• 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

Tripolygon

Member
I don't think he's talking about you, neither was I it was the other poster.
Springate never said Tier 2 VRS, he just said VRS. Also Tier 2 VRS has at least three different performance profiles according to The Coalition so that muddies the waters further. We know DF did a deep dive on Dirt 5, we know they then came out and said that Doom Eternal was the FIRST third party game to use Tier 2 VRS, so that strongly implies Dirt 5 didn't. We just don't have a definitive answer really.
We also know that what people were blaming on VRS turned out to be the Series S settings applied to the Series X version at launch, Codemasters apologised and fixed it.
Ok, you need to cut this shit out. First, it was Dirt 5 didn't have VRS because Digital Foundry did not catch it when NXGamer for instance accurately called VRS being in Dirt 5 then I even told you David Springate the technical director said they are using it. Now it's not Tier 2 VRS because Digital Foundry said Doom was the first 3rd party game to use VRS Tier 2? Does your whole argument hinge on what DF says? We have the answers from the developers with a freaking debug view showing how it works in the game, how much clarity do you need?

This is a definitive answer.


You might need a refresher on what changes were made to Dirt 5 between versions 1.03 and 2.0. Nobody was blaming VRS for the XSX version having slightly lower geometry, texture filtering, lower tesselation, and POM in some modes. VRS has nothing to do with any of those things.
 

Riky

My little VRR pleasure pearl goes vrrrooommm.
The clarity of the developer saying the Xbox Series version uses Tier 2 VRS, simple.
 

assurdum

Member

Has the fps advantage on seriesX, also there are other games like DMC5 which have higher fps on xsx on some modes, The photo mode in control also shows better fps on xsx.

But you think the PS5 + XSX are even in terms of hardware power, yes?

Well I think if you take all the comparisons, res and fps counts into consideration the XSX edges the PS5.
Its very close though.
There are games with resolution and fps advantage on ps5 too. Most of the games have higher resolution in XSX (though via DRS) but seems quite the minimum effort considered the higher CUs. What surprise me more, if XSX is so superior, why we have seen just Hitman 3 with some higher graphic setting where the list on ps5 is a bit wider?
 
Last edited:

Tripolygon

Member

Sosokrates

Gold Member
There are games with resolution and fps advantage on ps5 too. Most of the games have higher resolution in XSX (though via DRS) but seems quite the minimum effort considered the higher CUs. What surprise me more, if XSX is so superior, why we have seen just Hitman 3 with some higher graphic setting where the list on ps5 is a bit wider?
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.
 

Three

Member
I see in it Doom and that's about it. Xbox needs to do more to show off its advantage
Doom is a very well optimised first party game. One of the most lean and best ones out there. I agree that xbox needs to do more to show its advantage but if you're leaning on this presentation and think it's some untapped hardware secret sauce I have to tell you that unfortunately it's not.

The 33% saving UE4 vs UE5 mentioned here is unrelated to any hardware secret sauce some people have conjured up in the thread. The 0.36ms saving is actually coming from what Matt described in that tweet a long time ago (before Nanite was even announced).

It's about how the base pass saving is less compared to getting your geometry and rasterising it into a visibility buffer where you apply THE SAME VRS already being used in Gears 5. It's not some VRS hardware being compared here even though people try and make it about that for secret sauce console wars. It's what gains you get when your pipeline has changed to nanite and changing where you apply VRS. Before anybody asks this applies to other systems too. There is a GDC talk of a similar pipeline already in HFW in March for anyone actually interested in tech and not console wars.
 
Last edited:

Tripolygon

Member
No, lol, that table was proven false, missing a lot of games and gave wrong information.
No, not really, the thread was filled with fanboys quibbling over what slight meant, choosing screenshots that fit their narrative. For the majority of games, both consoles have identical performance, for the few others, they trade blows where one has a performance edge at resolution cost or vice versa. On select occasions, the Series X or PS5 takes the lead but on average they are indentical.
 
Last edited:

Riky

My little VRR pleasure pearl goes vrrrooommm.
The Nanite saving is not a comparison between UE4 and UE5 it doesn't say that at all, it's not being used in Gears 5.
What they are saying is the VRS quality setting they used in Gears 5 and applying it to the UE5 demo gives that saving.
That's obviously new as the demo itself is recent.
It's actually useful to watch the 2 hour presentation before commenting.
 
Last edited:

DaGwaphics

Gold Member
The price will put them in same ballpark, but a console that is the same price can still outperform the other console which is the same price.

