streaming system what's next.


What mip level did it ultimately sample? Seems like a very basic question. In a world before Sampler Feedback there’s no easy way to know. You could cobble together a heuristic. You can get to thinking about the sampling pattern, and make some educated guesses. But 1) You don’t have time for that, and 2) there’s no way it’d be 100% reliable.

Where exactly in the resource did it sample?
More specifically, what you really need to know is— which tiles? Could be in the top left corner, or right in the middle of the texture. Your streaming system would really benefit from this so that you’d know which mips to load up next. Yeah while you could always use HLSL CheckAccessFullyMapped to determine yes/no did-a-sample-try-to-get-at-something-nonresident, it’s definitely not the right tool for the job.

Direct3D Sampler Feedback answers these powerful questions.

At times, the accuracy of sampling information is everything. In the screencap shown below, this demo-scene compares a “bad” feedback approximation to an accurate one. The bad feedback approximation loads higher-detailed mips than necessary:

Although this demo comparison is a bit silly, it confirms something you probably suspected: good judgments about what to load next can mean dramatic memory savings. And even if you’re using a partial-mip-chain-based system, accurate sampler feedback can still allow you to make better judgments about what to load and when.
 

Panajev2001a

GAF's Pleasant Genius
This is definitely not accurate. The hardware almost certainly helps better inform the streaming system in the game. That's the whole point of Sampler Feedback. To say the hardware doesn't actually help you accurately prefetch what you need goes against everything
I was taking about PRT and addressing your comment about PRT being (potentially) less efficient than SFS and why I was agreeing with you (with some qualifications which I pointed out)!!! Feedback from the GPU is the bloody point of Sampler Feedback (streaming). I was saying the exact opposite of what you read :p.
 
Last edited:
Yes and also some people completely forget that prt+ is a thing, which is prt and sf combined, which accidentally is exactly what sfs is, except anyone is free to use it. Difference is that it’s not built in the hardware. So if you use the sfs multiplier for the xsx with some mental gymnastics you should do the same to ps5. So lets see, 17GB/s times 3 equals 51GB/s. Or no lets take the 22GB/s and times 3 equals 66gb/s. Oh my, look at that. The gap only widens in ps5’s favor.

Magic GIF by The Paley Center for Media

You're missing the entire point. I don't care how it compares to PS5, and most shouldn't either. I care about what it will do for Xbox Series X games. If it you makes you feel better, yes, if PS5 is using this feature it will always have the faster I/O.

Happy Jonah Hill GIF
 
I don't care how it compares to PS5, and most shouldn't either. I care about what it will do for Xbox Series X games.
It will definitely help series x games, I think the multiplier will be quite a bit lower than x3 in real usage scenarios but I’m sure the difference will be quite noticeable vs not using the tech. This generation of consoles really bring a lot of new and exciting stuff to the table, especially if you look vs last gen which was a weak hardware upgrade from the one before.
 
Top Bottom