Topic: General Xbox Series X|S Thread

Expand

@Senua Do you actually 'understand' what you spout out ad-nauseum or just jumping on the 'fanboy' bandwagon? You often 'quote' or post video's from one of the most bias and biggest Xbox fanboy on youtube - someone who has 'no' technical understanding of hardware. Mesh Shaders are a 'new' way of handling the Geometry pipeline over the traditional 'Primitive' shader method of previous generations but that doesn't mean that Sony's using Primitive shaders in the 'exact' same 'pipeline' as they have always been used. They designed their Geometry Engine to give Primitive Shaders a 'similar' level of control to the developers.

In this video from Sony's 'The Road to PS5' (timestamped for you), Mark Cerny describes the issue with Primitive Shaders in a traditional pipeline. The fact that they 'draw' all the triangles as that's handled by Hardware before the Software can process the Vertices - therefore Culling of triangles (not drawing ANY that won't be visible) is incredibly difficult - but their 'new' Geometry Engine allows devs to have full control of Vertices and procedurally adjusting the complexity - very much the exact same benefits Mesh Shading allows.

Point is, both companies have approached the SAME issues and sought the SAME solution but from a different angle. Sony have looked at the issues they have with the traditional pipeline and evolved primitive shaders and the pipeline to give Developers 'similar' control and results to 'Mesh Shading'. the Principals are very similar. The difference here though is that Sony's solution is very much Bespoke because they are developing for their 'bespoke' ecosystem. MS on the other hand, built the Series consoles to align more closely with PC and want DX12 Ultimate at the heart of these.

Its impossible to say that 'Mesh Shading' gives developers better results than Sony's Geometry Engine because we cannot compare ANY game directly as no game yet is using these. But just listening to Mark Cerny, the principal is virtually the same - but whether or not its as 'effective', as 'Performant' etc we really have to wait and see. Just because the 'traditional' pipeline used Primitive Shaders and 'Primitive' sounds more 'basic', doesn't mean that Sony's solution will be 'worse'. In some cases, it could be better too - like if the 'lead' platform is PS5 and developed around the Primitive Shaders, porting that across and adapting it to use 'Mesh Shaders' instead may give PS5 an advantage.

Of course, there are other 'big' differences too between the Hardware that can come into play - like GPU speed (more cycles per second mean each Core of the GPU can handle more instructions, calculate more vertices/intersections etc but being smaller, cannot handle as many simultaneous instructions). As we know, in an 'optimum' situation, the Series X can process 12.1bn instructions per second and the PS5, 10.2bn but the PS5 Processes more instructions per core because its faster. API's too can affect performance and Sony will be using different API's. There are numerous differences that can affect a 'direct' comparison that its impossible to compare directly - the only way would be if a game was made and optimised by the SAME person with equal familiarity of BOTH to use 'both' solutions on the SAME hardware to see which, if any looks, performs etc better and/or which is 'easier' to get up and running as well as it can too.

Anyway, regardless, we will have to wait to see how BOTH perform when they each use the Hardware to its fullest. Just because MS have adopted Mesh Shaders, doesn't mean that Sony haven't got an 'alternative' bespoke solution to achieve a Similar result. I am NOT going to say that Sony's bespoke solution is 'worse' (or Better either) because right now, no game has used Mesh Shading (and I don't believe any PS5 game has used their Bespoke solution either) so we have NO idea who, if any, as the advantage.

That being said, I would expect the Xbox to have the advantage long term because the PS5 is more Bespoke. Any multi-platform game will more likely go the Mesh Shader route because the majority of Hardware can utilise it and that 'familiarity' will help long term. Of course it may be relatively simple to go back and forth, both use vertices and polygons to cull and reduce/increase quality. Nanite is a form of Mesh Shading and Lumite, a form of RT - in engine as opposed to 'hardware' accelerated. Importing assets may well convert them to a 'format' ready to use Mesh Shading or Sony's Geometry engine. Assets are created from a polygon 'Mesh' anyway...

Expand

@BAMozzy

1. Do you actually 'understand' what you spout out ad-nauseum or just jumping on the 'fanboy' bandwagon?

Ans :- It’s a presentation taken from the Official Microsoft Hotchips Presentation.

