Next-Gen PS5 & XSX |OT| Console tEch threaD

Status
Not open for further replies.
And that proves.... What?
Instead of having to use more bandwidth to the SSD for assets, the Sampler Feedback Technique intelligently allows the GPU to grab assets when needed instead which will utilize less memory in the long run. This doesn't mean that the PS5 won't have the advantage as far as SSD streaming goes, but it allows the XSX to not be that far behind. I think the SSD on the PS5 was Sony's brute force attempt for great looking games and the 12TF with the Velocity architecture was Microsoft's brute force attempt for great looking games. I'm starting to think the differences will be negligible and there won't be a gigantic difference in the way games look, but I feel the XSX will have the slight advantage when it comes to additional effects, physics, and resolution but it wont make or break anything. Marketing and games are where things will really matter.
 
someone else here explained this quite good, with ms and Sony approaching the same issue (IO bottlenecks) a bit differently, one with hardware/software mix more focused on software (ms), and the other with hardware/software mix more focused on hardware (Sony). We need to wait and see which one works out better, might as well end with the difference in real world scenarios being negligible.

That's just not how any of this works in reality, though. Both teams would have profiled the entire IO stack from NAND flash to working cache and made software changes where possible, and if they still weren't hitting their target, moved parts of that software stack to fixed function accelerators.

The hardware and software aren't separate things. They aren't separate ways of attacking the same problem. That's now how it works. There are no separate hardware and software bottlenecks, the entire pipeline is profiled with telemetry.

That Sony developed the Coherency Engine and cache scrubbers to get more utilisation out of their CUs and edge closer to that theoretical 10.2TF maximum means they've thought about every single aspect of the IO pipeline as it's actually used in game. There is no software solution to selectively scrubbing a GPU cache. There is no software solution to keeping a cache and RAM version of the same data coherent.
They've basically eliminated a cache miss and refetch from RAM that is common to CPU and GPU processing which when it occurs causes the CU to basically stop working until the (relatively) much slower RAM has been able to refill the working cache again.
At the other end of the pipeline they've even got to the trouble of creating their own flash controller outside of and superior to the standard NVMe spec.
These aren't things you just add to a pipeline, but things you offload once the entire pipeline—hardware and software—has been analysed to see where the limits of software running on a generic programmable processor are hit and where you need to move the software logic to an ASIC accelerator.
 
Instead of having to use more bandwidth to the SSD for assets, the Sampler Feedback Technique intelligently allows the GPU to grab assets when needed instead which will utilize less memory in the long run. This doesn't mean that the PS5 won't have the advantage as far as SSD streaming goes, but it allows the XSX to not be that far behind. I think the SSD on the PS5 was Sony's brute force attempt for great looking games and the 12TF with the Velocity architecture was Microsoft's brute force attempt for great looking games. I'm starting to think the differences will be negligible and there won't be a gigantic difference in the way games look, but I feel the XSX will have the slight advantage when it comes to additional effects, physics, and resolution but it wont make or break anything. Marketing and games are where things will really matter.
So all Sony has to do is create an API that intelligently allows the GPU to perform better than the 10 tflops it has. I mean if it works for MS when it comes to close the 120% gap in ssd speed, it should be even easier for Sony to cut down the 18% gap in tflops with some intelligent and yet magical API.
 
And I would say sure, why not. I think some of the ideas in the XsX IO system are going to make it to PCs soon as well. I hope so in fact!

Why is it so incredible that a hardware feature that does not make sense to adopt in desktop cards right now (not talking about packed math, which by the way, is the GPU (analogous to vectorization in CPUs, so yes, ancient) eventually makes its way there once AMD decides to make it part of their roadmap?

So, you agree how utterly stupid it is to call people delusional for musing that possibility (and something that would happen years down the line anyway). Even people openly speculating about that long before RTG mentioned this idea from his conversation got attacked for that in this forum.


Everything is possible but it's best when base of reasoning is not third hand infos without clear responses/data.
Which feature will come to pc ?
i doubt it's complex i/o because it's a complete pipeline
geometry shader and RT are RDNA2.
is it tempest engine ? that seems not be the feature he was referring to.
So it's a major hidden feature ? why hide this feature and not the others ?
this make no sense
 
Last edited:
WHY WON'T ANYONE THINK OF THE KIDS!!!
x7L2VSNEiyAB5Ux7nxKmLo6yLyEJT6Jt5yhNCUpGMctTt1z4eTP2b1YqfVbk3FumGKBgaVgfoqmNjTK
 
