SADE, All Access, XCloud, and Game Pass. They've all been experiments for future revenue streams for Xbox, and the hardware affects all of them from the datacenter to the home, to subscriptions to SKUs.
-
Ever wanted an RSS feed of all your favorite gaming news sites? Go check out our new Gaming Headlines feed! Read more about it here.
Next-gen PS5 and next Xbox speculation launch thread |OT7| - nm
- Thread starter Mecha Meister
- Start date
- Status
Interesting. Thanks for your contributions so far.. is there a possibility we get Scarlet All digital edition at some point especially with their investment into Game Pass we should see MS make even more moves in that space.
I honestly think an all digital version of Scarlett down the line is a sure lock. They will try again in few years to see how market responds.
I think that was the thinking behind Lockhart. They have some interesting decisions to make that impact far more than just the consoles they sell. Microsoft also doesn't care about console numbers now. They made a strategic decision to link their console work to their cloud work to target all devices, but how long do you keep Xbox One S parts powering your datacenters? Do you then upgrade all of these to Anaconda parts, or is there some middle ground (Lockhart) that's based on the same CPU that provides different levels of xCloud access? Spencer has said XCloud will move to Scarlett, but we don't know when. You could say move to Xbox One X parts, but that has CPU constraints that will limit next-gen games. There's a lot to consider when your biggest competitors are starting from scratch (Google / Amazon).
I generally wonder about the sustainability of gaming servers. Can these servers be used for different things once the tech moved on? Will they go into trash? This also applies to Google it they drop stadia down the line because it doesn't move anything in terms of subscriptions.
It's a little bit off topic but I'd like to here if the companies shared some information on that. It's even worse for Sony when they have their Cell tech sitting in data centers and nobody can do anything with that tech.
Thanks for the reply. Its really getting interesting. Can't wait to see what everyone brings to the table next gen. Sony , MS and probably Nintendo.. Next gen might be one of the most interesting we've had in a while.
I really hope so, I bought only two physical game this whole generation
Yeah I'm completely down for an all digital Scarlett at a lower price
Not like anyone will listen to me but the information Klee gave is accurate. Also the differences between machines is within 10 percent. Last - Digital Foundry and pretty much everyone alive will be focusing on the difference in SSD performance.
Just a *thought.
Just a *thought.
At the end of the day it won't even matter, its going to be nothing like the pro to 1x difference which was not really a big deal.
You know what you were doing with the way you worded the first post, please stop that.
Oh it'll matter to THIS thread. We've spent over a year at each other's throats.
We have fallen warriors-TruthofKayle, SpinningBirdKick, and others. This is at this point about one thing and one thing only: bragging rights.
Lol no one will care put of a few warriors. The difference is so low, it means multiplatform game will be made like it is the same machine.
Some people will be happy because one piece of machine goes slightly faster than the other.
Imagine arguing which piece of plastic will be stronger and hoping for one of the companies to screw up when we're getting The Lion King graphics in real time either way.
People who do "brag" will be revealed as nothing more then fan folk, because this gen both platforms will have been on the weaker side, so if they brag about power they will have to acknowledge the weakness of there preferred platform last gen.
And you will probably be able to find contradicting statements on post history E.g "ps4 Pro/x1 power difference don't matter its the games etc" then they be like "haha ps5/scarlett is the 🐐 cos its more powerful"
This is not a good way for people to be.
Basically we need to rejoice both new consoles regardless of there power levels and be like:
It's impossible. A Lockheart with a third of the GPU perf as rumoured cannot render the same game without major patching of not just the game renderer code but also re-worked assets (e.g. LOD models, textures) appropriate for the lower resolution platform.
All this stuff is manual and requires coders and game artists to do manual work in view of a patch for their games.
*cracks his layman fingers*
With VRS you can decouple the Rasterization-Rate (How fine your pixelgrid is) from the Shading-Rate (shading/coloring the pixelgrid).
That's the difference to how we do it now where we scale both at the same time, either up or down:
Now as far as I understand it you wouldn't get more sub-pixel polygons with VRS since the amount of raw pixels per polygon is still the same but you can save shading costs when you let the pixel shader run for multiple raw pixels at once instead of each individually.
Overshading with small triangles is an issue, potentially with VRS you could save a lot but the question is at what visual cost?
If you need precise shading for small details you might not want to reduce the shading rate there.
The classic rasterization/shading pipeline is quite complex from a layman's perspective.
You have fixed function hardware for geometry generation but programmable shaders are also involved in the process.
Triangle culling can be done at different points with various precision and efficiency.
