Xbox Series X: A Closer Look at the Technology Powering the Next Generation

chipperrip

Member
Jan 29, 2019
171
You are talking about something different from what that person wrote. The XSX SoC can run four X1S games simultaneously.
Oh, I thought he was talking about quick resume. Where do they talk about simultaneous games that isn't quick resume?

That would be a cool way to splitscreen games that don't have local multiplayer.
 
Sep 19, 2019
1,179
Hamburg- Germany
Oh, I thought he was talking about quick resume. Where do they talk about simultaneous games that isn't quick resume?

That would be a cool way to splitscreen games that don't have local multiplayer.
As far as i know the XSX SOC can run 4 instances of X1S during Xcloud use. Quick resume is something different. i think you guys talk about 2 different things caused by the number 4 that both topics a driven by. :D
 

Scently

Member
Oct 27, 2017
592
Oh, I thought he was talking about quick resume. Where do they talk about simultaneous games that isn't quick resume?

That would be a cool way to splitscreen games that don't have local multiplayer.
Read the DF article. Its mentioned in there. The poster was asking how that is possible given that the XSX SoC only has 16gb of which 13.5 is accessible by a game. And since an X1S game uses up 5gb of ram, the question then becomes how does it run 4 X1S games simultaneously. It should be noted that this is about the XSX SoC on the Xcloud server so its possible that those can have a different memory configuration that the retail units.
 

chipperrip

Member
Jan 29, 2019
171
Read the DF article. Its mentioned in there. The poster was asking how that is possible given that the XSX SoC only has 16gb of which 13.5 is accessible by a game. And since an X1S game uses up 5gb of ram, the question then becomes how does it run 4 X1S games simultaneously. It should be noted that this is about the XSX SoC on the Xcloud server so its possible that those can have a different memory configuration that the retail units.
Thanks, I missed that first time around. Here's the relevant bit from the article.

There are customisations to the CPU core - specifically for security, power and performance, and with 76MB of SRAM across the entire SoC, it's reasonable to assume that the gigantic L3 cache found in desktop Zen 2 chips has been somewhat reduced. The exact same Series X processor is used in the Project Scarlett cloud servers that'll replace the Xbox One S-based xCloud models currenly being used. For this purpose, AMD built in EEC error correction for GDDR6 with no performance penalty (there is actually no such thing as EEC-compatible G6, so AMD and Microsoft are rolling their own solution), while virtualisation features are also included. And this leads us on to our first mic-drop moment: the Series X processor is actually capable of running four Xbox One S game sessions simultaneously on the same chip, and contains an new internal video encoder that is six times as fast as the more latent, external encoder used on current xCloud servers.
 

chipperrip

Member
Jan 29, 2019
171
Yeah that's it but it remains to be seen how its handled with regards to the memory. I suspect the memory configuration on the Xcloud servers will be 10 2gb chips for a total of 20gb. This solves the issue neatly; 5gb for each game instance.
Speculating wishfully for a moment:

If they could implement this on the consumer hardware for even just two concurrent games and have it work with backwards compatibility, that could lead to being able to split screen a huge library that doesn't have official split screen.

That would be a big deal for local multiplayer lovers such as myself.

Of course, there might be any number of modifications or additions to the xCloud hardware required that we aren't aware of for that particular use case of the XSX chip.
 

Colbert

Banned
Oct 27, 2017
6,174
Germany
I'm not a huge fan of cloud dependency for graphical features
The cloud dependency is pre-runtime of a game, not at run-time of a game. DLSS is the same!

AI should be the same.

Best example: Drivatars in Forza. Game play telemetry is used to update the drivatar AI async in the background. The game uses a data set that was available when you started the game. So at runtime that data set is static to your game while you in a race. The AI data only gets updated if a new data set was made available.
 
Last edited:

Axel Stone

Member
Jan 10, 2020
810
Speculating wishfully for a moment:

If they could implement this on the consumer hardware for even just two concurrent games and have it work with backwards compatibility, that could lead to being able to split screen a huge library that doesn't have official split screen.

That would be a big deal for local multiplayer lovers such as myself.