Instead of having to use more bandwidth to the SSD for assets, the Sampler Feedback Technique intelligently allows the GPU to grab assets when needed instead which will utilize less memory in the long run. This doesn't mean that the PS5 won't have the advantage as far as SSD streaming goes, but it allows the XSX to not be that far behind. I think the SSD on the PS5 was Sony's brute force attempt for great looking games and the 12TF with the Velocity architecture was Microsoft's brute force attempt for great looking games. I'm starting to think the differences will be negligible and there won't be a gigantic difference in the way games look, but I feel the XSX will have the slight advantage when it comes to additional effects, physics, and resolution but it wont make or break anything. Marketing and games are where things will really matter.
Good to see that someone understands.
And I will emphasize that the XSX does the texture filtering (i.e. avoiding the loading of unnecessary textures) through hardware;

 
Instead of having to use more bandwidth to the SSD for assets, the Sampler Feedback Technique intelligently allows the GPU to grab assets when needed instead which will utilize less memory in the long run. This doesn't mean that the PS5 won't have the advantage as far as SSD streaming goes, but it allows the XSX to not be that far behind. I think the SSD on the PS5 was Sony's brute force attempt for great looking games and the 12TF with the Velocity architecture was Microsoft's brute force attempt for great looking games. I'm starting to think the differences will be negligible and there won't be a gigantic difference in the way games look, but I feel the XSX will have the slight advantage when it comes to additional effects, physics, and resolution but it wont make or break anything. Marketing and games are where things will really matter.

Sampler Feedback for Streaming is some hardware filters on top of Sampler Feedback to make it more efficient. The 2x 3x multiplier part people keep quoting is at part of Sampler Feedback that isn't unique to Microsoft. It's basic tiling to reduce how much data is fetched, and should be compared against systems that don't do that. The optimisations Microsoft have made with Sample Feedback for Streaming don't add 2x-3x multipliers on top of Sampler Feedback. It makes Sampler Feedback itself more efficient for streaming in terms of latency and being ready in time, not in bandwidth reduction.
 
Everything is possible but it's best when base of reasoning is not third hand infos without clear responses/data.
Which feature will come to pc ?
i doubt it's complex i/o because it's a complete pipeline
geometry shader and RT are RDNA2.
is it tempest engine ? that seems not be the feature he was referring to.
So it's a major hidden feature ? why hide this feature and not the others ?
this make no sense
The short answer is, it would not be a major feature. Neither Sony nor MS advertised every single hardware feature to consumers.
 
Good to see that someone understands.
And I will emphasize that the XSX does the texture filtering (i.e. avoiding the loading of unnecessary textures) through hardware;



Yes it's offloading Sampler Feedback to hardware to help with latency, which is the "for Streaming" part of it. It doesn't reduce SSD bandwidth requirements by 2x-3x. That's what normal tiling or Sampler Feedback does. The "for streaming" customisations are being touted by some as a bandwidth multiplier when they are latency reduction, as well as a CPU offload.
 
Last edited:
So all Sony has to do is create an API that intelligently allows the GPU to perform better than the 10 tflops it has. I mean if it works for MS when it comes to close the 120% gap in ssd speed, it should be even easier for Sony to cut down the 18% gap in tflops with some intelligent and yet magical API.
Yeah, all they have to do is download more RAM.
 
Yes it's offloading Sampler Feedback to hardware to help with latency, which is the "for Streaming" part of it. It doesn't reduce SSD bandwidth requirements by 2x-3x. That's what normal tiling or Sampler Feedback does. The "for streaming" customisations are being touted by some as a bandwidth multiplier when they are latency reduction, as well as a CPU offload.
That's assuming that sampler feedback was working optimally in the first place. Was sampler feedback being used in the Xbox One X?
 
Last edited:
And that proves.... What?

It proves it's an evolution of a concept already known, hardware supported since GCN (PRT) and what MS is describing might very well be something that is RDNA2 and not a custom block specific to XSX. It's the new VRS. Also MS is describing technology as it relates to previous gen, not the competition. But I'll let you do you.
 
Last edited:
Sorry if this was already posted in here, I may have missed it

Looks like Bradly Halestorm had more comments for Gamingbolt, he had made the "Tempest Audio is most exciting new feature" comment a few days ago, I believe:


On Variable Frequency:

As Halestorm notes, "Variable frequency is an interesting discussion in the console world, primarily because it's never been a discussion in the console world. Sony's stated that the PS5 will alter CPU and GPU frequency in way that allows developers not to have to worry about it, which is what you would say if you were trying to court developers and also put them at ease.

"Microsoft is taking the opposite approach; they're touting the Series X's fixed frequencies. It's fascinating to see how the two companies are going about nudging developers into the coming gen."