2. They designed their Geometry Engine to give Primitive Shaders a 'similar' level of control to the developers.

Ans :- That literally no one knows and can’t even be confirmed from those Mark Cerny’s hollow words (where he makes false claims about Dolby Atmos Gaming (proven wrong later by Dolby) whereas even Microsoft Xbox S|X supports HRTF - EVIDENCE -> youtu.be/Ly2tekIbBZ8). So it’s good when Sony publicly markets the term “Geometry Engine” for their RDNA 1.5 (mistakenly confirmed by Sony Principal Graphics Engineer) but when Microsoft presents it (mesh shading geometry engine) in their Hotchips 2020 technical event it becomes a fanboy narrative. WOW! Why don’t you actually look up the technical documentation on how the mesh shading pipeline ACTUALLY WORKS before even referring it to as same as the one in the marketing video of Sony.
By fanboys if you mean channels like Rand Al Thor, WindowsCentral (Jez Corden) etc. they have been right almost every single time whereas YOU guys are still searching for that self proclaimed “INFINITY CACHE” on that die which has been ultimately proven to be missing in the Xray images. Also no evidence of specialised hardware for machine learning.

Also few educational videos for you recently released by Microsoft for developers.

Expand

@Senua I know exactly what that slide was from but you are confusing things and spouting BS that Rand Al Thor spouts in his ignorance about Hardware and who has been proven wrong COUNTLESS times when it comes to technical understanding. Being 'right' because they report on 'leaks' about Software doesn't mean they have the technical understanding of hardware. Rand Al Thor is a FANBOY Gamer who sits on his FAT ass playing GAMES on Xbox - hence he has 1m Gamer Score. Having 1m in gaming score does NOT mean he is educated on how the hardware works and has NO understanding of Hardware at all - He is like you!! Spouting off BS from an uneducated (at least on this topic) perspective in order to score 'brownie' points with like-minded idiots...

There is a BIG difference between the 'Geometry Engine', a bespoke part of the APU designed by Sony to transform the way Geometry has been handled in the past - and 'geometry engine' - a part of the Pipeline usually the first stage of the Pipeline - which is why its so 'inefficient' and why Sony have 'redesigned' it to give developers more control. You are confusing the two because Sony decided to call their Bespoke part 'Geometry Engine'. What is 'important' though is to understand what Sony are doing...

Which again shows your 'ignorance'. Sony have redesigned their 'Pipeline' to use something that ALL games use currently but in a 'new' way - in a 'similar' way to how Mesh Shading works. A 'mesh' is just a wire frame model made of Polygons and both MS and Sony, as well as EPIC, have looked at improving the efficiency of hardware and 'ease' for developers. Previous Pipelines would have the geometry engine create the scene using 'wire frames' or 'mesh'. Developers in the past would have to make 'lower' polygon count models for more 'distant' objects so that when they are called in at the very start, you don't go 'over' your Polygon budget and waste precious time on all that detail and complexity.

BOTH Sony and MS know that the from years of developing games and speaking to devs. Both have looked at addressing that. BOTH wanted to be able to cull ANY polygons/vertices/meshs that are 'not' visible to player (no point wasting resources on those) and BOTH wanted to be able to vary the complexity of their Wire Frame/Mesh based on distance from camera/size/resolution. No point having a polygon smaller than a 'pixel'.

All that RDNA 1.5 is BS narrative by Xbox fanboys. Yes Sony didn't want to use ALL the features of the Desktop RDNA 2.0 - same as Microsoft because Some are very PC specific. What's the 'point' of using EVERY feature if its taking up 'space' on your APU that you can use for something you need? They designed their own 'bespoke' Chip 'with AMD. The PS4 Pro had 'features' that weren't on the Polaris (OR ANY AMD GPU) because it was built specifically for Sony's console. Believe it or not, MS's GPU doesn't have ALL the features of RDNA2.0 either because they don't need ALL the desktop features. They wanted room for the things they needed for the 'Console' too.

