Support NeoGAF

[Coulomb_Barrier]

Well this confirm VRS, I can swear Cerny sometimes is not bes spokesperson mentioned so many things but was not able to even put in a page of the presentation things like that.

Be careful the trolls will come in here and say 'but that guy used to work for Sony so his tweet is not proof of anything, we only take tweets from Xbox employees seriously' lol.

 

[VanSiriusGabriel]

Not 6gb/s. 4.8 gb/s. At least be honest with the specs.

22 GB\s sounds even more honest.

 

[Metroiddarks]

Personally, I hope Halo is going to be a good game, want to try it on PC. Let's hope both will deliver on that front.

Man I need a good Halo .... I want to forget that Halo 5

 

[Coulomb_Barrier]

Oh Series X fans are not going to like this tweet:

"Wtf is a "6tflop console", you're just making this up. Teraflop measures one part of the pipeline of one discreet component of a console. You don't call your car a "12 gallon fuel tank saloon" like that means anything"

Matt Phillips, ex-Crytek, Travellers Tales programmer.

Twitter Publish

publish.twitter.com

 

[Bo_Hazem]

Man I need a good Halo .... I want to forget that Halo 5

I only "stained" my hands with an XboX 360 playing Assassin's Creed 1 of my cousin and friend (God bless his soul). PS3 didn't hit the market as quick, remember buying it at late 2007.

But, I wanna game at least for once on my PC, didn't feel anything worth wasting my time on (personal opinion).

 

[Metroiddarks]

Be careful the trolls will come in here and say 'but that guy used to work for Sony so his tweet is not proof of anything, we only take tweets from Xbox employees seriously' lol.

Usually I will use that fact (work for SIE) a way to see why he is so happy but this confirms a couple of things:

-The GPU of PS5 has all features of RDNA and Sony decides to move mesh shaders to a separate chip and rename it because for some strange reason the company always has a problem to share a name of some technology with its competition.
-Cernys presentation was horrible

 

[geordiemp]

Be careful the trolls will come in here and say 'but that guy used to work for Sony so his tweet is not proof of anything, we only take tweets from Xbox employees seriously' lol.

Here is a post from activision, maybe they are not so far apart...., the plot thickens.

 

[Shio]

If we compare raw numbers thn its very easy to compare and speculate to as which one is faster or better but we don't have the full picture. We are wrongly assuming every ounce of power or cycle will be used for the game but this is not so. Just one extra bit of info can change the whole picture. Xbox Series X uses one whole core which is 2 threads for the OS and also 2.5gb of RAM for the OS, we don't know what sony has done, if they had other ram and didnt take one core it would complicate the comparison. This is why as time gets closer to reveal and more of more smaller infos are revealed it will get very interesting. For me it would be amazing if sony said we have a seperate ram and chip for our OS and the rest is solely for the games.

 

[VanSiriusGabriel]

SSD is irrelevant, MS will likely release an Arcade sku of XsX with no SSD. This is the touted Lockheart model. You’ll be able to fit one as an option by plugging it into the provided slot. That’s how much more power the 1.8TF provides, it’s a game changer, SSD is just the new power of the cloud, mark my words!!!

I got their plan.
30 TF Lockheart with HDD confirmed.

It doesn't have to be more important to graphical fidelity, res etc. While it does eliminate most, if not all potential bottlenecks in those areas that can arise from storage, its main advantages lay in the fundamental functionality of the game world and the possibilities therein. Why are we stuck in this one track mind that it HAS to be important to just graphics? And I say that as a total graphics whore.

To quote a really recent post of mine...

"Both are great systems but they each had their priorities and Sony simply went further in this regard [SSD], it seems they've reached a threshold where there will be an even greater array of gameplay possibilities, particularly without compromising complexity.

When you can effectively teleport between two distinct, detailed environments in a second or load in new terrain faster than you can effectively traverse it, then the functional possibilities in a game environment become virtually limitless. Imagination and manpower will become the limit."

^ That, to me, is a far more valuable proposition than real world performance differential that will amount to ~10% (think ~2034p vs ~2160p).