On MS Cross-Gen Support:

"I don't mean to punt on this answer as well, but I'm not sure. The cross-gen stuff is interesting and is a very consumer-first strategy.

"I do think, however, that the further we get into this next console cycle – and thereby the further we get away from the current one – the more we'll see if there truly is an advantage in that cross-gen policy.

"Right now, it's smart to promote cross-gen within your own family of consoles; after all, not everyone will upgrade to a PS5 or XSX this November or even next year. But 4 years down the road, will developers still be as willing to allocate resources to ensuring that their game runs on a console from the previous generation that is, at that point, nearly half a decade old? I'm not so certain."
 
OK, let's forget about the fly-through thing (even though it is a strong one but this evidence isn't enough of a confirmation for you).

There were some other cut-frames, like the one when it revealed 500 statues, what's up with that? They are surely much more demanding on the system than a cut-frame to a wall, yet they didn't have any "pop-in", come on man, you can't get out of this. 😂🤣🤣🤣

I'm talking about this one. 🤣🤣🤣🤣


nREGS6A.jpg

Did you know that all those 500 statues are using the same textures?
and that anything not in view and that means the back space of an object is culled and not render?
 
It proves it's an evolution of a concept already known, hardware supported since GCN (PRT) and what MS is describing might very well be something that is RDNA2 and not a custom block specific to XSX. It's the new VRS. Also MS is describing technology as it relates to previous gen, not the competition. But I'll let you do you.
And why do you come to that conclusion? The Eurogamer article clearly stated that on the Xbox One X, memory utilization efficiency is low. Considering that games have been using PRT for a while, it's quite safe to assume, that what they discovered on the Xbox One X, was with PRT. Not without.
When you use the sampler feedback to then load the required textures, rather than pre-loading partial textures, you get the advertised 2x - 3x efficiency increase.
As far as I'm aware, at least.
 
Did you know that all those 500 statues are using the same textures?
and that anything not in view and that means the back space of an object is culled and not render?
Yet it is rendering the same amount of polygons whether you are looking at them or that wall, and even if we used your logic, the overall screen-space is still viewing all the textures and polygon counts will exceed that wall.
 
Because you're parroting a marketing bullet point about, and even worse you're expanding it beyond its use. It's applied to textures, therefore it's not an effective overall 2x or 3x multiplier in bandwidth and RAM, you keep spreading BS. And MS is so vague about this, that it's not even possible to understand just how different it is from partially resident textures that has been hardware supported since GCN.


"Use of sampler feedback with streaming is sometimes abbreviated as SFS. It is also sometimes called sparse feedback textures, or SFT, or PRT+, which stands for "partially resident textures".

Is nvidia also lying?

"Sampler Feedback shares the same philosophy as Variable Rate Shading: work smarter to reduce GPU load, to improve performance. Sampler Feedback improves computing between the texture and shader hardware. Developers and engines can use Sampler Feedback to implement sophisticated texture streaming algorithms, allowing large open worlds with beautiful high-resolution textures, without immersion-breaking stutter or texture pop-in. Sampler Feedback can also be used to implement texture space shading, for more efficient rendering."

 
And why do you come to that conclusion? The Eurogamer article clearly stated that on the Xbox One X, memory utilization efficiency is low. Considering that games have been using PRT for a while, it's quite safe to assume, that what they discovered on the Xbox One X, was with PRT. Not without.
When you use the sampler feedback to then load the required textures, rather than pre-loading partial textures, you get the advertised 2x - 3x efficiency increase.
As far as I'm aware, at least.

The 2x-3x bandwidth multiplier being touted doesn't relate to an advantage over just-in-time streaming versus keeping them resident in RAM, but fine grained control over how much of that texture needs to be streamed into RAM in the first instance, and intelligently fetching more as required. This isn't new tech, but offloading part of that logic to a hardware accelerator would help with latency (required for just-in-time streaming) and to a lesser extent CPU utilisation.

SFS isn't helping Microsoft stream in textures at an effective 2x-3x 4.8GB/s, and it's not even like more typical Sampler Feedback is some fixed speed metric anyway. In a lot of cases it won't be applicable, in others it will be much more marginal etc.
I'm not sure about Xbox One, but PS4 hasn't ever needed to pull in an entire texture file if it only needed some tiles from it.
The video linked here showed Gears of War struggling with texture streaming, though?


Is nvidia also lying?

