• 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.

Microsoft Game Stack VRS update (Series X|S) - Doom Eternal, Gears 5 and UE5 - 33% boost to Nanite Performance - cut deferred lighting time in half

Panajev2001a

GAF's Pleasant Genius
Its just talk like the above is pretty console warrish, its not related to the topic.
The first was a reply to you about a remark you made, the second one… fair, something better taken to a PM (something to Riky and not to the thread). I still think it is ironic and defensive as an attitude (show me the same circling of the wagons in pony threads, what is your line between HW talk and comparison with other HW?).
 

Sosokrates

Report me if I continue to console war
The first was a reply to you about a remark you made, the second one… fair, something better taken to a PM (something to Riky and not to the thread). I still think it is ironic and defensive as an attitude (show me the same circling of the wagons in pony threads, what is your line between HW talk and comparison with other HW?).

Its one thing to talk about hardware its another thing to talk like you did in the 2nd quote.
 

Sosokrates

Report me if I continue to console war
I agree it was more personal than console warrish, but point taken on that sentence.
It wasn't just that sentence

Your analysis also only sees XSX with tricks under its sleeves and the competition with nothing left to improve.
The ps5 has nothing to do with xbox's VRS tech, bringing it up like you is just bizarre.

Its like

fCvLBe6.gif



Its seem more of a personal thing you have with Riky.
 

Panajev2001a

GAF's Pleasant Genius
It wasn't just that sentence


The ps5 has nothing to do with xbox's VRS tech, bringing it up like you is just bizarre.

Its like

fCvLBe6.gif



Its seem more of a personal thing you have with Riky.
No, not purely around VRS in and of itself unless you look at impact with third party titles, but a sentence in a post replying to him was a bit more directed than general.
 

VFXVeteran

Banned
You seem unable to read, I quoted that because it directly says VRS Tier 2 and it came from the very same morning before this presentation was released where obviously Tier 2 VRS has not gone anywhere, why you're so hurt by this who knows.

Anyway back on topic,

Kmg0EDc.png





I don't think on its own its large enough but we're anticipating a series of features being used together that will help Series consoles improve over time, if the average is still around 15% on its own then that's not insignificant you must agree in development terms, any extra performance is welcome. Since UE5 will probably be the most used engine over the next decade then as the slide above shows and we're in early days it's all helpful.
I don't doubt UE5 will be the most used engine, but having said that developers will still tinker with it to the point where you get this same psuedo-random results between the two platforms.

The main thing to focus on is what does 15% improvement in performance buy you? Certainly not more RT as the delta difference of non-RT and RT is too big and thus wouldn't be a choice. More detail in textures? Nope because bandwidth is needed there not shader execution. Nanite? Not likely as it will need to be consistent across platforms. So I would say it would best be spent in more stable FPS. You MIGHT be able to sample better shadows or something like that but 15% isn't enough for any BIG apparently visual differences.
 

winjer

Gold Member
I don't doubt UE5 will be the most used engine, but having said that developers will still tinker with it to the point where you get this same psuedo-random results between the two platforms.

The main thing to focus on is what does 15% improvement in performance buy you? Certainly not more RT as the delta difference of non-RT and RT is too big and thus wouldn't be a choice. More detail in textures? Nope because bandwidth is needed there not shader execution. Nanite? Not likely as it will need to be consistent across platforms. So I would say it would best be spent in more stable FPS. You MIGHT be able to sample better shadows or something like that but 15% isn't enough for any BIG apparently visual differences.

15% might not seem important, but is always helpful. It can be the difference between maintaining a constant frame rate, or having small drops below 30 or 60 fps. If you don't have a VRR display and console/PC, this means screen tearing and judder.
Consider that consoles don´t have hardware for ray traversal, so t's done in shaders. This means 15% more shader units can be used for rays casting. And this means slightly better reflections, shadows or GI.
It also might mean higher resolution for the rest of the image. If we are spending less shaders on parts of the image that don't matter, then e can spend those shaders improving where it matters most. So it might be used to get 15% higher screen resolution.
It will always depend on where the devs want to spend that extra available shading power, but it will make things a bit better.

It's not a huge difference. It's only 15%. But it helps that one bit.
It's basically like getting one extra TFLOP out of these machines.
 

Three

Member
You seem unable to read, I quoted that because it directly says VRS Tier 2 and it came from the very same morning before this presentation was released where obviously Tier 2 VRS has not gone anywhere, why you're so hurt by this who knows.
You quoted that simply because it had the words "VRS tier 2" in it? Really?
Not only am I able to read but I'm able to read context too because I wasn't born yesterday.

You quoted that because you think you had a gotcha in a console warring argument you started and doubled down on in another thread discussing VRS closing a console war gap that didn't materialise.

I'm not hurt by your comments but it's the fact that you ignore the context of the discussion you yourself started then act like a victim of console warring when it's you perpatrating it.

You should be talking about the approach, about how this method could benefit all games/systems with a similar deffered pipeline, yet you're here talking about "xbox wins", xbox "exclusive circuitry", and quoting people in discussions you've had in the past where you stated xbox will smoke the competition due to these 'features'. It's because you don't even know what "Tier 2 VRS" means here. You don't even seem to grasp where the optimisation comes from.

Sosokrates Sosokrates Riky drags Panajev into this thread by quoting him regarding Rikys own past console war comments.
Panajev comes here and discusses what Riky quoted him about and you're asking panajev to go somewhere else because he's console warring? Panajev is even nice enough to apologise for it.

Riky thinks his console war comments are vindicated. Half the comments on the first page are referencing Rikys console war comments of the past. Riky is clearly more concerned about "xbox wins" instead of optimisation wins for the industry in this thread. There is a clear console war mentality that you're ignoring and you've got the wrong guy for
 
Last edited:

Riky

$MSFT
You quoted that simply because it had the words "VRS tier 2" in it? Really?
Not only am I able to read but I'm able to read context too because I wasn't born yesterday.

You quoted that because you think you had a gotcha in a console warring argument you started and doubled down on in another thread discussing VRS closing a console war gap that didn't materialise.

I'm not hurt by your comments but it's the fact that you ignore the context of the discussion you yourself started then act like a victim of console warring when it's you perpatrating it.

You should be talking about the approach, about how this method could benefit all games/systems with a similar deffered pipeline, yet you're here talking about "xbox wins", xbox "exclusive circuitry", and quoting people in discussions you've had in the past where you stated xbox will smoke the competition due to these 'features'. It's because you don't even know what "Tier 2 VRS" means here. You don't even seem to grasp where the optimisation comes from.

Sosokrates Sosokrates Riky drags Panajev into this thread by quoting him regarding Rikys own past console war comments.
Panajev comes here and discusses what Riky quoted him about and you're asking panajev to go somewhere else because he's console warring? Panajev is even nice enough to apologise for it.