Also, yes, the additional SSD grunt won't bring a great deal more to multiplatform games. But why can't they innovate for the sake of their first party games? It's not like they're thin on the ground or a load of stinkers.

Exactly what I said before. No proof here, but to me PS5 pushes more what hasn't been upgraded since a lot of time, delivering the most difference potentially, which is what should define next gen.

You may say optimistic, but many devs are sharing the same idea. Won't be one system better at all, but it'll depend on every game and its variations.

Why would I be optimistic? It's the same technique done on Horizon Zero Dawn but on steroids. And the gap between 12>10? Seems you missed it, but here:

2080 (10.1 TF) vs 2080 Super (11.9 TF) at 4K with same settings, watch the fps.





1-5fps with no optimization for locked 60fps and with the same memory, no Kraken magic, no ultra fast SSD on one side. It's not even funny to talk about it.

I always liked you because you're smart and like my same box.

It’s pro vs X all over again with Series X coming out on top. That’s fast vs wide design, it is what it is . It’s a lotus sports car philosophy vs a Bugatti Veyron ... everyone knows which one wins in pure power.

But carry on .

It’s clear to me PS5 was designed for 399 in mind with beter performance for 399 and a 2019 release date. And Cerny did a excellent job with these guides in place. To bad component pricing got in the way.

Xbox is best all around.

Maybe Xbox is better all around but the 36 GPU isn't enough to justify what you say, we didn't have anything on the other components, this PS5 one year ago at 399 even a bit inferior would have been a dream, because it wouldn't have 3000 series right at the gate.
A top tier PC in practice for 399? Yeah not happening. Also no point in going cheap without MS, it would have probably been 499 and it wouldn't be too much.
Even now I don't expect 399, it would be crazy.

 

[Metroiddarks]

Oh Series X fans are not going to like this tweet:

"Wtf is a "6tflop console", you're just making this up. Teraflop measures one part of the pipeline of one discreet component of a console. You don't call your car a "12 gallon fuel tank saloon" like that means anything"

Matt Phillips, ex-Crytek, Travellers Tales programmer.

Twitter Publish

publish.twitter.com

That is like in PS3 era the play fans love to says his CELL processor was so powerfull and I was yeah but is hard to use and you know a cpu is no a gpu right ? jajaja

I am still sure XSX will have a little better graphics maybe the difference will not so notificiable except maybe when use many rays in scene that well is a different story.

 

[Metroiddarks]

I only "stained" my hands with an XboX 360 playing Assassin's Creed 1 of my cousin and friend (God bless his soul). PS3 didn't hit the market as quick, remember buying it at late 2007.

But, I wanna game at least for once on my PC, didn't feel anything wasting my time on (personal opinion).

I have the special edition of xbox one of halo 5 ....

 

[pasterpl]

People tend to forget about this I regards to xsex;

So if a game never had to load pages that are ultimately never actually used, that means a 2-3x multiplier on the effective amount of physical memory, and a 2-3x multiplier on our effective IO performance.

A technique called Sampler Feedback Streaming - SFS - was built to more closely marry the memory demands of the GPU, intelligently loading in the texture mip data that's actually required with the guarantee of a lower quality mip available if the higher quality version isn't readily available, stopping GPU stalls and frame-time spikes. Bespoke hardware within the GPU is available to smooth the transition between mips, on the off-chance that the higher quality texture arrives a frame or two later. Microsoft considers these aspects of the Velocity Architecture to be a genuine game-changer, adding a multiplier to how physical memory is utilised.

Our second component is a high-speed hardware decompression block that can deliver over 6GB/s,

I think that until mid gen refresh xbsex will be the better console to play games, it takes time for devs to fully understand new architecture. And then when the xbsex ssd becomes bottleneck we will get a pro version of both consoles.

we need to remember that developers will need time to switch from hdd mindset to ssd and then move towards fully utilising available speeds.