"Sampler Feedback shares the same philosophy as Variable Rate Shading: work smarter to reduce GPU load, to improve performance. Sampler Feedback improves computing between the texture and shader hardware. Developers and engines can use Sampler Feedback to implement sophisticated texture streaming algorithms, allowing large open worlds with beautiful high-resolution textures, without immersion-breaking stutter or texture pop-in. Sampler Feedback can also be used to implement texture space shading, for more efficient rendering."


Sampler Feedback != Sampler Feedback for Streaming
 
Last edited:
Replying to this specifically....



Until devs chime in, I see a play of words. "Velocity Architecture is all custom to Xbox ". When pushed to say if there's anything preventing others, he dodges.

Again, sources other than MS ambiguous answers and then we will be talking.
 
Instead of having to use more bandwidth to the SSD for assets, the Sampler Feedback Technique intelligently allows the GPU to grab assets when needed instead which will utilize less memory in the long run. This doesn't mean that the PS5 won't have the advantage as far as SSD streaming goes, but it allows the XSX to not be that far behind. I think the SSD on the PS5 was Sony's brute force attempt for great looking games and the 12TF with the Velocity architecture was Microsoft's brute force attempt for great looking games. I'm starting to think the differences will be negligible and there won't be a gigantic difference in the way games look, but I feel the XSX will have the slight advantage when it comes to additional effects, physics, and resolution but it wont make or break anything. Marketing and games are where things will really matter.

Theres only a tiny thing against this post and that's, well... that devs are completely disagreeing.

We have Matt from Resetera saying io on PS5 vs Xbox difference is significant.

We have this from a third party dev..

It is easy. It is useless to have 12 boxes if they do not fit through the door all together.

You have 12 boxes to fill. So you can't pass all the boxes at once. You must decide which boxes will pass and which will not. That is handled by a coordinator. And the coordinator tells the delivery man which boxes to take.

Mrs. XSX wants to make the move as soon as possible, but it turns out that only 8 boxes can fit on the door at a time. The coordinator is fast, and also uses a box compressor so that 10 boxes can go through instead of 8, but there are several drawbacks. The compressor can only compress the red boxes, and the coordinator also has to coordinate many other things, street traffic, people passing through the door, the space in the room where the boxes are stored, the noise of neighbors who distract the delivery man, search and select what the boxes are filled with, etc. Also, the delivery man is not so fast and is very distracted filling and transporting boxes. So it passes the 10 boxes (not 12) at a certain speed "1x". The lady demands that the boxes arrive, but they do not arrive as quickly as the lady would like, since although she has many boxes, the system is not capable of managing all of them properly.

On the other hand we have Mrs. PS5. You only have 10 boxes to fill. But its door is twice as big, enough for all its boxes to enter at once and there is room for people to also enter and exit through the door. Furthermore, the coordinator has the ability to automatically discard unnecessary boxes, so he doesn't waste time checking boxes that are not going to be used. In addition, anyone in the environment can do the job of the coordinator or the delivery man (even at the same time). The compressor is not that new, but it can compress all boxes, whether they are red or blue. All. And the delivery man is more than twice as fast and manages to pass the boxes at the speed of "2.5x" in the worst case, and "5x" on many occasions. In addition, if someone is left free or without work, they can help to distribute boxes with the delivery man or coordinate work with the coordinator. All this makes this removal company the most efficient ever seen and that the number of boxes available is irrelevant. For that moving system, 12 boxes are not needed, with 10 you can do the same job (and more or better in some cases). Having more boxes would only make the price of the move more expensive without needing any of it.

Of course, having more boxes available always helps to advertise yourself as a top removal company compared to the competition, even if your removal company is normal and ordinary. But it is only that, a smokescreen.

That does not mean that XSX is bad, far from it, it is an extraordinary machine. But PS5 has an efficiency NEVER seen before.

It is true that on PC there are more powerful cards or more powerful systems, but you know that these cards are never used properly, they draw raw power, but they are never used. It is the scourge of PC, an ecosystem that is too varied and unusable. In addition to exorbitant prices.

And I've always been a PCLover, but things as they are, what I've seen on PS5 I only remember something similar when 3DFX and its Glide came out. Its astonishing speed leaves you speechless

We also have this from Moores law is dead and his sources... he says the difference won't be negligible...


33E5uWr.jpg
 
Does anybody know what is the correct frequency of the XSeX CPU with SMT enabled, is it 3.6 or 3.66GHz? You can find both numbers depending on the website you check, . Microsoft´s rounding of numbers does not help either (e.g. 12TF vs 12.155TF, etc...).
 
When pushed to say if there's anything preventing others, he dodges.

I think alot of what makes up the Velocity Architecture is also part of the PS5s I/O system. Sony may call their features by a different name but in the end it's the same thing across both platforms.