Of course, there might be any number of modifications or additions to the xCloud hardware required that we aren't aware of for that particular use case of the XSX chip.
That does sound pretty amazing to be fair. Halo 5 split screen at last!
 
Jun 18, 2018
886
Speculating wishfully for a moment:

If they could implement this on the consumer hardware for even just two concurrent games and have it work with backwards compatibility, that could lead to being able to split screen a huge library that doesn't have official split screen.

That would be a big deal for local multiplayer lovers such as myself.

Of course, there might be any number of modifications or additions to the xCloud hardware required that we aren't aware of for that particular use case of the XSX chip.
Agreed, this would be amazing!!!
 

PianoBlack

Member
May 24, 2018
699
It's probably just saving an image of the game instance from ram to the drive. Sort of like an emulator save state, or the hibernate (not sleep) function on Windows.

When stored, the game is effectively not taking up any resources except storage.

Their limit of 4 games is likely just an arbitrary choice to avoid reserving too much storage space for this feature.
Huh? It runs four games simultaneously. That's why it's interesting. It's not running one game and putting three into a save state.

Edit sorry missed your other posts.
 

Colbert

Banned
Oct 27, 2017
6,174
Germany
M3rcy THEVOID
Ha ha, I mixed up a commentator with his picture to be the author. Thought it was a contributor.
Sorry LOL

M3rcy
Regarding your first response:
You can't say for certain at what clock the PS5 will run until we see game performances or somebody comes out with profiling data. So nothing in this area is wrong or right. That is why I will not participate in discussions argue about it.
 
Last edited:

EBomb

Member
Oct 25, 2017
389
Speculating wishfully for a moment:

If they could implement this on the consumer hardware for even just two concurrent games and have it work with backwards compatibility, that could lead to being able to split screen a huge library that doesn't have official split screen.

That would be a big deal for local multiplayer lovers such as myself.

Of course, there might be any number of modifications or additions to the xCloud hardware required that we aren't aware of for that particular use case of the XSX chip.
Independent instances in local split screen has some awesome speed run competition possibilities.
 

M3rcy

Member
Oct 27, 2017
662
M3rcy THEVOID
Ha ha, I mixed up a commentator with his picture to be the author. Thought it was a contributor.
Sorry LOL

M3rcy
Regarding your first response:
You can't say for certain at what clock the PS5 will run until we see game performances or somebody comes out with profiling data. So nothing in this area is wrong or right. That is why I will not participate in discussions argue about it.
That's not what I'm referring to. He completley mis-characterized the way boost is used. Not just on PS5, but on desktop as well! It was a complete failure of understanding all around.

Edit - As an example:

For example, in highly dynamic environments where explosions are happening, there is no “boost” to help keep the frames stable. Effectively, developers have to optimize their entire game to defined performance benchmark and if they attempt to render a scene that taxes the system, they have to scale down the scene rather than push the clock speed higher to maintain stability.
That's not what the PS5 does. The PS5 operates at boost clocks by default and then when overloaded, as first option, tries to shift power from the CPU/GPU (depending on which chip is loaded) if it has power to spare. If there isn't sufficient power available to shift from the other part of the chip to prevent the overall chip from consuming too much power, the system drops the CPU and/or GPU clocks to keep the system's power consumption under control.

Put another way, PlayStation developers should be targeting games with the performance characteristics of a 9.2TFlop console and if a scene needs some extra muscle to keep frames stable, the console can briefly kick up the horsepower.
Completely and utterly wrong. Not sure how it could be more incorrect, in fact.
 
Last edited:

Colbert

Banned
Oct 27, 2017
6,174
Germany
That's not what I'm referring to. He completley mis-characterized the way boost is used. Not just on PS5, but on desktop as well! It was a complete failure of understanding all around.

Edit - As an example:



That's not what the PS5 does. The PS5 operates at boost clocks by default and then when overloaded, as first option, tries to shift power from the CPU/GPU (depending on which chip is loaded) if it has power to spare. If there isn't sufficient power available to shift from the other part of the chip to prevent the overall chip from consuming too much power, the system drops the CPU and/or GPU clocks to keep the system's power consumption under control.



Completely and utterly wrong. Not sure how it could be more incorrect, in fact.
The wording in the article is a little bit weird when he talks about the special effect but boost on the PS5 is not working the same way as on a Desktop PC anyway.