 

[geordiemp]

People tend to forget about this I regards to xsex;





I think that until mid gen refresh xbsex will be the better console to play games, it takes time for devs to fully understand new architecture. And then when the xbsex ssd becomes bottleneck we will get a pro version of both consoles.

we need to remember that developers will need time to switch from hdd mindset to ssd and then move towards fully utilising available speeds.

Looks like 3rd party devs have already started the SSD journey....,,,maybe you need to rethink what is going to happen ? SSD generation begins......GE is the Ps5 version of VRS. by the way.

 

[Bo_Hazem]

I always liked you because you're smart and like my same box.

 

[MT231290]

There's a lot of confusion on why SSD is so important for next-gen and how it will change things.
Here I will try to explain the main concepts.
TL;DR fast SSD is a game changing feature, this generation will be fun to watch!

It was working fine before, why do we even need that?
No, it wasn't fine, it was a giant PITA for anything other than small multiplayer maps or fighting games.
Let's talk some numbers. Unfortunately not many games have ever published their RAM pools and asset pools to the public, but some did.
Enter Killzone: Shadowfall Demo presentation.
We have roughly the following:

Type Approx. Size, % Approx. Size, MB
Textures 30% 1400
CPU working set 15% 700
GPU working set 25% 1200
Streaming pool 10% 500
Sounds 10% 450
Meshes 10% 450
Animations/Particles 1% 45

*These numbers are rounded sums of various much more detailed numbers presented in the article above.

We are interested in the "streaming pool" number here (but we will talk about others too)
We have ~500MB of data that is loaded as the demo progresses, on the fly.
The whole chunk of data that the game samples from (for that streaming process) is 1600MB.
The load speed of PS4 drive is (compressed data) <50MB/sec (uncompressed is <20MB/sec), i.e. it will take >30sec to load that at least.

It seems like it's not that big of a problem, and indeed for demo it is. But what about the game?
The game size is ~40GB, you have 6.5GB of usable RAM, you cannot load the whole game, even if you tried.
So what's left? We can either stream things in, or do a loading screen between each new section.
Let's try the easier approach: do a loading screen
We have 6.5GB of RAM, and the resident set is ~2GB from the table above (GPU + CPU working set). We need to load 4.5GB each time. It's 90 seconds, pretty annoying, but it's the best case. Any time you need to load things not sequentially, you will need to seek the drive and the time will increase.
You can't go back, as it will re-load things and - another loading screen.
You can't use more than 4.5GB assets in your whole gaming section, or you will need another loading screen.
It gets even more ridiculous if your levels are dynamic: left an item in previous zone? Load time will increase (item is not built into the gaming world, we load the world, then we seek for each item/item group on disk).
Remember Skyrim? Loading into each house? That's what will happen.
So, loading screens are easy, but if your game is not a linear, static, theme-park style attraction it gets ridiculous pretty fast.

How to we stream then?
We have a chunk of memory (remember 500Mb) that's reserved for streaming things from disk.
With our 50MB/sec speed we fill it up each 10 sec.
So, each 10 sec we can have a totally new data in RAM.
Let's do some metrics, for example: how much new shit we can show to the player in 1 min? Easy: 6*500 = 3GB
How much old shit player sees each minute? Easy again: 1400+450+450+45=~ 2.5GB
So we have a roughly 50/50 old to new shit on screen.
Reused monsters? assets? textures? NPCs? you name it. You have the 50/50 going on.

But PS4 has 6.5GB of RAM, we used only 4.5GB till now, what about other 2GB?
Excellent question!
The answer is: it goes to the old shit. Because if we increase the streaming buffer to 1.5GB it still does nothing to the 50MB/sec speed.
With the full 6.5GB we get to 6GB old vs 3GB new in 1 minute. Which is 2:1 old shit wins.

But what about 10 minutes?
Good, good. Here we go!
In 10 min we can get to 30GB new shit vs 6GB old.
And that's, my friends, how the games worked last gen.
You're as a player were introduced to the new gaming moments very gradually.
Or, there were some tricks they used: open doors animation.
Remember Uncharted with all the "let's open that heavy door for 15sec?" that's because new shit needs to load, players need to get to a new location, but we cannot load it fast.