Riky thinks his console war comments are vindicated. Half the comments on the first page are referencing Rikys console war comments of the past. Riky is clearly more concerned about "xbox wins" instead of optimisation wins for the industry in this thread. There is a clear console war mentality that you're ignoring and you've got the wrong guy for

You're obviously upset but you can't change history as it's all here to see and I've quoted it, up to the very day these presentations were made these people were trying to ridicule and downplay claiming it had come and gone and was just marketing terms.
You're obsession with console wars never comes with the quotes, maybe because it's all in your head.
Xbox "wins" are just ways of getting extra performance from fixed hardware, you can take it whatever way you like.
People can obviously see how much stick I've got for believing what MS said about increasing performance and the proof has arrived, they are just referencing that.
Anyway I'm done talking about it and unless you'e got something to say about the actual topic then there is no point talking about it anymore.
 

Sosokrates

Report me if I continue to console war
You quoted that simply because it had the words "VRS tier 2" in it? Really?
Not only am I able to read but I'm able to read context too because I wasn't born yesterday.

You quoted that because you think you had a gotcha in a console warring argument you started and doubled down on in another thread discussing VRS closing a console war gap that didn't materialise.

I'm not hurt by your comments but it's the fact that you ignore the context of the discussion you yourself started then act like a victim of console warring when it's you perpatrating it.

You should be talking about the approach, about how this method could benefit all games/systems with a similar deffered pipeline, yet you're here talking about "xbox wins", xbox "exclusive circuitry", and quoting people in discussions you've had in the past where you stated xbox will smoke the competition due to these 'features'. It's because you don't even know what "Tier 2 VRS" means here. You don't even seem to grasp where the optimisation comes from.

Sosokrates Sosokrates Riky drags Panajev into this thread by quoting him regarding Rikys own past console war comments.
Panajev comes here and discusses what Riky quoted him about and you're asking panajev to go somewhere else because he's console warring? Panajev is even nice enough to apologise for it.

Riky thinks his console war comments are vindicated. Half the comments on the first page are referencing Rikys console war comments of the past. Riky is clearly more concerned about "xbox wins" instead of optimisation wins for the industry in this thread. There is a clear console war mentality that you're ignoring and you've got the wrong guy for
Where has riky said xbox is winning or wins?

He seems to be praising xbox and there technologies in a xbox thread, nothing wrong with that.
 

VFXVeteran

Banned
15% might not seem important, but is always helpful. It can be the difference between maintaining a constant frame rate, or having small drops below 30 or 60 fps. If you don't have a VRR display and console/PC, this means screen tearing and judder.
Consider that consoles don´t have hardware for ray traversal, so t's done in shaders. This means 15% more shader units can be used for rays casting.
That's what I'm saying. That's going to be negligible and unnoticeable. I would not use 15% shader throughput for just more intersections for a better approximation on something like shadows (they will still be too low resolution) or reflections or GI.

It also might mean higher resolution for the rest of the image.
I would doubt this because if you raise the pixel count by a little on the final framebuffer, you'd have to do it for all the framebuffers in the rendering pipeline - which could push that 15% gain to losses. And if we only apply it on reconstruction, it won't make such a big difference when there are several sublevel samples that need to be made across the entire graphics pipeline (i.e. depth buffers, w-buffers, post FX, etc.)

It will always depend on where the devs want to spend that extra available shading power, but it will make things a bit better.
This is the reality. I don't doubt it will make things a bit better but not enough to actually "see" those differences such that the look is dramatic.

It's not a huge difference. It's only 15%. But it helps that one bit.
It's basically like getting one extra TFLOP out of these machines.
I would rather take the side of conservativeness and state the reality. I think this would avoid a lot of arguments from console owners of a different platform if you get my drift.
 

winjer

Gold Member
That's what I'm saying. That's going to be negligible and unnoticeable. I would not use 15% shader throughput for just more intersections for a better approximation on something like shadows (they will still be too low resolution) or reflections or GI.

I would doubt this because if you raise the pixel count by a little on the final framebuffer, you'd have to do it for all the framebuffers in the rendering pipeline - which could push that 15% gain to losses. And if we only apply it on reconstruction, it won't make such a big difference when there are several sublevel samples that need to be made across the entire graphics pipeline (i.e. depth buffers, w-buffers, post FX, etc.)

This is the reality. I don't doubt it will make things a bit better but not enough to actually "see" those differences such that the look is dramatic.

I would rather take the side of conservativeness and state the reality. I think this would avoid a lot of arguments from console owners of a different platform if you get my drift.

For all you said, you have already been proven wrong by what the Coalition did with Gears 5.
The game, on Series S/X consoles, has gained significant performance. And with no noticeable loss of quality.
Once again, it's only 10-15%. But it's a gain. Add all the other gains devs are doing with optimizations, and things start to add up to great values.
This is what's all about, squeezing all available performance, with every trick in the book.
And that talk about losses, is just wrong. Watch the videos posted by MS, to understand what people are talking about.
 

Sega Orphan

Banned
Whos saying "xsx is so suprior"

I mean its really close, but all in all the xsx has better overall visuals more of the time.
The consoles are closer in apecs than any other generation before. While the XSX has the hardware advantage, however slight, Sony has a far more mature tool set and development environment than the XSX does. They are far more mature. The change over curve from PS4 to PS5 is alot less than from Xbox One to Xbox Series. That was one of the aims of Cerny, which he spoke about with the time to triangle.
You can see that in the first generation of games released. MS was behind on that. They even admit that. They claim it was because they were waiting on the full RDNA 2 and DX12 U tech. That makes sense as AMD RDNA 2 and DX12U are joined at the hip.

The possible benefits to Xbox is they have nice tech like Sampler Feedback Streaming, VRS, Mesh Shaders, XVA and ML on board the console, but these need to be adopted and exploited for it to be realised. VRS is a pretty easy to adopt tech, however others like Mesh Shaders, ML and SFS will need to be built in to game engines. There is no guarantees that they will be. Hopefully with the majority of PC games being DX12, this may happen, however there is one thing you need to understand, and that's parity. No third party dev is going to put more effort into one console to use their full abilities at the expense of the other console. They want to have the same quality on both. If they have a more performant game on console A compared to console B, they get to look forward to all the social media fallout. Cries of lazy devs, their twitter fees full of rabies infected fanboys of one console or the other complaining about them, the online petitions and not to mention actual correspondence from the console maker asking for a please explain.

The only hope you have this gen of either console fully getting exploited will be first party games. Sony's devs will push their SSD, lower level APIs and primitive shaders to get the best they can, while MS will be pushing their DX12 extensions.

There is actually one interesting point that may make the XSXs DX12U extensions get exploited a bit earlier, and that is due to the XSS. We have Memory issues on the XSS from what some devs have said, and one of the positives of the DX12U is Sampler Feedback Streaming. This will absolutely help with Memory management on the XSS, and some devs may well get into this early to evleviate the Memory issues they are having. This experience will mean it could also get applied to XSX games as well for even better performance.