Regarding loading times, its depends on the game, it depends how they are designed.

Of course. I was referring to the narrative where Xbox can never "win" in any comparison, because any "win" that does occur isn't as big as it should be. While simultaneously the PS5 can never "lose" a comparison even when the statistics are at odds with that, because it "isn't losing by much". It's a nonsensical narrative, but funny to watch play out none the less. Some of the comments read like the XSX is a 1k system getting whipped by a system half the price. There is absolutely zero pressure on either system other than to be on par, they are the same price. With that said, I'm sure engineers behind both feel some pride when a specific game responds well to the choices made for the box etc. and obviously with two different systems one of them is going to come out a frame or two on top every time.
 
Last edited:

DaGwaphics

Gold Member
I don't think I've seen anyone say there is a negative in any threads, raytracing or here. The comparison is usually made by people trying to shitpost about the PS5 both in the raytracing thread and here. I can pinpoint where the first comparison was made in both threads if you like (and yes, it's on the first page).

All I'm seeing on the first page is a post of @Panajev2001a quoted by Riky Riky that only calls out differences in opinion on the subject at hand (no mention of the competition there at all). And of course @ assurdum assurdum doing his usual. 🤷‍♂️
 

Three

Member
All I'm seeing on the first page is a post of @Panajev2001a quoted by Riky Riky that only calls out differences in opinion on the subject at hand (no mention of the competition there at all). And of course @ assurdum assurdum doing his usual. 🤷‍♂️
Not sure what you're reading here or if you're willfully acting like you don't understand the context.
Panajev isn't even on the first page in this thread. Riky decided to quote Panajev from an entirely different thread where, you guessed it, it was a comparison thread. Riky decided to drag panajev into this one for you know what reason, this wasn't even related to the SFS in any way but he dragged him here anway. Then the first blatant PS5 comparison is from Darsxx.
 
Last edited:

DaGwaphics

Gold Member
Not sure what you're reading here or if you're willfully acting like you don't understand the context.
Panajev isn't even on the first page in this thread. Riky decided to quote Panajev from an entirely different thread where, you guessed it, it was a comparison thread. Riky decided to drag panajev into this one for you know what reason, this wasn't even related to the SFS in any way but he dragged him here anway. Then the first blatant PS5 comparison is from Darsxx.

The statistical numbers from Doom, when Doom and Id are part of the video/discussion at hand? The context there is clearly about the IQ offered by Doom with or without VRS.
 

Riky

My little VRR pleasure pearl goes vrrrooommm.
The thread isn't about "SFS" don't know where you get that from.
To say it's not related when he actually used the word VRS in that quote is quite bizarre.
Regardless the information of its increase in performance and how it is being taken further now is all in that 2 hour presentation, there is simply no disputing it with all that proof.
Hopefully we'll see a flood of games using this feature that will benefit Xbox and PC games.
 

Ozriel

Member
There are games with resolution and fps advantage on ps5 too. Most of the games have higher resolution in XSX (though via DRS) but seems quite the minimum effort considered the higher CUs. What surprise me more, if XSX is so superior, why we have seen just Hitman 3 with some higher graphic setting where the list on ps5 is a bit wider?

Just like you said, these are still early days for the consoles.



Let’s not move goalposts now.
 

VFXVeteran

Banned
The thread isn't about "SFS" don't know where you get that from.
To say it's not related when he actually used the word VRS in that quote is quite bizarre.
Regardless the information of its increase in performance and how it is being taken further now is all in that 2 hour presentation, there is simply no disputing it with all that proof.
Hopefully we'll see a flood of games using this feature that will benefit Xbox and PC games.
I think the speed up is coming from the cost of shader usage (i.e. the 1.09ms down to 0.74ms?). I don't believe it's a hardware function but suppose that it is.. I still don't think the increase in performance is large enough to get super excited over. It might allow for a few more complex shaders but all in all, I'd rather use it for stability in frametimes.
 
Last edited:

Three

Member
The statistical numbers from Doom, when Doom and Id are part of the video/discussion at hand? The context there is clearly about the IQ offered by Doom with or without VRS.
Where are you getting this imaginary context of him discussing Doom statistics from? The chain of discussion from the other thread Riky dragged in here is from a Cyberpunk thread and it stemmed from his silly idea that Series S will give PS5 a run for its money due to secret sauce RDNA2 features like "tier 2 hardware VRS" which made little difference in Cyberpunk:

Did Series S push PS5 pretty hard in this game or do we have to wait more time?