So, what about SSDs then?
We will answer that later.
Let's ask something else.

What about 4K?
With 4K "GPU working set" will grow 4x, at least.
We are looking at 1200*4 = 4.8GB of GPU data.
CPU working set will also grow (everybody wants these better scripts and physics I presume?) but probably 2x only, to 700*2 = ~1.5GB
So overall the persistent memory will be well over 6GB, let's say 6.5GB.
That leaves us with ~5GB of free RAM in XSeX and ~8GB for PS5.

Stop, stop! Why PS5 has more RAM suddenly?
That's simple.
XSeX RAM is divided into two pools (logically, physically it's the same RAM): 10GB and 3.5GB.
GPU working set must use the 10GB pool (it's the memory set that absolutely needs the fast bandwidth).
So 10 - 4.8 = 5.2 which is ~5GB
CPU working set will use 3.5GB pool and we will have a spare 2GB there for other things.
We may load some low freq data there, like streaming meshes and stuff, but it will hard to use in each frame: accessing that data too frequently will lower the whole system bandwidth to 336Mb/sec.
That's why MSFT calls the 10GB pool "GPU optimal".

But what about PS5? It also has some RAM reserved for the system? It should be ~14GB usable!
Nope, sorry.
PS5 has a 5.5GB/sec flash drive. That typically loads 2GB in 0.27 sec. It's write speed is lower, but not less than 5.5GB/sec raw.
What PS5 can do, and I would be pretty surprised if Sony won't do it. Is to save the system image to the disk while the game is playing.
And thus give almost full 16GB of RAM to the game.
2GB system image will load into RAM in <1 sec (save 2GB game data to disk in 0.6 sec + load system from disk 0.3 sec). Why keep it resident?
But I'm on the safe side here. So it's ~14.5GB usable for PS5.

Hmm, essentially MSFT can do that too?
Yep, they can. The speeds will be less sexy but not more than ~3sec, I think.
Why don't they do it? Probably they rely on OS constantly running on the background for all the services it provides.
That's why I gave Sony 14.5GB.
But I have hard time understanding why 2.5GB is needed, all the background services can run on a much smaller RAM footprint just fine, and UI stuff can load on-demand.

Can we talk about SSD for games now?
Yup.
So, let's get to the numbers again.
For XSeX ~5GB of "free" RAM we can divide it into 2 parts: resident and streaming.
Why two? Because typically you cannot load shit into frame while frame is rendering.
GPU is so fast, that each time you ask GPU "what exact memory location are you reading now?" will slow it down to give you an answer.

But can you load things into other part while the first one is rendering?
Absolutely. You can switch "resident" and "streaming" part as much as you like, if it's fast enough.
Anyway, we got to 50/50 of "new shit" to "old shit" inside 1 second now!
2.5GB of resident + 2.5GB of streaming pool and it takes XSeX just 1 sec to completely reload the streaming part!
In 1 min we have 60:1 of new/old ratio!
Nice!

What about PS5 then? Is it just 2x faster and that's it?
Not really.
The whole 8GB of the RAM we have "free" can be a "streaming pool" on PS5.

But you said "we cannot load while frame is rendering"?
In XSeX, yes.
But in PS5 we have GPU cache scrubbers.
This is a piece of silicon inside the GPU that will reload our assets on the fly while GPU is rendering the frame.
It has full access to where and what GPU is reading right now (it's all in the GPU cache, hence "cache scrubber")
It will also never invalidate the whole cache (which can still lead to GPU "stall") but reload exactly the data that changed (I hope you've listened to that part of Cerny's talk very closely).

But it's free RAM size doesn't really matter, we still have 2:1 of old/new in one frame, because SSD is only 2x faster?
Yes, and no.
We do have only 2x faster rates (although the max rates are much higher for PS5: 22GB/sec vs 6GB/sec)
But the thing is, GPU can render from 8GB of game data. And XSeX - only from 2.5GB, do you remember that we cannot render from the "streaming" part while it loads?
So in any given scene, potentially, PS5 can have 2x to 3x more details/textures/assets than XSeX.
Yes, XSeX will render it faster, higher FPS or higher frame-buffer resolution (not both, perf difference is too low).
But the scene itself will be less detailed, have less artwork.