There is fixed function HW which does the rasterization of the scene, also with varying granularity, pixel shaders are actually doing the shading processing in pixel-quads (2x2) and not individual pixels (with VRS it get's more complex to understand) since that way you get derivatives for texture filtering, you have HW for Coverage/Z/Stencil-Testing of the pixel-quad-packages, you have to follow the submission order to guarentee correct results, some stuff in the pipeline is getting serialized because of it.
I'm not in the know if and how much VRS and Mesh Shaders ultimately could help to improve the efficiency of the whole pipeline if micro polys are the main issue.
I would guess that VRS won't help much here but Mesh Shaders and feeding the machine only with visible triangles and with knowing the triangle topology you could shade meshes with far less unnecessary computations.
But performance is of course a general resource.
If you do stuff more efficiently you can also spent more on other areas or afford more of sub-optimal rendering for better visuals.
I know my posting boils down to basically this:
But there is google and I know that there is at least one (older) patent from AMD about improving the performance in the case of micro polygon rasterization, if you really want to know about the bottlenecks and potential improvements then there might be some answers.
I have more or less a list of 20 patents I want to read but no the time for all of it.
Bookmarked!
Cheers, Locuza. I was looking forward to this. It was indeed illuminating and I would be interested to get that patent reading list too. Would make for some good bedtime reading 😄
Posting speculation as fact or insider info without qualification is frowned upon in here and can in the worst case net you a ban.
So either state clearly that the above posts are your own speculation or if not where you heard said info, or you'll likely be reported to a mod.
Are you a mod?
Obviously I am joking, many people here do. You need to calm down.
Reminds me of Gies, unsurprisingly from Verge's sister site. Wonder if they have the same source.
Universe brain tinfoil hat theory: what if Gies IS his source?
Last edited:
I bet u will be all ready with your goat emojis if ps5 is more powerful 😉
It is funny that people just assume that such a small hardware difference in TF specs on paper and current dev kit specs automatically translates one for one into real world game performance difference such as Xbox Scarlett may drop an extra frame or two here or there is all. We literally have no idea how these retail consoles will actually perform with shipping games that have been optimized on final dev kits and such. Klee's source is a 3rd party developer working on a game that is out there a good ways too, so they very likely don't even have their own game remotely optimized regardless of considering these consoles at this point.
As everyone always rightly points out these are closed systems where software and API optimizations can make a difference more than PC. I am not talking about Xbox suddenly having secret sauce and magic that gives a large performance delta, rather it is entirely plausible that you could still end up with some or many games being slightly more consistent in frame rate or something on Scarlett for whatever reason even if final hardware tweaks such as clocks don't change the small difference much. The simple fact is that we are a long way from having any clue about what Digital Foundry or anyone else will be saying about specific 3rd party game performance on these consoles at launch.
As everyone always rightly points out these are closed systems where software and API optimizations can make a difference more than PC. I am not talking about Xbox suddenly having secret sauce and magic that gives a large performance delta, rather it is entirely plausible that you could still end up with some or many games being slightly more consistent in frame rate or something on Scarlett for whatever reason even if final hardware tweaks such as clocks don't change the small difference much. The simple fact is that we are a long way from having any clue about what Digital Foundry or anyone else will be saying about specific 3rd party game performance on these consoles at launch.
Imagine arguing which piece of plastic will be stronger and hoping for one of the companies to screw up when we're getting The Lion King graphics in real time either way.
You are so insisting on it that it may really happen some day. :p
API \ software optimizations don't stop even when the consoles launch .
This goes for both MS and Sony so people just going by what they think the hardware difference will bring .
Plus you know devs will come up with new stuff during the gen like they always do to try and get more out the systems.
Last edited:
One thing that is interesting is that if Sony do go for uber powerful $499, it's certainly a change in strategy. One can't really blame ms if they are slightly less powerful on a $499 console.
This is true, particularly if the hardware is really so close such as less than 10% difference in raw TF or something.
User Warned: Hostility
You know what they say, if you believe hard enough anything is possible. Just look at bcatwilly above, I admire his dedication for a brand
If it's anything like last gen, he likely had something DM'ed to him, but it's probably 3rd hand and unverifiable.
Change in what strategy? Sony had the most powerful console this gen as well.
It clearly wasn't obvious, which is why I tried to warn you to clarify your post. You knew exactly what you were doing.
But it's no skin off my back. By all means continue posting like that and see how long you last.
Tech moves on for all the hardware, data centers or not, and GPU instances aren't some special unique snowflake. Most of the time data-centers are not at the cutting edge of hw that exists on the market anyway.