Which rdna2 features are left to exploit? If this game has it the VRR and the VRS, what’s left to help close the gap on the PS5? Mash Shaders, or is the secret gravy sauce on top of the mash?

We have to be realistic here. Systems are close and thats unlikely to change.

VRS gives the least performance increase
SFS and Mesh Shaders the hardware assisted version will give huge performance increases, although PS5 does feature primitive shaders which should give similar results to Mesh Shaders.

The best is yet to come.

SFS giving a huge performance increase (especially over PRT, but keep overselling SFS)? Much bigger than VRS? What kind of contrived scenario are you taking where improvement to virtualised texture streaming bring more performance than reducing shading rate? You mentioned performance and not quality mind you.

Simple, the least of the increases we got this gen is in amount of Ram and memory bandwith, SFS is directed precisely at that.

You are still banking on that 2-3x memory bandwidth advantage without a quoted basis of comparison (aka no virtual texture streaming) apparently trivial to obtain. It is getting closer to mid 2022 btw.
Sure… 2-3x advantage over what other consoles can do with their PRT implementation…

Considering how few games pushed Xbox One X and PS4 Pro properly, you are also choosing the wrong comparison in terms of bandwidth improvement to the devs baseline. Memory size fine, but again you mention “performance” not “detail/IQ”.

PRT was rarely used, having SFS baked into the GDK and the hardware is a different ball game.

Where do you get these hard stats? What do you think having the feature available in the SDK (the API exposes the feature as it did PRT) does? Virtual texturing/texture streaming schemes have been used for the past 20 years, people were doing partial texture updates / single mip map level streaming on PS2!

Why are you overhyping this so much? First it was VRS tier 2, came and went, now apparently it is SFS, etc… maybe the talk of PS5 throttling and it’s customisations being over engineered pointless bits of propaganda (which urioispy materialised) while on the other side you have been saying the SDK and tools have yet to be used properly for the last year and a half (post launch). Mysteriously the PS5 SDK has also no growth potential and was maximised day 0.

The last one is the one he quoted here. He was leaning on SFS to save him. Why do you think he quoted it here? Because he thinks VRS gives the least boost? Or that he thinks this is related to SFS? Why are you defending this nonsense?

Edit: apologies to all who are now dragged in here from the other thread
 
Last edited:

VFXVeteran

Banned
Where are you getting this imaginary context of him discussing Doom statistics from? The chain of discussion from the other thread Riky dragged in here is from a Cyberpunk thread and it stemmed from his silly idea that Series S will give PS5 a run for its money due to secret sauce RDNA2 features like VRS which made little difference in Cyberpunk:

















The last one is the one he quoted here. Why do you think he quoted it? Because he thinks VRS gives the least boost? Or that he thinks this is related to SFS? Why are you defending this nonsense?
Three, you are a developer. Why are you arguing so much? You act like me sometimes.. :D
 

Sosokrates

Gold Member
Its a shame a developer won't come out and flat out say why the xsx and ps5 are trading blows, people can speculate on here the reasons why, but at the end of the day no one knows.
 

DaGwaphics

Gold Member
Where are you getting this imaginary context of him discussing Doom statistics from? The chain of discussion from the other thread Riky dragged in here is from a Cyberpunk thread and it stemmed from his silly idea that Series S will give PS5 a run for its money due to secret sauce RDNA2 features like "tier 2 hardware VRS" which made little difference in Cyberpunk:

I was referring to the guy that compared Doom on XSX to PS5. He didn't compare the systems as much as he just compared Doom with and without VRS. Darsx or something like that. :messenger_grinning_smiling:
 

DaGwaphics

Gold Member
Its a shame a developer won't come out and flat out say why the xsx and ps5 are trading blows, people can speculate on here the reasons why, but at the end of the day no one knows.

What would they have to say? Results are typical for two similarly priced consoles/GPUs. Even the GC and Xbox traded blows in certain situations, the boxes appear to have slightly different strengths and weaknesses.
 
Last edited:

Sosokrates

Gold Member
What would they have to say? Results are typical for two similarly priced consoles/GPUs. Even the GC and Xbox traded blows in certain situations, the boxes appear to have slightly different strengths and weaknesses.

A GPU with more tflops and memory bandwidth using the same architecture should always perform better then a gpu with less.

It would just be interesting to know why on some games the xsx performs better and others perform better on PS5.
 