OMG, can MSFT do something about it?
Of course they will, and they do!
What are the XSeX advantages? More ALU power (FLOPS) more RT power, more CPU power.
What MSFT will do: rely heavily on this power advantage instead of the artwork: more procedural stuff, more ALU used for physics simulation (remember, RT and lighting is a physics simulation too, after all).
More compute and more complex shaders.

So what will be the end result?
It's pretty simple.
PS5: relies on more artwork and pushing more data through the system. Potentially 2x performance in that.
XSeX: relies more on in-frame calculations, procedural. Potentially 30% performance in that.
Who will win: dunno. There are pros and cons for each.
It will be a fun generation indeed. Much more fun than the previous one, for sure.

 

[Metroiddarks]

People tend to forget about this I regards to xsex;





I think that until mid gen refresh xbsex will be the better console to play games, it takes time for devs to fully understand new architecture. And then when the xbsex ssd becomes bottleneck we will get a pro version of both consoles.

we need to remember that developers will need time to switch from hdd mindset to ssd and then move towards fully utilising available speeds.

Yes and not, yes is XSX is very easy to learn and use those new features yes why not but also needs to have good sales.

If the PS5 has better sales or if its features is easier well they use first the features of that console.

Finally, if that compression tech of XSX is only software at not is propertary of a third company which you need to pay for use , then is not exactly something unique as you can have it also at least you need some specific hardware to do it correctly.

 

[M-V2]

Ok so I see some laughable statements, "the difference is in Xbox SX favor" when the difference is 16%

But

"SSD is exactly the same you won't see a difference but 2sec in loading times" when the SSD is 2X faster (120+%) on ps5

I mean quit the nonsense lol

 

[VanSiriusGabriel]

There's a lot of confusion on why SSD is so important for next-gen and how it will change things.
Here I will try to explain the main concepts.
TL;DR fast SSD is a game changing feature, this generation will be fun to watch!

It was working fine before, why do we even need that?
No, it wasn't fine, it was a giant PITA for anything other than small multiplayer maps or fighting games.
Let's talk some numbers. Unfortunately not many games have ever published their RAM pools and asset pools to the public, but some did.
Enter Killzone: Shadowfall Demo presentation.
We have roughly the following:

Type Approx. Size, % Approx. Size, MB
Textures 30% 1400
CPU working set 15% 700
GPU working set 25% 1200
Streaming pool 10% 500
Sounds 10% 450
Meshes 10% 450
Animations/Particles 1% 45

*These numbers are rounded sums of various much more detailed numbers presented in the article above.

We are interested in the "streaming pool" number here (but we will talk about others too)
We have ~500MB of data that is loaded as the demo progresses, on the fly.
The whole chunk of data that the game samples from (for that streaming process) is 1600MB.
The load speed of PS4 drive is (compressed data) <50MB/sec (uncompressed is <20MB/sec), i.e. it will take >30sec to load that at least.

It seems like it's not that big of a problem, and indeed for demo it is. But what about the game?
The game size is ~40GB, you have 6.5GB of usable RAM, you cannot load the whole game, even if you tried.
So what's left? We can either stream things in, or do a loading screen between each new section.
Let's try the easier approach: do a loading screen
We have 6.5GB of RAM, and the resident set is ~2GB from the table above (GPU + CPU working set). We need to load 4.5GB each time. It's 90 seconds, pretty annoying, but it's the best case. Any time you need to load things not sequentially, you will need to seek the drive and the time will increase.
You can't go back, as it will re-load things and - another loading screen.
You can't use more than 4.5GB assets in your whole gaming section, or you will need another loading screen.
It gets even more ridiculous if your levels are dynamic: left an item in previous zone? Load time will increase (item is not built into the gaming world, we load the world, then we seek for each item/item group on disk).
Remember Skyrim? Loading into each house? That's what will happen.
So, loading screens are easy, but if your game is not a linear, static, theme-park style attraction it gets ridiculous pretty fast.