And it is even not the same as Smartshift would work on a Laptop.

Cerny himself pointed to the fact, that to make sure that all consoles run with the same predictable performance a dev have to select a profile offered by the SDK which is a target setup of CPU and GPU clock. So the dynamic clock is constrained to a certain range it operates in.

Anyway, this is all kind of off-topic: We are in a Xbox thread.
 

M3rcy

Member
Oct 27, 2017
662
The wording in the article is a little bit weird when he talks about the special effect but boost on the PS5 is not working the same way as on a Desktop PC anyway.

And it is even not the same as Smartshift would work on a Laptop.

Cerny himself pointed to the fact, that to make sure that all consoles run with the same predictable performance a dev have to select a profile offered by the SDK which is a target setup of CPU and GPU clock. So the dynamic clock is constrained to a certain range it operates in.

Anyway, this is all kind of off-topic: We are in a Xbox thread.
Hey, you're the one that introduced the XBSX vs. PS5 article with the embarrassingly bad technical errors in it into the dicussion. :)
 

Colbert

Banned
Oct 27, 2017
6,174
Germany
Hey, you're the one that introduced the XBSX vs. PS5 article with the embarrassingly bad technical errors in it into the dicussion. :)
All I did was introducing another take about how the consoles compare. Another POV. It was not meant to start a deep dive in PS5 system behavior. The PS5 details can be still discussed at the other thread imo, right?
 

M3rcy

Member
Oct 27, 2017
662
All I did was introducing another take about how the consoles compare. Another POV. It was not meant to start a deep dive in PS5 system behavior.
The PS5 details can be still discussed at the other thread imo, right?
I was a bad take. I said so and you challenged me on it. I responded with support for why it was a bad take. I couldn't really do that without explaining what it got so very very wrong.
 

Fredrik

Member
Oct 27, 2017
6,282
Brad: - ”What I am saying is that you can’t compare the TFlop output of next-gen consoles to existing hardware as it’s not the complete picture. For next-gen consoles, it’s a fair comparison but when Microsoft does finally announce Lockhart, don’t lock-in on the raw performance as it’s not the complete story.”

Lol Lockhart will turn into a lovely trainweck here when/if it’s announced. Unless I’m completely out of the loop on Lockhart it’s 100% based on the XSX, but scaled down. So dodging the teraflop difference is every bit as wrong here as doing it with PS5 vs XSX, or even worse. No? Seems like common sense to me.


My highly uneducated take:

Same generation and GPU architecture, teraflops matters. It’s still theroretical but everything is so similar that the theory nevertheless shows itself in framerates.

Different generation and GPU architectures, teraflops matters in theory but rarely in practise. Everything is so different that it’s really just a basic math calculation from shaders and clockspeed.

There must be thousands of GPU tests to verify this from whenever AMD or Nvidia has released a new super duper variant without actually going into a new generation. Same architecture but different CU count and clockspeed etc.
 

Dimajjio

Member
Oct 13, 2019
632
So does that new video encoder in the XSX APU mean that it can be used on your local console for Twitch streaming and xCloud remote play? Potentially higher quality video with lower bandwidth?
 

OneBadMutha

Member
Nov 2, 2017
4,917
Brad: - ”What I am saying is that you can’t compare the TFlop output of next-gen consoles to existing hardware as it’s not the complete picture. For next-gen consoles, it’s a fair comparison but when Microsoft does finally announce Lockhart, don’t lock-in on the raw performance as it’s not the complete story.”

Lol Lockhart will turn into a lovely trainweck here when/if it’s announced. Unless I’m completely out of the loop on Lockhart it’s 100% based on the XSX, but scaled down. So dodging the teraflop difference is every bit as wrong here as doing it with PS5 vs XSX, or even worse. No? Seems like common sense to me.


My highly uneducated take:

Same generation and GPU architecture, teraflops matters. It’s still theroretical but everything is so similar that the theory nevertheless shows itself in framerates.

Different generation and GPU architectures, teraflops matters in theory but rarely in practise. Everything is so different that it’s really just a basic math calculation from shaders and clockspeed.