Three

Member
I was referring to the guy that compared Doom on XSX to PS5. He didn't compare the systems as much as he just compared Doom with and without VRS. Darsx or something like that. :messenger_grinning_smiling:
I'm not sure what sort of comparison you were expecting then
The first couple of pages is filled with discussion about the video presented, with competitors never being compared/contrasted by anyone, only a certain element starts to try and bring that in.

He was the one who brought the comparison of PS5 to XSX, talked about how he's unsure about how much VRS could help Xbox in comparison. Said PS5 doesn't have hardware to do it and how there are other games that lack it on PS5 etc.
If it was about image quility between VRS off vs on in Doom you can easily do that on a PC and don't need to stipulate on whether the PS5 version uses some kind of VRS or not.

Who do you think was the element that introduced the comparison then if not Riky indirectly and Darsxx directly?
 
Last edited:

DaGwaphics

Gold Member
A GPU with more tflops and memory bandwidth using the same architecture should always perform better then a gpu with less.

It would just be interesting to know why on some games the xsx performs better and others perform better on PS5.

That's like looking at Hitman 3 and asking why PS isn't doing better because a GPU with higher triangles per second and pixel fill should always perform better. In desktop GPUs the lineups are always structured in a way where almost every statistical metric associated with GPUs increases with the TF count. The consoles are using custom GPUs.
 

Sosokrates

Gold Member
That's like looking at Hitman 3 and asking why PS isn't doing better because a GPU with higher triangles per second and pixel fill should always perform better. In desktop GPUs the lineups are always structured in a way where almost every statistical metric associated with GPUs increases with the TF count. The consoles are using custom GPUs.

Which is why it would be good hearing the reasons why from devs.
 

DaGwaphics

Gold Member
I'm not sure what sort of comparison you were expecting then


He was the one who brought the comparison of PS5 to XSX, talked about how he's unsure about how much VRS could help Xbox in comparison. Said PS5 doesn't have hardware to do it and how there are other games that lack it on PS5 etc.
If it was about image quility between VRS off vs on in Doom you can easily do that on a PC and don't need to stipulate on whether the PS5 version uses some kind of VRS or not.

Who do you think was the element that introduced the comparison then if not Riky indirectly and Darsxx directly?

I was referring to your post #263 specifically. That user's first reference to PS5 #44 was just comparing the results of Doom with/without VRS. It would be silly to compare results on PC when only the XSX version has the VRS integration, what PC gpu do we compare it to, what settings do we run the game at. PS5 is a much more natural comparison seeing that it appears to be using near identical visual settings with the exception of VRS. I'm not seeing anything from him on the first page (#36 and #46) that isn't simply discussing the thread subject. Any console warring you see there seems imagined.

You are doing a good job of illustrating the core point I was making (with the description of Darsxx's page one posts), that the mere discussion of techniques/technologies associated with Xbox is triggering for some. Yet these users don't seem triggered when discussing PS related tech. Go figure.

On a side note, this is probably the weirdest exchange I've ever had on this board. Proving that people will in fact spend time arguing about literally anything. That's not a knock on you T Three any more than myself, the fact that we have been arguing different interpretations of a fellow user right here on the same board is kind of hilarious.

Snoop Dogg Reaction GIF
 
Last edited:

Boglin

Member
On a side note, this is probably the weirdest exchange I've ever had on this board. Proving that people will in fact spend time arguing about literally anything.

I wish more people would argue in order to fix a specific misunderstanding or to clear up context like you and T Three have been doing. It actually has the potential outcome of people getting on the same page rather than arguing passed each other endlessly.
 

01011001

Gold Member
A GPU with more tflops and memory bandwidth using the same architecture should always perform better then a gpu with less.

It would just be interesting to know why on some games the xsx performs better and others perform better on PS5.

it's called lead platform.
the fact that there were games on PS2 that ran better than the same game on Xbox says enough about how it can make a huge difference for which console a developer actually puts the time in.
we are talking about a hardware power difference anywhere between 2x and 3x here, yet games sometimes ran worse on that 3x more powerful system

you could wave that one away by saying "well architectural differences could bring out these advantages for one system and not the other"
but that doesn't explain the following...
more recently. where PS4 Pro games run either the same or better than the same game on One X when the One X is literally the same hardware family and architecture but better in every imaginable way... makes no sense, but it happened
we have here a faster CPU of the same architecture... we got a GPU that is 1.4x more powerful and a huge memory size + bandwidth advantage... yet, some games run exactly identically and others even managed to run worse on it
 
Last edited:

Riky

My little VRR pleasure pearl goes vrrrooommm.
The chain of discussion from the other thread Riky dragged in here
First it was VRS tier 2, came and went,

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,




I think the speed up is coming from the cost of shader usage (i.e. the 1.09ms down to 0.74ms?). I don't believe it's a hardware function but suppose that it is.. I still don't think the increase in performance is large enough to get super excited over. It might allow for a few more complex shaders but all in all, I'd rather use it for stability in frametimes.