The short of it is this gen is going to be 99% of games looking identical with maybe one have a coupe of FPS advantage here, maybe a pixel or more extra on here, which absolutely no one will be able to see unless they do the 400x zoom and pixel count or run a FPS counter with them alongside each other.
No console is going to end up with an inferior version, which after the PS3 vs 360 era of poor PS3 versions, and the last gen of subpar Xbox One versions is great news.
Everyone's a winner.
 
Last edited:

VFXVeteran

Banned
For all you said, you have already been proven wrong by what the Coalition did with Gears 5.

The game, on Series S/X consoles, has gained significant performance. And with no noticeable loss of quality.
Once again, it's only 10-15%. But it's a gain. Add all the other gains devs are doing with optimizations, and things start to add up to great values.
This is what's all about, squeezing all available performance, with every trick in the book.
And that talk about losses, is just wrong. Watch the videos posted by MS, to understand what people are talking about.
So you believe over this entire generation that performance gains for the Xbox will be so drastic that all the games will look a generation ahead of what we are seeing now? Surely you don't believe that. Do you?

You will never see a developer create double the performance of hardware. Bandwidth will always be the limit no matter what a programmer tries to do. Also, these squeezing out performance with every trick in the book for graphics features has been done already. We are not dealing with assembly language programming here. In other words, you aren't going to increase the number of specular lobes for a hair shader for free if the algorithm only has enough bandwidth for 1 specular lobe. RT is the only unknown area that can be explored more - but again - bandwidth will be the limit there.

I get the sense from both camps that developers have yet to "discover" the new hardware as they are all brute-forcing things even though the foundation for the APIs of each platform has been created last generation. To me, that's just wishful thinking.
 
Last edited:

Sosokrates

Report me if I continue to console war
Exactly I'm pretty happy they are having involvement in the engine, it obviously helped in the Matrix Demo. I wonder if we see some sort of Test project before Gears 6.

Theres been leaks and reports that they are working on multiple projects, one which is coming before gears 6
 

winjer

Gold Member
So you believe over this entire generation that games for the Xbox will be so drastic that all the games will look a generation ahead of what we are seeing now? Surely you don't believe that. Do you?

You will never see a developer create double the performance of hardware. Bandwidth will always be the limit no matter what a programmer tries to do. Also, these squeezing out performance with every trick in the book for graphics features has been done already. We are not dealing with assembly language programming here. RT is the only unknown area that can be explored more - but again - bandwidth will be the limit there.

Yes, I believe games will look better a few years from now. That has ALWAYS happened in all consoles.
And the reason is that devs get to know the hardware better. Use new techniques. Use new tricks. Create new tools.
VRS is just one of these new techniques. Alone, it will make a small difference. But with other techniques, it will make games look even better.
No, not every trick in the book has been done already. We are now getting more games with VRS. And games will start using new geometry engines, with Primitive or Mesh Shaders. Soon we'll start seeing games using Epic's TAA Gen5. And wider use of TAAU. And FSR. And maybe upscalers like XeSS running on DP4A.
And games using Sampler Feedback Streaming. And the list of new tech is constantly increasing. Even just this week we found a new patent from Sony to improve Ray-Tracing performance.

And it's not necessary to code to hardware to create better looking games. We are just seeing the beginning of a new generation of compilers, using machine Learning, that will make things faster and more efficient.

That talk about bandwidth is cute. As if this is the only thing that matters for performance.
But you do realize that if you have to shade 15% less fragments, that's less memory access you have to do. Consider that VRS can be applied to many graphical effects. Just watch MS and the Coalition explain this.
 

Three

Member
You're obviously upset but you can't change history as it's all here to see and I've quoted it, up to the very day these presentations were made these people were trying to ridicule and downplay claiming it had come and gone and was just marketing terms.
You're obsession with console wars never comes with the quotes, maybe because it's all in your head.
Xbox "wins" are just ways of getting extra performance from fixed hardware, you can take it whatever way you like.
People can obviously see how much stick I've got for believing what MS said about increasing performance and the proof has arrived, they are just referencing that.
Anyway I'm done talking about it and unless you'e got something to say about the actual topic then there is no point talking about it anymore.
There is a screenshot of your console warring post. That's what you get stick for. You get stick because you go around talking about 2x performance advantage due to SFS, Series S trading blows with PS5 etc.

The "Tier VRS 2 came and went" was in a discussion talking about that console war gap comment of yours. You want me to believe that you quoted it here because it has the words 'VRS Tier 2' in it, completely ignoring context entirely?
Where has riky said xbox is winning or wins?

He seems to be praising xbox and there technologies in a xbox thread, nothing wrong with that.
Everywhere if you look at any of his posts.
Only if you want to completely ignore what he's talking about does it seem like he's just praising xbox.
My post here explains where he is approaching this from. Make sure you look at the screenshot of what Riky said in the first quote. There would be no need to discuss "xbox wins" or "xbox exclusive circuitry" like he does here if he wasn't approaching it from that angle . These optimisations are not xbox specific in any way. Even his idea of "Tier 2 VRS" isn't xbox specific either. Well it is in the sense that it's the term MS used in DX12 for driver/API compatibility but screen space image VRS can be done even on last gen consoles. He doesn't get that though. He's more concerned about "xbox wins" and "xbox exclusive circuitry" to care.
 

Sosokrates

Report me if I continue to console war
The consoles are closer in apecs than any other generation before. While the XSX has the hardware advantage, however slight, Sony has a far more mature tool set and development environment than the XSX does. They are far more mature. The change over curve from PS4 to PS5 is alot less than from Xbox One to Xbox Series. That was one of the aims of Cerny, which he spoke about with the time to triangle.
You can see that in the first generation of games released. MS was behind on that. They even admit that. They claim it was because they were waiting on the full RDNA 2 and DX12 U tech. That makes sense as AMD RDNA 2 and DX12U are joined at the hip.

The possible benefits to Xbox is they have nice tech like Sampler Feedback Streaming, VRS, Mesh Shaders, XVA and ML on board the console, but these need to be adopted and exploited for it to be realised. VRS is a pretty easy to adopt tech, however others like Mesh Shaders, ML and SFS will need to be built in to game engines. There is no guarantees that they will be. Hopefully with the majority of PC games being DX12, this may happen, however there is one thing you need to understand, and that's parity. No third party dev is going to put more effort into one console to use their full abilities at the expense of the other console. They want to have the same quality on both. If they have a more performant game on console A compared to console B, they get to look forward to all the social media fallout. Cries of lazy devs, their twitter fees full of rabies infected fanboys of one console or the other complaining about them, the online petitions and not to mention actual correspondence from the console maker asking for a please explain.

The only hope you have this gen of either console fully getting exploited will be first party games. Sony's devs will push their SSD, lower level APIs and primitive shaders to get the best they can, while MS will be pushing their DX12 extensions.

There is actually one interesting point that may make the XSXs DX12U extensions get exploited a bit earlier, and that is due to the XSS. We have Memory issues on the XSS from what some devs have said, and one of the positives of the DX12U is Sampler Feedback Streaming. This will absolutely help with Memory management on the XSS, and some devs may well get into this early to evleviate the Memory issues they are having. This experience will mean it could also get applied to XSX games as well for even better performance.