There must be thousands of GPU tests to verify this from whenever AMD or Nvidia has released a new super duper variant without actually going into a new generation. Same architecture but different CU count and clockspeed etc.
GPU power difference matters to visual fidelity. It will impact quality of resolution, textures, lighting and reflections. That stuff matters to some people more than others. It's not about downplaying anything. Different people have different priorities. It's only important to understand the facts so that people can prioritize with good info.
 

PianoBlack

Member
May 24, 2018
699
Brad: - ”What I am saying is that you can’t compare the TFlop output of next-gen consoles to existing hardware as it’s not the complete picture. For next-gen consoles, it’s a fair comparison but when Microsoft does finally announce Lockhart, don’t lock-in on the raw performance as it’s not the complete story.”

Lol Lockhart will turn into a lovely trainweck here when/if it’s announced. Unless I’m completely out of the loop on Lockhart it’s 100% based on the XSX, but scaled down. So dodging the teraflop difference is every bit as wrong here as doing it with PS5 vs XSX, or even worse. No? Seems like common sense to me.

My highly uneducated take:

Same generation and GPU architecture, teraflops matters. It’s still theroretical but everything is so similar that the theory nevertheless shows itself in framerates.
Agreed that it will be a trainwreck on this forum, but disagree with the way you're framing it.

Speaking very generally, teraflops do matter between PS5 and XSX because they are targeting the same resolutions and framerates. XSX will either produce a better image at same resolution and framerate as PS5, or the same quality of image at a higher resolution/framerate, or some combination thereof. PS5 designers and owners would probably prefer not to give up an advantage on any of these aspects, but the TF difference is a rough indicator that, usually, they will to some degree. (Whether that difference is large enough to care about is another issue entirely.)

Lockhart will explicitly target 1/4 of the resolution of XSX and PS5. So criticizing its teraflop count will make less sense - it is intentionally producing a lower resolution image in hopes of maintaining similar quality and framerate with fewer teraflops, and will be openly sold as such. But this distinction will not be drawn by most posters and many threads will be shitshows :)
 

xem

Member
Oct 31, 2017
564
Agreed that it will be a trainwreck on this forum, but disagree with the way you're framing it.

Speaking very generally, teraflops do matter between PS5 and XSX because they are targeting the same resolutions and framerates. XSX will either produce a better image at same resolution and framerate as PS5, or the same quality of image at a higher resolution/framerate, or some combination thereof. PS5 designers and owners would probably prefer not to give up an advantage on any of these aspects, but the TF difference is a rough indicator that, usually, they will to some degree. (Whether that difference is large enough to care about is another issue entirely.)

Lockhart will explicitly target 1/4 of the resolution of XSX and PS5. So criticizing its teraflop count will make less sense - it is intentionally producing a lower resolution image in hopes of maintaining similar quality and framerate with fewer teraflops, and will be openly sold as such. But this distinction will not be drawn by most posters and many threads will be shitshows :)
It should be criticized based on value. If im getting nearly the same performance and resolution on a cheaper ps5 than XSX then it will be the best value per performance. If lockhart is 299 and gives 1/4 of said resolution, that value is much less in my eyes. Price will dictate all 3 of these consoles IMO.
 

anexanhume

Member
Oct 25, 2017
7,338
It should be criticized based on value. If im getting nearly the same performance and resolution on a cheaper ps5 than XSX then it will be the best value per performance. If lockhart is 299 and gives 1/4 of said resolution, that value is much less in my eyes. Price will dictate all 3 of these consoles IMO.
I think the price sensitive consumers for Lockhart will only care about the price of entry. Otherwise, Pro, X1X would have dominated once they debuted.
 

xem

Member
Oct 31, 2017
564
I think the price sensitive consumers for Lockhart will only care about the price of entry. Otherwise, Pro, X1X would have dominated once they debuted.
Sort of. It all depends on the price tier. If the break down is $100 between each one. 299>399>499 then thats more of reasonable decision but when it was between a ps4 at 199 and pro at 399 around holiday price, then that value is not there IMO. there just isnt enough performance gain for 200 more dollars. But then comparing a 1X to ps4 Pro id argue that performance gain is easily worth the extra 100 with all the back compat enhancements and more native 4k titles avail. Its going to be interesting how the market will shake out. Series X is looking very powerful and i also wonder if they cant squeeze more out of that closer to launch, getting near 13TF.
 