I also never said that Sony's solution actually works EXACTLY the same and will give the 'same' performance. What I said was the principal is very similar AND ultimately does a similar thing to Mesh Shading for developers. However, until we actually see it in operation, we don't know how it compares to Mesh Shading. We have seen more 'demo's' and information about Mesh Shading but as yet, we still haven't got game that uses it. That's more than we have from Sony but doesn't mean that Sony are lying about their hardware. That's like calling MS a liar because they promised Mesh Shading yet we still haven't seen any game utilising this. Even if its as good as Mesh Shading, it still doesn't mean the PS5 would match the Xbox 'exactly' on visual quality AND Performance because there are too many other differences too - like GPU size and speed, CPU speed, API's etc etc - therefore its still going to be 'impossible' to compare. But the Principal advantages Mesh Shading offer are very similar to Sony's Proprietary solution. Unless they can make that Mesh Shading Demo run on PS5 in both methods, as well as the traditional method, you cannot compare at all accurately. You can get an idea of how both compare to their relative hardware.

I really do not see the point in shovelling BS about Sony or trying to make out that I am somehow a Sony Fanboy trying to find something to cling on to in 'hope'. I am a GAMER first and foremost with a reasonable technical understanding of hardware. I couldn't care less what 'brand' name is on the 'box' but I do care about providing fair and accurate information and hate FANBOYS spouting BS about things they clearly have NO understanding of - especially if that ignorance is in some way to belittle the other and/or make the other look vastly superior to.

There is NO way to know right now how well Mesh Shading works in a Game, no way to know if Sony's Geometry Engine compares at all because we haven't got ANY games to compare, any tech demo's from Sony running the the same demo. All I said is that in principal, the description that Mark Cerny himself gave would indicate that they have developed their own bespoke version of what Mesh Shaders offer.

As for Dolby Atmos, Mark Cerny didn't say they would support it - after all that costs money to buy a Dolby Atmos licence and Sony developed their own proprietary 3D audio solution - essentially a Rival to Dolby. What he said was that their Tempest Engine was superior to Dolby Atmos as that can only handle 32 objects where as theirs can handle hundreds - going on to say each raindrop can have its own 'sound'.

Dolby then came out to correct Mark Cerny - saying Atmos can handle hundreds of objects too. Dolby issued a Statement after: “It is not correct that Dolby Atmos handles 32 objects maximum. As a technology, Dolby Atmos can support hundreds of objects simultaneously. Having said that, we refer to some advice from game developers who have already created content with Atmos support: objects are fantastic tools, but you should use them sparingly as regards their number in activity within the same scene. Too many moving objects can create a confusing sound environment.“

However, this again is another example of Sony creating their 'own' bespoke solution - doesn't mean its 'worse' than Windows Sonic, DTS-X or Dolby Atmos - its just Sony's proprietary 3D audio to 'rival' those. That's another part of the 'custom' APU that's not on PC's - like the dedicated decompression area that both Sony and MS have built into their APU. Both coming up with a 'similar' solution to an issue that affected last gen and their 'weak' CPU's.

What you are doing is like Sony Fanboys jumping on Xbox because they didn't mention dedicated Audio processing at the reveal of Xbox so 'assumed' that the speed advantage of the CPU was negated by having to process Audio too and that 3D audio would be a Game Changer - forgetting that the XB1 had a dedicated Audio processor (because XB1 had Kinect and needed to process Voice too) so it wasn't 'new' for Xbox - also forgetting the triggers already provided some haptic feedback on XB1 because MS didn't mention it for Series X whilst Sony spent a portion of time on both their Tempest engine and DS5's Haptics.

Its that ignorance that I object to. Its that lack of understanding that I object to. Its that spreading of BS that I object to. Its just petty and pathetic from petty and pathetic 'fanboys' that give gamers a bad name and reputation.

Its no point putting up all those video's and slides from MS as I have watched them, understood them and excited for the future of gaming on Xbox. However, I am also excited about the future of Playstation too and lets not forget, the PS uses AMD too so could take advantage of some of these features too...

Incidentally, if you want to use the die shot, then why not use the one explaining why some features are NOT necessary for a Console...