The short of it is this gen is going to be 99% of games looking identical with maybe one have a coupe of FPS advantage here, maybe a pixel or more extra on here, which absolutely no one will be able to see unless they do the 400x zoom and pixel count or run a FPS counter with them alongside each other.
No console is going to end up with an inferior version, which after the PS3 vs 360 era of poor PS3 versions, and the last gen of subpar Xbox One versions is great news.
Everyone's a winner.
Interesting points.

Yes its interested how both approaches differed where PS5 dev environment was an improved PS4 dev environment, the benifits of one sku.

Xbox went from the XDK which xbox specific to GDK which is "general"and covers xbox one, pc and xbox series.

I think because of the XSS ram limitation ironically we will see Innovations in ssd streaming. It does have only 8gb of ram, but like its bigger brothers the SSD with alleviate that somewhat.
 

Sosokrates

Report me if I continue to console war
There is a screenshot of your console warring post. That's what you get stick for. You get stick because you go around talking about 2x performance advantage due to SFS, Series S trading blows with PS5 etc.

The "Tier VRS 2 came and went" was in a discussion talking about that console war gap comment of yours. You want me to believe that you quoted it here because it has the words 'VRS Tier 2' in it, completely ignoring context entirely?

Everywhere if you look at any of his posts.
Only if you want to completely ignore what he's talking about does it seem like he's just praising xbox.
My post here explains where he is approaching this from. Make sure you look at the screenshot of what Riky said in the first quote. There would be no need to discuss "xbox wins" or "xbox exclusive circuitry" like he does here if he wasn't approaching it from that angle . These optimisations are not xbox specific in any way. Even his idea of "Tier 2 VRS" isn't xbox specific either. Well it is in the sense that it's the term MS used in DX12 for driver/API compatibility but screen space image VRS can be done even on last gen consoles. He doesn't get that though. He's more concerned about "xbox wins" and "xbox exclusive circuitry" to care.
Thats one post over a year ago.

I think its time to let it go.

Xbox series hardware does have an advantage in VRS because of the RB+ units which are different to RDNA1 cards and the PS5.

k5Np0uV.png
 

VFXVeteran

Banned
Yes, I believe games will look better a few years from now. That has ALWAYS happened in all consoles.
They don't look drastically better since PS4 era when custom hardware became more off-the-shelf hardware.

And the reason is that devs get to know the hardware better. Use new techniques. Use new tricks. Create new tools.
All tools aren't created over from scratch. Most pipeline tools will remain the same. The hardware last gen isn't much different from the hardware this generation. That's just a fact.

VRS is just one of these new techniques. Alone, it will make a small difference. But with other techniques, it will make games look even better.
What other techniques?

No, not every trick in the book has been done already. We are now getting more games with VRS. And games will start using new geometry engines, with Primitive or Mesh Shaders.
What do you think mesh shaders will do that will make the geometry look dramatically different from the older method? If you say it will increase performance, I'll say ok. But that doesn't make it look different.

Soon we'll start seeing games using Epic's TAA Gen5. And wider use of TAAU. And FSR. And maybe upscalers like XeSS running on DP4A.
OK. So where is that making the final rendered image look better? A native 4k image is visually better than any of those reconstruction techniques. Why not just have enough bandwidth to render a native 4k image?

And games using Sampler Feedback Streaming. And the list of new tech is constantly increasing. Even just this week we found a new patent from Sony to improve Ray-Tracing performance.
But people are exaggerating their own visions of what they think games will look like when last gen they weren't a generation ahead of the earlier games.

And it's not necessary to code to hardware to create better looking games. We are just seeing the beginning of a new generation of compilers, using machine Learning, that will make things faster and more efficient.
ML is in software for the consoles. It's not hardware like the Nvidia boards.

That talk about bandwidth is cute. As if this is the only thing that matters for performance.
But you do realize that if you have to shade 15% less fragments, that's less memory access you have to do. Consider that VRS can be applied to many graphical effects. Just watch MS and the Coalition explain this.
OK. So this VRS has just got you under a spell dude. It's not going to give you the results you expect.

You guys have very very unrealistic expectations for these consoles. I saw this last gen when people thought we would be at CGI levels. I remember the kite demo from Epic and people swore it would come to last gen games. It didn't. And here we are 2yrs into next gen and thinking we'll have enough silicon and technical expertise to push visuals to unrealistic expectations yet again.

Even the Matrix demo wasn't vastly better looking than today's games. Every generation Epic puts up a demo and people think by the end of the generation, despite most companies only able to put out ONE big title each generation, will look like it. I'll revisit this conversation every year to see the big differences evolving from the various game developers and remind people of their expectations.
 

Riky

$MSFT
There is a screenshot of your console warring post. That's what you get stick for. You get stick because you go around talking about 2x performance advantage due to SFS, Series S trading blows with PS5 etc.

The "Tier VRS 2 came and went" was in a discussion talking about that console war gap comment of yours. You want me to believe that you quoted it here because it has the words 'VRS Tier 2' in it, completely ignoring context entirely?

Everywhere if you look at any of his posts.
Only if you want to completely ignore what he's talking about does it seem like he's just praising xbox.
My post here explains where he is approaching this from. Make sure you look at the screenshot of what Riky said in the first quote. There would be no need to discuss "xbox wins" or "xbox exclusive circuitry" like he does here if he wasn't approaching it from that angle . These optimisations are not xbox specific in any way. Even his idea of "Tier 2 VRS" isn't xbox specific either. Well it is in the sense that it's the term MS used in DX12 for driver/API compatibility but screen space image VRS can be done even on last gen consoles. He doesn't get that though. He's more concerned about "xbox wins" and "xbox exclusive circuitry" to care.

Still no quotes, you screenshot one quote that was in a "speculation" thread, you know what that means? Speculation, look it up.
As for "2x performance" or "exclusive circuitry" bring the quotes, stop derailing the thread, it's only about me in your head, it's sad. Bring up the exact quotes where I said those exact phrases or be quiet.
 

Riky

$MSFT
Thats one post over a year ago.

I think its time to let it go.

Xbox series hardware does have an advantage in VRS because of the RB+ units which are different to RDNA1 cards and the PS5.

k5Np0uV.png

Yes they have the DX12 API fallback position, but it works for Vulcan to, they can't grasp the hardware part or they choose to ignore it.
 

Three

Member
Thats one post over a year ago.

I think its time to let it go.

Xbox series hardware does have an advantage in VRS because of the RB+ units which are different to RDNA1 cards and the PS5.
I've let it go already and I only laugh at at his old post. He hasn't changed though. The issue is that Riky continues to double down on it by arguing with others about it to this day and drags those arguments into this thread to argue some more, but you're blaming others for that and suggesting he's innocently hyped for tech instead. S/He's not.