EvilBoris

HDTVtest
Verified
Oct 29, 2017
8,048
It should be criticized based on value. If im getting nearly the same performance and resolution on a cheaper ps5 than XSX then it will be the best value per performance. If lockhart is 299 and gives 1/4 of said resolution, that value is much less in my eyes. Price will dictate all 3 of these consoles IMO.
ah the old $ per pixel calculation.
 

Axel Stone

Member
Jan 10, 2020
810
We don’t know any prices as yet.

FWIW, I do think that if we end up in a $499/$399/$299 situation, then yes, the value proposition is in Sony’s favour.

If it’s, say $449/$399/$249 I think things look quite different.
 

PianoBlack

Member
May 24, 2018
699
I always get every console...but since sony focuses on sp games I'll likley wait until most are out before thinking about it. XSX on the other hand day 1.
Yep. With Game Pass I'm missing out on new games that come all the time, not just MS games. And I'm already subbed anyway. With Sony, I'll just wait until they have enough neat exclusives at $20 for me to buy in. Microsoft has completely changed my game consumption behavior.
 
Oct 27, 2017
3,580
Another opinion piece on how next gen consoles compare. I post this here as Thurrot is usually posting MS related stuff.

Why is Brad Sams making up a "typical" power output scenario with the PS5? We don't actually know what typical would be in a real-world application yet, but going by what Cerney himself has said 10+ TFs seems to be the "typical" power output.
 

Fabtacular

Member
Jul 11, 2019
298
Yep. With Game Pass I'm missing out on new games that come all the time, not just MS games. And I'm already subbed anyway. With Sony, I'll just wait until they have enough neat exclusives at $20 for me to buy in. Microsoft has completely changed my game consumption behavior.
So in another thread I was thinking about a scenario of where a person had decided they would buy both consoles: one at launch, and the other at the mid-gen refresh. Which would be the best order to buy them in?

I like your thought here. Multiplats are probably going to perform better on XSX at least in the short-term, and you'll enjoy all of the benefits of Gamepass for the first-party MS stuff. Then, by the time the PS5 Pro comes out, you'll pay bargain-bin prices on the PS-exclusives you missed, and you'll get to play superior versions of them that take advantage of the Pro power.
 

Fredrik

Member
Oct 27, 2017
6,282
Agreed that it will be a trainwreck on this forum, but disagree with the way you're framing it.

Speaking very generally, teraflops do matter between PS5 and XSX because they are targeting the same resolutions and framerates. XSX will either produce a better image at same resolution and framerate as PS5, or the same quality of image at a higher resolution/framerate, or some combination thereof. PS5 designers and owners would probably prefer not to give up an advantage on any of these aspects, but the TF difference is a rough indicator that, usually, they will to some degree. (Whether that difference is large enough to care about is another issue entirely.)
Sounds reasonable and lines up perfectly with PC GPU tests.
I don’t think a difference might be seen if a game is well-optimized to a locked framerate though, unless you’re Digital Foundry and notice that the magic they did really wasn’t magic at all.
But that’s a whole other thing. It’s the all too common scenario when a game is struggling to maintain a locked framerate that will show the hardware differences the most. That’s when pure hardware specs usually matters.


Lockhart will explicitly target 1/4 of the resolution of XSX and PS5. So criticizing its teraflop count will make less sense - it is intentionally producing a lower resolution image in hopes of maintaining similar quality and framerate with fewer teraflops, and will be openly sold as such.
That sounds good in theory but don’t really make teraflops matter any less, the target resolution is just lower. It happens on Xbox One today, and Switch, even PS4 Pro. And if the strategy is to simply target 1080p instead of 4K, then what happens on Lockhart when a game comes along that can’t quite reach 4K at the targeted framerate on the XSX so they go for a lower resolution?
720p on Lockhart? 20fps? Downgraded visuals?

I just don’t think there is any magic that makes hardware limitations disappear, there is only smoke and mirrors and clever compromises.