Thanks for sharing, the lack of the FPU extensions for Zen 2 are logical. AVX2 will kill performance, heat, frequency anyway and should not be on the CPU.

Again, as I have said many times, a Console has very different needs and cuts are normal. This is NOT a desktop device.

Sony obviously decided to shave off the bits they didn't want in favour of either replacing them with features they want and/or keeping the size and therefore cost down. Sony themselves never claimed they had 'Infinity Cache' - just another one of those rumours that grew because some ignorant fanboy who cannot separate AMD desktop parts from 'bespoke, custom APU's designed specifically for their respective Consoles and their clients needs. If MS/Sony don't want/need desktop features, they remove them to free up more 'space' to put the things they actually want/need for a dedicated 'gaming' machine.

As I said above, we cannot tell yet whether Sony's proprietary solution is 'exactly' the same to the 'end users' - Gamers. Gamers won't care whether a game is using Mesh Shaders or the Geometry Engine, they care about the Visual Quality and Performance. In that Mesh Shading demo, the frame rate jumped nearly 9x - that's like going from CB4k/30 to 4k/120 - double the resolution and 4x the frame rate (an 8x boost) - although that Demo, like the Asteroids one, was designed to showcase the advantages of Mesh Shading - a stuffed room of extremely high polygon count statues, millions of high polygon count Asteroids... Thus demonstrating the 'smoother' transitions (no pop in of higher polygon models and/or better textures) as objects get closer and the boost to performance by 'culling' polygons not in the players vision so the actual benefits in game may not be quite so high.

Anyway, the point I am making is that you cannot categorically say that Sony's Geometry engine is inherently worse because we do not know. It could well be a more efficient way of doing essentially the same thing or maybe does it quicker because the GPU is faster. Maybe it isn't the 'best' solution but until games start coming out that are utilising BOTH so we can get at least 'some' idea on consoles, we just don't know! Yes the PS5 is using Primitive Shaders but that doesn't mean they are using them in the exact same way as the PS4/XB1 for example as Mark Cerny himself states in his Road to PS5.

I have no issue with you sharing DF video's for example as they independent fact based analysis of games running on their respective hardware. If you want to make 'something' of the results, regardless of which has the 'better' resolution and/or performance as that is fact. You can't argue that 1440p is better than 1800p for example or one having a more consistent frame rate - the evidence is right there to 'prove' how that 'specific' piece of Software runs on the hardware and when games do release using these consoles and their 'features' to the fullest, we will have a better idea of the design choices each company made.

Like I said, Sony's PS4 Pro had features that the desktop equivalent didn't offer - more 'advanced' than the RX480 it was 'based' on - and had some 'bespoke' parts not seen on ANY AMD GPU as well as some parts that were missing. Series X too has some RDNA2.0 features, some that are older and some that the desktop will never have.

Even the 'solution' to having 'high Res' textures because you are targeting high res visuals and don't want your texures to look 'soupy' (PS4 Pro...) has been tackled in different ways. Sonys answer was to increase the bandwidth to get those textures streaming in fast - where as MS went at it from 'two' angles - first 'increase' the bandwidth too, but also realised that often, only a 'small' part of the whole texture is visible so why not only stream 'part' of that texture - thus reduce the amount of data you actually 'need' to stream. If you only need around 50% of the texture file, you reduce the amount of texture data by half. If the PS5 has to stream 10GB of texture files, Xbox, using Sampler Feedback may only need 5GB. Of course, there are 'pro's and con's' to both - Sony's bandwidth is always there but on Xbox, devs would need to use Sampler Feedback to offset that. However, if they do, they also free up RAM, free up resources because you don't have all those large texture files embedded in etc. Does that give Xbox a 'massive' advantage? Maybe - if developers embrace it and use it. Its just a different solution to the same problem...

Whether one ends up being 'better' overall or not, we just do not know. So to say Xbox has this 'big' advantage that will see a '9x' gain over PS5 because PS5 only uses 'Primitive Shaders' is complete fanboy conjecture. The whole point of the Geometry Engine is to change the 'traditional' graphics Pipeline - eliminate some of the inefficiencies and issues that pipeline created - something Mesh Shaders too looks to address...