Regarding RB+ this isn't what the results in the video are about at all. It isn't a case of them not having RB+ then having it so again why the console war comparison/discussion by him and now you? The gains are not console specific and the VRS described can be done on all consoles and PC and the percentage gain would not change. It's a pipeline optimisation so why discuss 'xbox wins' and 'xbox exclusive circuitry' at all do you think?
 
Last edited:

Riky

$MSFT
This has already been explained a hundred times, VRS is not exclusive to Xbox or even this generation, however hardware support for Tier 2 is for certain PC cards and Series consoles.
Hence Halo Infinite uses Tier 2 on Series consoles but does not on Xbox One, same for Gears, same for Doom Eternal, where id stated it in their interview, I believe them over a raging fanboy.
Go back to the DF quote about it, it explains it fully and the presentation by The Coalition on moving Gears Tactics from Tier 1 to Tier 2 and the implementation in Gears 5. All this is freely available to read.
 

Three

Member
Still no quotes, you screenshot one quote that was in a "speculation" thread, you know what that means? Speculation, look it up.
As for "2x performance" or "exclusive circuitry" bring the quotes, stop derailing the thread, it's only about me in your head, it's sad. Bring up the exact quotes where I said those exact phrases or be quiet.
So you can go on record and say I was only speculating and I was wrong?
Because it doesn't seem like you do that to me. you double down on that nonsense by clinging on to the next thing like wait for SFS or 'Tier 2 VRS part 2: the coalition edition'. Let it go.

What are you requesting quotes for exactly? you've talked about xbox specific hardware and xbox wins all over this forum.

Edit: Define "tier 2 hardware VRS" instead of asking me to read DF. Try not to blow your braincells while you do it.
 
Last edited:

Riky

$MSFT
So you can go on record and say I was only speculating and I was wrong?
Because it doesn't seem like you do that to me. you double down on that nonsense by clinging on to the next thing like wait for SFS or 'Tier 2 VRS part 2: the coalition edition'. Let it go.

What are you requesting quotes for exactly you've talked about xbox specific hardware and xbox wins all over this forum.

Your still talking about me? Your obsession runs deep, bring the quotes about two specific sayings you claim I said.

"you go around talking about 2x performance advantage due to SFS"

Bring the quote where I said Xbox has a 2x performance advantage over PS5, it can't be hard to do.

"Xbox exclusive circuitry"

You've used this several times in this thread, I don't recall using that phrase ever, so bring the quote.

This thread isn't about me, it's about over 2 hours of talks about the performance benefits for developers with Tier 2 VRS, try talking about that. Your obsession is quite scary.
 
Last edited:

winjer

Gold Member
They don't look drastically better since PS4 era when custom hardware became more off-the-shelf hardware.

Games do look better now, than when the generation started, with the PS4 and X1.
But of course, they are still in the same generation. You can't expect a generational leap just from software optimization. We have to be realistic.

All tools aren't created over from scratch. Most pipeline tools will remain the same. The hardware last gen isn't much different from the hardware this generation. That's just a fact.

We are already seeing a few games taking advantage of the rendering pipelines. There are already new tools and engines being developed.
Just look at UE5 and all it brings, such as Nanite, Lumen, Super Resolution, etc.

What other techniques?

The ones I listed, for example.

What do you think mesh shaders will do that will make the geometry look dramatically different from the older method? If you say it will increase performance, I'll say ok. But that doesn't make it look different.

Much greater amount of polygons rendered per second. Meaning more detailed worlds. Higher detailed characters. Longer LODs. Etc.

OK. So where is that making the final rendered image look better? A native 4k image is visually better than any of those reconstruction techniques. Why not just have enough bandwidth to render a native 4k image?

4K is a pipe dream, as it takes to much processing power to render. Consoles are budget options, they are not am RTX 3090.
So the solution is to render at a lower resolution and then upscale. TAAU does an incredible job, second only to DLSS, at this point.
And DLSS 2.X has already proven it's worth many times over. Have you tried DLSS 2.x?

And again that talk about bandwidth. That's just one of the bottleneck to reach 4K. Besides, adding more bandwidth is not that simple.
Adding more channels makes a console more expensive, because of more traces to route in the PCB. And more memory controllers occupying die space on the SoC. And this will also mean greater power consumption. It's not a linear scale.

But people are exaggerating their own visions of what they think games will look like when last gen they weren't a generation ahead of the earlier games.

You are stuck on 2020, thinking things won't improve with the new generation.
We are already seeing some games taking advantage of some new features of the hardware. More will come.

ML is in software for the consoles. It's not hardware like the Nvidia boards.

DP4A is not merely hardware. But regardless, I'm talking about a compiler being improved by ML.
After compilation, during runtime, ML is no longer necessary to run.
Seriously, do you even know what a compiler does?

OK. So this VRS has just got you under a spell dude. It's not going to give you the results you expect.

It has already given the results in several games. That is 10-15% performance improvement.

You guys have very very unrealistic expectations for these consoles. I saw this last gen when people thought we would be at CGI levels. I remember the kite demo from Epic and people swore it would come to last gen games. It didn't. And here we are 2yrs into next gen and thinking we'll have enough silicon and technical expertise to push visuals to unrealistic expectations yet again.

Even the Matrix demo wasn't vastly better looking than today's games. Every generation Epic puts up a demo and people think by the end of the generation, despite most companies only able to put out ONE big title each generation, will look like it. I'll revisit this conversation every year to see the big differences evolving from the various game developers and remind people of their expectations.

I don't know what games you have been playing this last gen that are on par with the Matrix demo.
 

DaGwaphics

Member
For all you said, you have already been proven wrong by what the Coalition did with Gears 5.
The game, on Series S/X consoles, has gained significant performance. And with no noticeable loss of quality.
Once again, it's only 10-15%. But it's a gain. Add all the other gains devs are doing with optimizations, and things start to add up to great values.
This is what's all about, squeezing all available performance, with every trick in the book.
And that talk about losses, is just wrong. Watch the videos posted by MS, to understand what people are talking about.

Agreed. Add to that the fact that Id did say they have already achieved higher DRS resolutions with previous iterations of the tech while stating that the newest optimizations were better. There will definitely be some tangible results, which is great to see.
 
Last edited:

John Wick

Member
You are milking the same sentence to the last drop you can and trying to stretch it and twist it. Spent months saying XYZ tech has not been used, discovered it was used, still did not explain why it was not universally applied, moving on to other saviours. Which it is ironic coming from you because you were utterly convinced XSX would smoke PS5 before launch (12 TFLOPS, true RDNA2, fixed vs variable clocks, GPU clocks not being as important as people say nor CU to cache ratio, etc…) and now you are predicting proper VRS and SFS use to come in 2+ years from now.
Your analysis also only sees XSX with tricks under its sleeves and the competition with nothing left to improve.
I think what GG has done with HFW even though crossgen is proof enough what the PS5 will be capable off.
Totally agree people like pRiky think only the XSX will improve but PS5 is already maxed. First it was the tools argument,,Sony started early, SX has full RDNA2, now 15 months post launch majority of the XSX features aren't being used or still haven't been implemented in the GDK. We should tell Cerny to pause and take break......
 