In the end I understand that the end user might not care anyway depending on why they bought the cheaper console. But that’s not why I quoted what Brad said, I just found it interesting that the teraflops count would matter between PS5 and XSX but not between XSX and Lockhart. They might actually matter in exactly the same way, just not on the same degree, all depending on how much the higher spec console is pushed and how clever the dev is when it comes to hiding the required compromises on the lower spec console.
DF will have plenty of work next gen that’s for sure! ;)
 

Trup1aya

Member
Oct 25, 2017
7,029
It should be criticized based on value. If im getting nearly the same performance and resolution on a cheaper ps5 than XSX then it will be the best value per performance. If lockhart is 299 and gives 1/4 of said resolution, that value is much less in my eyes. Price will dictate all 3 of these consoles IMO.
This only makes sense if you base your value solely on resolution. Someone who is fine with 1080p might rather spend that $100 on a couple games, or an extra controller. Maybe someone is buying it as a second console, or maybe they prefer the smaller form factor it promises.

I'm not sure how anyone can determine that a $100 savings isn't generally worth the loss in resolution. The ps4 pro and the x1x both get handedly outsold buy their 4k big brothers... with the difference being $100 and 1/4 the pixels
 

EvilBoris

HDTVtest
Verified
Oct 29, 2017
8,048
Resolution is going to not be as a big a metric for performance and image quality this time round, not when VRS is able to drop back and simultaneously maintain resolution
 

JaggedSac

Member
Oct 25, 2017
2,032
Burbs of Atlanta
It should be criticized based on value. If im getting nearly the same performance and resolution on a cheaper ps5 than XSX then it will be the best value per performance. If lockhart is 299 and gives 1/4 of said resolution, that value is much less in my eyes. Price will dictate all 3 of these consoles IMO.
PS5 isn't going to be much cheaper than the XSX, if at all in my opinion.
 

PianoBlack

Member
May 24, 2018
699
That sounds good in theory but don’t really make teraflops matter any less, the target resolution is just lower. It happens on Xbox One today, and Switch, even PS4 Pro. And if the strategy is to simply target 1080p instead of 4K, then what happens on Lockhart when a game comes along that can’t quite reach 4K at the targeted framerate on the XSX so they go for a lower resolution?
720p on Lockhart? 20fps? Downgraded visuals?
Sure, why not? That's the whole point of the system - you pay less for worse visuals. But nobody will be targeting less than 1440p on XSX and PS5 so this isn't a huge concern, IMO. There will be plenty of room to reduce resolution, plus reconstruction techniques are quite good.

Also I sincerely doubt we see 20FPS specifically given the fast CPU in Lockhart. 720p and downgraded effects are much more likely.

I just don’t think there is any magic that makes hardware limitations disappear, there is only smoke and mirrors and clever compromises.
I don't understand what you think is smoke and mirrors. Lockhart is targeting 1/4 resolution with 1/3 the GPU power. In general I'd expect that to work out fine. I'm sure some other effects will be reduced here and there (e.g. ray tracing) but overall it's the same architecture as XSX.

In the end I understand that the end user might not care anyway depending on why they bought the cheaper console. But that’s not why I quoted what Brad said, I just found it interesting that the teraflops count would matter between PS5 and XSX but not between XSX and Lockhart. They might actually matter in exactly the same way, just not on the same degree, all depending on how much the higher spec console is pushed and how clever the dev is when it comes to hiding the required compromises on the lower spec console.
DF will have plenty of work next gen that’s for sure! ;)
Again, I don't think it's exactly the same, because PS5 is not being positioned as a value machine that will cost less but offer less visual quality.
 

Morgan J

Member
Oct 26, 2017
1,217
I always get every console...but since sony focuses on sp games I'll likley wait until most are out before thinking about it. XSX on the other hand day 1.
I'm doing the same as you.
Worldwide, absolutely, but in the markets that MS is strong in they could be a lot more competitive with Sony this gen.
If Microsoft gets more momentum after release, then absolutely!
Sure, but that doesn't necessarily mean the overall console marketshare won't increase for MS compared to this gen.
I'm not arguing against that, just that Sony has a wide fanbase. But I do MS gets a higher piece of the market.