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.
 
Last edited:

Panajev2001a

GAF's Pleasant Genius
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,






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.
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.
 

Riky

My little VRR pleasure pearl goes vrrrooommm.
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.

You keep saying things with no evidence as I quoted, why has everything got to be some war with you and a comparison with other products.
This thread is about improvements to Xbox Series games, no need to keep dragging up other consoles that don't feature the hardware.
I'm pleased we will see some improvements to especially first party games that can't be compared anyway as they are exclusive, if you don't even own an Xbox I don't see why this bothers you.
 

Sosokrates

Gold 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.

Go console war somewhere else.
 

Panajev2001a

GAF's Pleasant Genius
it's called lead platform.
the fact that there were games on PS2 that ran better than the same game on Xbox says enough about how it can make a huge difference for which console a developer actually puts the time in.
we are talking about a hardware power difference anywhere between 2x and 3x here, yet games sometimes ran worse on that 3x more powerful system

you could wave that one away by saying "well architectural differences could bring out these advantages for one system and not the other"
but that doesn't explain the following...
more recently. where PS4 Pro games run either the same or better than the same game on One X when the One X is literally the same hardware family and architecture but better in every imaginable way... makes no sense, but it happened
we have here a faster CPU of the same architecture... we got a GPU that is 1.4x more powerful and a huge memory size + bandwidth advantage... yet, some games run exactly identically and others even managed to run worse on it

I think XOX and PS4 Pro are not great examples because third parties mostly phoned it in with these consoles, going as far as what brute forcing the game would allow them too. In most cases XOX was clearly ahead pretty much in everythi by that mattered. Which was expected in a console that was $100 more expensive and came out almost a year later (which matters).

Maybe the push towards a system that could afford devkits to be sent out early and achieve closer to target performance sooner helped? Maybe the focus on evolutionary tooling/libraries helped (we will see with XSX’s successor as the transition to GDK will be over)? Maybe some of the customisations did prove to be useful and smart bets?
There may be some truth to the lead platform take, but I would argue that it had been earned by how the system was designed, how they worked with devs, and how they invested on some key elements (not sure why people make fun of “time to triangle” as it was explained… maximising that as well as allowing some puzzles to be explored by devs in second and third waves of software is good system architecture design).

XSX strategy (minus XSS which it is no surprise to say I am not a huge fan of) was different and has its perks: their BC story was far far better in most accounts (loading times saw a bigger boost than what PS5 gave to non optimised titles, graphics and framerate updates were possible and more impactful, etc…). You could get more software across PC, Xbox, and cloud (XSX being the basis of the cloud blades was not a bad choice either… seeing MSFS hitting Xbox was also super appreciated).
 

Sosokrates

Gold Member
Look I am not warring here, nor trying to, and I was quoted and replied to the quote accusing me of besmirching / undervaluing VRS which I am not. So calm your horses.
Which it is ironic coming from you because you were utterly convinced XSX would smoke PS5 before launch

Its just talk like the above is pretty console warrish, its not related to the topic.
 

Panajev2001a

GAF's Pleasant Genius
why has everything got to be some war with you and a comparison with other products.
Comparison to other products can matter as when you look at something you compare it with what was there before and what it is the state of the art around it, but not trying to war here so fair some of that can be curbed on my end. Still going to argue why something feels oversold when it is, this is separate from the times you try to use things to console war and when you are just talking about exclusive software without relation to other consoles or pretend to.

Quite unsure why you felt quoting me trying to prove a point or mock was when your intention is just to talk about improvement in games without warring…

I'm pleased we will see some improvements to especially first party games that can't be compared anyway as they are exclusive
I am pleased about improvements in those games too.

, if you don't even own an Xbox I don't see why this bothers you.

Who told you I do not own an Xbox Series X? I actually got one before my PS5 arrived. You seem to search my post history a lot, but somehow missed that?
 
Top Bottom