Three

Member
Your still talking about me? Your obsession runs deep, bring the quotes about two specific sayings you claim I said.

"you go around talking about 2x performance advantage due to SFS"

Bring the quote where I said Xbox has a 2x performance advantage over PS5, it can't be hard to do.

"Xbox exclusive circuitry"

You've used this several times in this thread, I don't recall using that phrase ever, so bring the quote.

This thread isn't about me, it's about over 2 hours of talks about the performance benefits for developers with Tier 2 VRS, try talking about that. Your obsession is quite scary.
You quoted me and I quoted you so I'm not talking about you but to you here. You didnt use that exact circuitry phrase, others did, which you liked and you continued on with the exact same stuff:

This thread is about improvements to Xbox Series games, no need to keep dragging up other consoles that don't feature the hardware

And that's were you're trying to make incorrect claims and steer things to console wars. This thread isn't about xbox series x games it's about game optimisation results and it's about an optimisation that can be done on all hardware and multiplatform games.

"Xbox exclusive circuitry" is a tongue in cheek reference to your idea that "Tier 2 VRS", screen space VRS in this case, can't be done on other hardware. In reality even a PS4 using compute shaders can do it believe it or not.

This boost from nanite and VRS isn't a feature that PCs and PS5 are missing and the percentage gain would not be different. I even gave you an example with a GDC talk already out using a similar pipeline. You think this is about xbox series secret sauce though and you try your hardest to make tech like VRS and SFS about wars all the time. I remember you arguing about PRT+ vs SFS x2 gains back in the day but can't be bothered dig up that exact post. I just remember you went around making claims like this all the time
... that's before VRS, SFS and Mesh Shaders. Gap will get bigger.
Things don't turn out how you expected when the actual results of games came in with VRS though and you still argue that a gap will materialise. One thing you're right about is that the conversation has run its course.
 
Last edited:

VFXVeteran

Banned
Games do look better now, than when the generation started, with the PS4 and X1.
But of course, they are still in the same generation. You can't expect a generational leap just from software optimization. We have to be realistic.
Looking better is to general. You guys are acting like it will make huge leaps in differences of rendered games when it won't. That's why we are arguing.

We are already seeing a few games taking advantage of the rendering pipelines. There are already new tools and engines being developed.
Just look at UE5 and all it brings, such as Nanite, Lumen, Super Resolution, etc.
Ok. You just ignored my comment. Of course there will be new tools. I said most pipelines are already in place. It takes years to develop a game from scratch. Most companies will only have ONE game out this entire generation. They don't have time to iterate on their game with more graphics features to produce a second game. So if anything, you guys are looking at next-gen having the graphics you want. NOT this generation.

Much greater amount of polygons rendered per second. Meaning more detailed worlds. Higher detailed characters. Longer LODs. Etc.
If you add polygons to be rendered it slows the entire rendering pipeline down because they have to be shaded and lit. Again, the hardware can't afford it when concessions are already being made with simple rendering pipelines. The consoles this generation (and even the Nvidia GPUs) are just too weak for what you guys are expecting.

4K is a pipe dream, as it takes to much processing power to render. Consoles are budget options, they are not am RTX 3090.
So the solution is to render at a lower resolution and then upscale. TAAU does an incredible job, second only to DLSS, at this point.
And DLSS 2.X has already proven it's worth many times over. Have you tried DLSS 2.x?
DLSS is for the PC NOT the consoles. We are talking about consoles here. AMD missed the boat on that one. Which is why I'm arguing your expectations are too high for this hardware.

You are stuck on 2020, thinking things won't improve with the new generation.
We are already seeing some games taking advantage of some new features of the hardware. More will come.
What game is taking advantage of new features that looks better than anything else? I'm curious.

DP4A is not merely hardware. But regardless, I'm talking about a compiler being improved by ML.
After compilation, during runtime, ML is no longer necessary to run.
Seriously, do you even know what a compiler does?
Ah.. now we want to make personal attacks about what I know?

I don't know what games you have been playing this last gen that are on par with the Matrix demo.
Aside from Nanite technology smoothing out polygonal edges, the Matrix demo is doing nothing new that hasn't already been implemented before. The shaders are all old based off of last gens PBR shaders. The lighting is using RT just like many other games (i.e. Metro, UE4's Ascent, Control, etc.). FX are conventional and equal to most other games' FX. Animation isn't that good. Textures aren't as high res as Crysis remake (which holds the crown right now on 8K texture maps). So what other technology in this demo is significantly better than any other game which shows a graphics feature in the hardware that shows a significant change over games already released this generation?
 
Last edited:

onesvenus

Member
"Xbox exclusive circuitry" is a tongue in cheek reference to your idea that "Tier 2 VRS", screen space VRS in this case, can't be done on other hardware. In reality even a PS4 using compute shaders can do it believe it or not.

This boost from nanite and VRS isn't a feature that PCs and PS5 are missing and the percentage gain would not be different
Isn't it true that Xbox has hardware VRS that's not in PCs or the PS5? I thought that was confirmed by Microsoft.

Even if what you say about the performance gains being the same doing it via hardware or via a compute shader (sources on that?), freeing the computer shader to do it via hardware would allow other things to be done. I don't think you can argue it's the same thing (and sorry if I misunderstood your point)
 

Three

Member
Looking better is to general. You guys are acting like it will make huge leaps in differences of rendered games when it won't. That's why we are arguing.


Ok. You just ignored my comment. Of course there will be new tools. I said most pipelines are already in place. It takes years to develop a game from scratch. Most companies will only have ONE game out this entire generation. They don't have time to iterate on their game with more graphics features to produce a second game. So if anything, you guys are looking at next-gen having the graphics you want. NOT this generation.

If you add polygons to be rendered it slows the entire rendering pipeline down because they have to be shaded and lit. Again, the hardware can't afford it when concessions are already being made with simple rendering pipelines. The consoles this generation (and even the Nvidia GPUs) are just too weak for what you guys are expecting.


DLSS is for the PC NOT the consoles. We are talking about consoles here. AMD missed the boat on that one. Which is why I'm arguing your expectations are too high for this hardware.


What game is taking advantage of new features that looks better than anything else? I'm curious.


Ah.. now we want to make personal attacks about what I know?


Aside from Nanite technology smoothing out polygonal edges, the Matrix demo is doing nothing new that hasn't already been implemented before. The shaders are all old based off of last gens PBR shaders. The lighting is using RT just like many other games (i.e. Metro, UE4's Ascent, Control, etc.). FX are conventional and equal to most other games' FX. Animation isn't that good. Textures aren't as high res as Crysis remake (which holds the crown right now on 8K texture maps). So what other technology in this demo is significantly better than any other game which shows a graphics feature in the hardware that shows a significant change over games already released this generation?
I think you're coming from the perspective that these incremental improvements won't be huge leaps in games compared to some paradigm shift with raytracing. You would be right, the incremental improvements won't be huge leaps in rendered games but many small steps still get you somewhere.
 