For most part that's factored into the extortionary pricing of cloud-compute, which is one of the reasons that make things like 'multi-instance games-per-user' completely unaffordable unless you're one of those people willing to pay a LOT more for your gaming time then you do today.
The issue at the heart of this is that as long as local-compute keeps advancing at the rates it has, cloud will not be cost-competitive for the use cases we talk about here, as you will always have local options that either perform better, or do it for (significantly)less.
GPUs are flexible enough now to allow generating command/display-lists without CPU intervention, but these kind of pipelines have their own trade-offs. As for LOD management, it also comes down to a lot more than draw-call counts, and available I/O bandwidth is one of the sources of visible popup artifacts.
No.
VRS respects polygon boundaries (being MSAA derived concept) so it's even the other way around - micro-polygons will further ruin what benefits you get from it.
What you could look for is for polygon-size/density distribution that is more optimal to the underlying rasterization hardware (which MeshShading 'could' have a hand in) but fundamental inefficiencies of micro-polygons idling parts of the hardware remain. Worth noting that MeshShading-like processing has been a viable option on current gen consoles, so it's not that much of a step forward either.
DF has shown a very good aptitude to adapt to narrower scope of differences to analyze(as does their core-audience, let's be honest), I think they'll still find multiple things to talk about even with the gap reducing.
There is no $399 strategy, it's about offering a price point that's acceptable by the market in regards to the specs, $499 for what the next gen consoles will have is more than fine.
Sure, posting logical things about how we don't have a clue what real world game performance will be for these consoles at launch is definitely brand promotion. I mean come on people, Klee himself recently made the same type of post about software and API optimizations by Xbox that could make a possible difference.
Truth is for first few years they talk about new tech , engine changes and what may be possible later on .
So they cover for a good while even if the gap close.
From what Tom Warren said about the purpose of the dev kits shape , it looks I was right with my cooling concept sketch weeks ago.. like beginning of October ago..
For some reason I cannot quote myself there so I'll copy it here again..
Okay since the topic with those dev kits came up again, i allow myself to post something here what i posted on reddit a couple weeks ago:
For some reason I cannot quote myself there so I'll copy it here again..
Okay since the topic with those dev kits came up again, i allow myself to post something here what i posted on reddit a couple weeks ago:
So i made up a crude Sketch of that PS5 Dev Kit internals using one of those Patents
I think that, as you can see, that those both "Arms" of that "V" are actually 2 Heatsinks. I assume here that on the outside are 1-3 Fans wich suck cool Air through that heatsink. Both Heatsinks are connected with the APU by heatpipes . To me , it seems (if iam right)to be a smart solution. It completly sepaerates the PS5 into 2 Areas one side wich produce Heat and one side that gets rid of it. Also it is a very clear Language in terms of Airflow.Like : Cool from front and top. Warm to sides and back. Very clean, very no bullshit approach.That pink line i put into to indicate roughly the hight of a vanilla PS4. Wich suggests that all that funny upper part of the body are only there for the cooling solution. One could cool a PS4 with such a Method , basicly double its height.
Wich brings me right to the Final Release PS5. People look at that Patent Sketch an be like : ,,I never put such a thing under my TV"But what People do not understand is that this Dev kit Design can be transformed pretty easily into a release PS5.
Imagine the front with less USB Ports and no funny lights above the Disk insert. Then remove in your mind those scary looking vents on the sides and use more moderate looking ones. The next step is to cover those 2 openings of the "V" in front and on top with some kind of air pervious plastic grid. And there you go, the coming PS5 could actually stay with the same design Language as a PS4 Pro.So what you are thinking - me right or wrong :D
PS - i hope i made not too many grammar mistakes since iam a non native english speaker
Offering a new gen console at 2 different price points with $100 between is a different stratagy.
I agree $499 is more then fine, but it is a different strategy.
A 8tflop, 8core Zen2, 16gb gddr6 and 1tb ssd @ $399
Is a different proposition then
A 11tflop, 8core Zen2, 20gb ram 1-2tb ssd @$499
Yes for what I read from dev mesh shading like was done using compute shader current gen.
I have a feeling that some of the Digital Foundry comparisons could end up with observations such as "The PS5 does seem to load this area of the game a little faster than the Xbox" and "The frame rate is a little more consistent on the Xbox with ray tracing enabled, particularly in these certain areas of the game".
So basically, stuff that's meaningless and won't matter when we're fighting a dragon in Rockstar's historically accurate medieval game
- Status