Both probably have proprietary technology that achieve similar results in a different way.

Obviously one I/O is going to better than the other since they have different hardware/software in it.
 
Theres only a tiny thing against this post and that's, well... that devs are completely disagreeing.

We have Matt from Resetera saying io on PS5 vs Xbox difference is significant.

We have this from a third party dev..



We also have this from Moores law is dead and his sources... he says the difference won't be negligible...


33E5uWr.jpg

Pretty damning if true. Especially the part where he states that the PS5s I/O solution is a generation ahead of the Xboxs. Microsoft shouldn't let this happen due to how big the I/O delta will be if true.
 
Pretty damning if true. Especially the part where he states that the PS5s I/O solution is a generation ahead of the Xboxs. Microsoft shouldn't let this happen due to how big the I/O delta will be if true.
They are just prioritizing resolution and effects over pop-in. It's a legitimate call either way, I wouldn't call it a mistake.
 
Last edited:
They are just prioritizing resolution and effects over pop-in. It's a legitimate call either way, I wouldn't call it a mistake.

The difference seems pretty big though compared to the GPU delta. There's a part where he claims that Microsoft was going to go with an even slower SSD until they heard about Sonys plan.
 
When it's all assembled and tested Sony got 5.5gb/s and MS got 2.4gb/s. Maybe both companies expected the performance for themselves, but I think MS got wind of Sony's claimed figures and concluded this might be a marketing problem.
I have heard the highlighted part quite a few times in last few days

So would it not be considered as corporate espionage?
 
I think alot of what makes up the Velocity Architecture is also part of the PS5s I/O system. Sony may call their features by a different name but in the end it's the same thing across both platforms.

Both probably have proprietary technology that achieve similar results in a different way.

Obviously one I/O is going to better than the other since they have different hardware/software in it.

I tend to agree that they will be fundamentally similar. It's just curious to see the difference in processing's power dedicated to that though. XSX uses 1/10th of a core while PS5 has a dedicated block that's equivalent to 1-2 cores. I guess Sony is anticipating handling magnitudes more I/O ops in next gen games.
 
Pretty damning if true. Especially the part where he states that the PS5s I/O solution is a generation ahead of the Xboxs. Microsoft shouldn't let this happen due to how big the I/O delta will be if true.

I don't think Microsoft has been at all negligent. They probably just envisaged flash memory as being about load times and less pop-in. Just like everyone else did before Sweeney unzipped his trousers.
Sony and Epic Games have been cooking this new paradigm up for a while, and it's probably no coincidence it wasn't publicly revealed until this close to release.

Sony have potentially played a blinder by partnering with pretty much the premier third-party engine like this. I wonder if Microsoft was also approached to work on future rendering technologies, and whether they'd have been as interested in investing the same resources as Sony have done.
 
The 100GB number is just referring to the game installation. There is no artificially imposed limit. The entire game data is quickly accessible, without the seek times by virtue of being installed in an SSD.

Well let's look at the actual text shall we?

"Enter Xbox Velocity Architecture, which features tight integration between hardware and software and is a revolutionary new architecture optimized for streaming of in game assets. This will unlock new capabilities that have never been seen before in console development, allowing 100 GB of game assets to be instantly accessible by the developer. "

Now "100gb of game assets [...] instantly accessible" is clearly suggesting there is something about that 100gb capacity that is significant.


The wording of it "instantly" accessible suggests to me that they are referring to a different access paradigm to this 100gb than the rest of the SSD. Otherwise why not say 1tb of game assets instantly accessible the same as Sony has?


But let's say for a moment you are right and the 100gb refers to a game install - then this would imply there is a 100gb install limit for games. That seems .... rather small.


Considering the only number MS mentioned in reference to their SSD storage was this 100gb capacity (they didn't mention access speeds or other capabilities for example) its rather strange that they were unable to convey what this 100gb number refers to.

I'm confident they wouldn't refer to a 100gb limit on game installs in a marketing piece, in fact I'm confident that no firmware architect would allow such a limitation.


EDIT: And there is this in the Eurogamer analysis that confirms my interpretation that the 100gb is a direct reference block. Which raises the question what happens if the game data creeps over 100gb...

" the game package that sits on storage essentially becomes extended memory, allowing 100GB of game assets stored on the SSD to be instantly accessible by the developer."
 
Last edited:
Good to see that someone understands.
And I will emphasize that the XSX does the texture filtering (i.e. avoiding the loading of unnecessary textures) through hardware;


I swear I saw this screenshot for million times, I mean you guys don't have anything else to show to prove your point?? 🤣🤣
 
Status
Not open for further replies.
Top Bottom