DaGwaphics

Member
So you believe over this entire generation that performance gains for the Xbox will be so drastic that all the games will look a generation ahead of what we are seeing now? Surely you don't believe that. Do you?

It would be difficult to not see that as the likely outcome. If the generation continues to produce software that looks near identical to the cross-gen phase, that would be something unique that has never happened before. Games in the last years of the PS360 generation are head and shoulders above similar releases in the first couple years of that cycle, as are games from later in the X1/PS4 cycle. We haven't seen the type of levels that can be designed with the SSDs as a guarantee, same with the increased CPU and some of the modern GPU features. At this point, the software seems to be just brute forcing X1/PS4 games with maybe some RT on top.
 
Last edited:

John Wick

Member
You quoted that simply because it had the words "VRS tier 2" in it? Really?
Not only am I able to read but I'm able to read context too because I wasn't born yesterday.

You quoted that because you think you had a gotcha in a console warring argument you started and doubled down on in another thread discussing VRS closing a console war gap that didn't materialise.

I'm not hurt by your comments but it's the fact that you ignore the context of the discussion you yourself started then act like a victim of console warring when it's you perpatrating it.

You should be talking about the approach, about how this method could benefit all games/systems with a similar deffered pipeline, yet you're here talking about "xbox wins", xbox "exclusive circuitry", and quoting people in discussions you've had in the past where you stated xbox will smoke the competition due to these 'features'. It's because you don't even know what "Tier 2 VRS" means here. You don't even seem to grasp where the optimisation comes from.

Sosokrates Sosokrates Riky drags Panajev into this thread by quoting him regarding Rikys own past console war comments.
Panajev comes here and discusses what Riky quoted him about and you're asking panajev to go somewhere else because he's console warring? Panajev is even nice enough to apologise for it.

Riky thinks his console war comments are vindicated. Half the comments on the first page are referencing Rikys console war comments of the past. Riky is clearly more concerned about "xbox wins" instead of optimisation wins for the industry in this thread. There is a clear console war mentality that you're ignoring and you've got the wrong guy for
Couldn't have said it any better.
Anyday now once VRS and SFS are implemented the XSS will push the PS5 hard.
 

Riky

$MSFT
You quoted me and I quoted you so I'm not talking about you but to you here. You didnt use that exact circuitry phrase, others did, which you liked and you continued on with the exact same stuff:



And that's were you're trying to make incorrect claims and steer things to console wars. This thread isn't about xbox series x games it's about game optimisation results and it's about an optimisation that can be done on all hardware and multiplatform games.

"Xbox exclusive circuitry" is a tongue in cheek reference to your idea that "Tier 2 VRS", screen space VRS in this case, can't be done on other hardware. In reality even a PS4 using compute shaders can do it believe it or not.

This boost from nanite and VRS isn't a feature that PCs and PS5 are missing and the percentage gain would not be different. I even gave you an example with a GDC talk already out using a similar pipeline. You think this is about xbox series secret sauce though and you try your hardest to make tech like VRS and SFS about wars all the time. I remember you arguing about PRT+ vs SFS x2 gains back in the day but can't be bothered dig up that exact post. I just remember you went around making claims like this all the time

Things don't turn out how you expected when the actual results of games came in with VRS though and you still argue that a gap will materialise. One thing you're right about is that the conversation has run its course.

You're not talking about me? You've talked about nothing else, huge rants just about me in fact, I'm flattered my opinion matters so much to you.

You used a "tongue in cheek" quote, or in fact you made up one, it's not a quote if I didn't say it. You claim you can't be bothered to dig up quotes now, in other words you lied and couldn't find them because they don't exist, sad.

As per the joint AMD/MS statement, the Series consoles are the only ones with hardware support for Tier 2 VRS


"Xbox Series X|S are the only next-generation consoles with full hardware support for all the RDNA 2 capabilities AMD showcased today.
"
Take it up with AMD, their logo is atop of the statement, not me. You keep trying to confuse software VRS which I've already said can be done elsewhere and last gen to muddy the waters, the statement is clear, the analysis from DF is clear,

"It's also interesting to note that Xbox Series consoles use the hardware-based tier two VRS feature of the RDNA2 hardware, which is not present on PlayStation 5. VRS stands for variable rate shading, adjusting the precision of pixel shading based on factors such as contrast and motion. Pre-launch there was plenty of discussion about whether PS5 had the feature or not and the truth is, it doesn't have any hardware-based VRS support at all"

Again take it up with DF, not me.

The results are what I said they would be, as per Doom Eternal, 17% difference between tflop power, but up to 30% resolution difference,

"Next up, there's the 120Hz mode, which works best on HDMI 2.1 displays, allowing for the game's full resolution to successfully resolve at full frame-rate. This looks to offer something akin to the last-gen Doom Eternal experience at twice the performance level. Xbox Series X operates at a dynamic 1800p, while PlayStation 5 tops out at 1584p - and it is visibly blurrier"

Again take it up with DF, it's the only third party game to use Tier 2 VRS that appears on both systems, id did an interview about it, they said they wish every platform supported it, were they wrong? You think you know better? Take it up with them and educate them.

I'm only pointing out what other more respected sources have said, nothing more nothing less.
 
Last edited:

John Wick

Member
Still no quotes, you screenshot one quote that was in a "speculation" thread, you know what that means? Speculation, look it up.
As for "2x performance" or "exclusive circuitry" bring the quotes, stop derailing the thread, it's only about me in your head, it's sad. Bring up the exact quotes where I said those exact phrases or be quiet.
Pot Kettle Black
 

ckaneo

Member
Yes, I believe games will look better a few years from now. That has ALWAYS happened in all consoles.
And the reason is that devs get to know the hardware better. Use new techniques. Use new tricks. Create new tools.
VRS is just one of these new techniques. Alone, it will make a small difference. But with other techniques, it will make games look even better.
No, not every trick in the book has been done already. We are now getting more games with VRS. And games will start using new geometry engines, with Primitive or Mesh Shaders. Soon we'll start seeing games using Epic's TAA Gen5. And wider use of TAAU. And FSR. And maybe upscalers like XeSS running on DP4A.
And games using Sampler Feedback Streaming. And the list of new tech is constantly increasing. Even just this week we found a new patent from Sony to improve Ray-Tracing performance.

And it's not necessary to code to hardware to create better looking games. We are just seeing the beginning of a new generation of compilers, using machine Learning, that will make things faster and more efficient.

That talk about bandwidth is cute. As if this is the only thing that matters for performance.
But you do realize that if you have to shade 15% less fragments, that's less memory access you have to do. Consider that VRS can be applied to many graphical effects. Just watch MS and the Coalition explain this.
I think games will look better, but I feel we are pretty close to reach art/man power graphical limitations as opposed to technical ones. So games will be way more technically impressive in lighting, effects, models, physics, 4k etc but someone still has to create it all
 

winjer

Gold Member
Looking better is to general. You guys are acting like it will make huge leaps in differences of rendered games when it won't. That's why we are arguing.
Don't care about what "others" are saying.
I'm arguing about my statements.

Ok. You just ignored my comment. Of course there will be new tools. I said most pipelines are already in place. It takes years to develop a game from scratch. Most companies will only have ONE game out this entire generation. They don't have time to iterate on their game with more graphics features to produce a second game. So if anything, you guys are looking at next-gen having the graphics you want. NOT this generation.

I don't understand what you are referring about with pipelines. But in this discussion, I'm talking about the rendering pipeline.
And in this generation we have some of the biggest changes ever. For once, the whole geometry engine is revamped. Especially with Mesh Shaders and Amplification Shaders, replacing the whole geometry part of the rendering pipeline.
Then we have ray-tracing, that can replace or enhance several parts of the rendering pipeline. Be it shadows, reflections or global Illumination.
Then we have the file system and API for SSDs. For example, on the PS5 there's low level and high level access and game-makers can choose whichever flavour they want - but it's the new I/O API that allows developers to tap into the extreme speed of the new hardware. The concept of filenames and paths is gone in favor of an ID-based system which tells the system exactly where to find the data they need as quickly as possible. Developers simply need to specify the ID, the start location and end location and a few milliseconds later, the data is delivered. Two command lists are sent to the hardware - one with the list of IDs, the other centering on memory allocation and deallocation - i.e. making sure that the memory is freed up for the new data.
We could also talk about MS Direct Storage and Sampler Feedback Streaming.
So, as you can see, there is a lot of new improvements in the rendering pipeline.

About tools, yes there are also new tools. UE5 is just one of the most prominent examples.

If you add polygons to be rendered it slows the entire rendering pipeline down because they have to be shaded and lit. Again, the hardware can't afford it when concessions are already being made with simple rendering pipelines. The consoles this generation (and even the Nvidia GPUs) are just too weak for what you guys are expecting.

One of the advantages of Mesh Shaders and even primitive Shaders is the ability to better cull unseen geometry, at an early stage. This means less resources spent on rendering invisible polygons, but also less overdraw for pixel shading. So it's a net win all around.
Also consider that the thing that spends most of shader work, is fragment shading and this depends mostly on render resolution. Not geometry.
Again, don´t try to make me defend other people's arguments. I defend my own.

DLSS is for the PC NOT the consoles. We are talking about consoles here. AMD missed the boat on that one. Which is why I'm arguing your expectations are too high for this hardware.

DLSS 1.9 was running on shaders, using DP4A. So will XESS.
Regardless, have you seen what TAAU can do? I have and although it's not as good as DLSS, it is still something with great results.

What game is taking advantage of new features that looks better than anything else? I'm curious.

The PS5 versions of Spider-man with lots of ray-tracing effects. Ratchet and Clank with great effects of ray-tracing, fur rendering, SSD for level loading.
Watch Dogs Legion with ray-tracing. Doom Eternal and Gear 5 with VRS.
Metro Exodus Enhanced with real time Global Illumination.
And several others. I would advise you subscribe to Digital Foundry, they have plenty of videos talking about these things.

Ah.. now we want to make personal attacks about what I know?
But do you know what is a compiler? And why ML is important to improve performance and efficiency of compiled code?

Aside from Nanite technology smoothing out polygonal edges, the Matrix demo is doing nothing new that hasn't already been implemented before. The shaders are all old based off of last gens PBR shaders. The lighting is using RT just like many other games (i.e. Metro, UE4's Ascent, Control, etc.). FX are conventional and equal to most other games' FX. Animation isn't that good. Textures aren't as high res as Crysis remake (which holds the crown right now on 8K texture maps). So what other technology in this demo is significantly better than any other game which shows a graphics feature in the hardware that shows a significant change over games already released this generation?

No other game is doing better use of geometry at this point, than UE5 with Nanite. No other game has a better TAA than Epic's TAA Gen5. No other game has better temporal upscaler than Epic's Super Resolution, sides from DLSS.
It has probably the best use of SDFs to render Global illumination. It has one of the most impressive streaming and LOD scaling in any game engine.
It is bringing so many new great things. You can't just focus on what it does like previous tech.
 

winjer

Gold Member
BTW, here is a little test with UE5.
One of these images was rendered native at 1440p, rendering at 33 fps on a 2070S. The other was rendered at 1080p, upscaled with Epic's Super Resolution, running at 43 fps.
Can you tell which one is which? And if so, do you think there is such a big difference? Or did you have to pixel peep to find the differences?
Today, asides DLSS, there is nothing on the market doing anything as good as this.

51909656386_7b4760993e_o_d.png


51909749858_42f3d11ced_o_d.png
 

Three

Member
Isn't it true that Xbox has hardware VRS that's not in PCs or the PS5? I thought that was confirmed by Microsoft.
I doubt MS tried to confirm anything regarding the PS5 though I could be wrong and some silly quote exists out there, I would be surprised though.

When people talk about "hardware VRS" they are mainly talking about API/drivers.
When people talk about "software VRS" they are talking about writing their own outside of the API.

What's faster/better will depend on the performance of that API vs your own one and what you're trying to achieve. If your team is talented enough a software one might allow some optimisations that you couldn't do with the API but the API has the benefit of being easier for devs.

You don't need to be sceptical of the % gain being equal when you think about it. This optimisation as a percentage wouldn't change because of the factor being changed. It's not going from software VRS to hardware VRS. What has changed that would make it hardware specific or percentage gains higher on specific hardware? The answer is nothing. It's describing using visibility buffers for VRS at a later stage (nanite vs UE4). Any gains would apply anywhere. There is really good material here on it with software VRS explained in part 3. It's basically what is described in the MS video but in more detail
 
Last edited:

Leyasu

Banned
BTW, here is a little test with UE5.
One of these images was rendered native at 1440p, rendering at 33 fps on a 2070S. The other was rendered at 1080p, upscaled with Epic's Super Resolution, running at 43 fps.
Can you tell which one is which? And if so, do you think there is such a big difference? Or did you have to pixel peep to find the differences?
Today, asides DLSS, there is nothing on the market doing anything as good as this.

51909656386_7b4760993e_o_d.png


51909749858_42f3d11ced_o_d.png
Bottom picture looks sharper to me without my glasses on
 

Riky

$MSFT
Isn't it true that Xbox has hardware VRS that's not in PCs or the PS5? I thought that was confirmed by Microsoft.

Even if what you say about the performance gains being the same doing it via hardware or via a compute shader (sources on that?), freeing the computer shader to do it via hardware would allow other things to be done. I don't think you can argue it's the same thing (and sorry if I misunderstood your point)

"The Series X, on the other hand, uses the RB+ design seen on RDNA 2 which doubles the throughput from four to eight 32-bit pixels per cycle, along with sixteen depth samples, in addition to advanced H/W VRS support. The PS5 relies on a software/API variant of VRS"

Yes there is a difference, it's at a hardware level.
 
Top Bottom