How to we stream then?
We have a chunk of memory (remember 500Mb) that's reserved for streaming things from disk.
With our 50MB/sec speed we fill it up each 10 sec.
So, each 10 sec we can have a totally new data in RAM.
Let's do some metrics, for example: how much new shit we can show to the player in 1 min? Easy: 6*500 = 3GB
How much old shit player sees each minute? Easy again: 1400+450+450+45=~ 2.5GB
So we have a roughly 50/50 old to new shit on screen.
Reused monsters? assets? textures? NPCs? you name it. You have the 50/50 going on.

But PS4 has 6.5GB of RAM, we used only 4.5GB till now, what about other 2GB?
Excellent question!
The answer is: it goes to the old shit. Because if we increase the streaming buffer to 1.5GB it still does nothing to the 50MB/sec speed.
With the full 6.5GB we get to 6GB old vs 3GB new in 1 minute. Which is 2:1 old shit wins.

But what about 10 minutes?
Good, good. Here we go!
In 10 min we can get to 30GB new shit vs 6GB old.
And that's, my friends, how the games worked last gen.
You're as a player were introduced to the new gaming moments very gradually.
Or, there were some tricks they used: open doors animation.
Remember Uncharted with all the "let's open that heavy door for 15sec?" that's because new shit needs to load, players need to get to a new location, but we cannot load it fast.

So, what about SSDs then?
We will answer that later.
Let's ask something else.

What about 4K?
With 4K "GPU working set" will grow 4x, at least.
We are looking at 1200*4 = 4.8GB of GPU data.
CPU working set will also grow (everybody wants these better scripts and physics I presume?) but probably 2x only, to 700*2 = ~1.5GB
So overall the persistent memory will be well over 6GB, let's say 6.5GB.
That leaves us with ~5GB of free RAM in XSeX and ~8GB for PS5.

Stop, stop! Why PS5 has more RAM suddenly?
That's simple.
XSeX RAM is divided into two pools (logically, physically it's the same RAM): 10GB and 3.5GB.
GPU working set must use the 10GB pool (it's the memory set that absolutely needs the fast bandwidth).
So 10 - 4.8 = 5.2 which is ~5GB
CPU working set will use 3.5GB pool and we will have a spare 2GB there for other things.
We may load some low freq data there, like streaming meshes and stuff, but it will hard to use in each frame: accessing that data too frequently will lower the whole system bandwidth to 336Mb/sec.
That's why MSFT calls the 10GB pool "GPU optimal".

But what about PS5? It also has some RAM reserved for the system? It should be ~14GB usable!
Nope, sorry.
PS5 has a 5.5GB/sec flash drive. That typically loads 2GB in 0.27 sec. It's write speed is lower, but not less than 5.5GB/sec raw.
What PS5 can do, and I would be pretty surprised if Sony won't do it. Is to save the system image to the disk while the game is playing.
And thus give almost full 16GB of RAM to the game.
2GB system image will load into RAM in <1 sec (save 2GB game data to disk in 0.6 sec + load system from disk 0.3 sec). Why keep it resident?
But I'm on the safe side here. So it's ~14.5GB usable for PS5.

Hmm, essentially MSFT can do that too?
Yep, they can. The speeds will be less sexy but not more than ~3sec, I think.
Why don't they do it? Probably they rely on OS constantly running on the background for all the services it provides.
That's why I gave Sony 14.5GB.
But I have hard time understanding why 2.5GB is needed, all the background services can run on a much smaller RAM footprint just fine, and UI stuff can load on-demand.

Can we talk about SSD for games now?
Yup.
So, let's get to the numbers again.
For XSeX ~5GB of "free" RAM we can divide it into 2 parts: resident and streaming.
Why two? Because typically you cannot load shit into frame while frame is rendering.
GPU is so fast, that each time you ask GPU "what exact memory location are you reading now?" will slow it down to give you an answer.

But can you load things into other part while the first one is rendering?
Absolutely. You can switch "resident" and "streaming" part as much as you like, if it's fast enough.
Anyway, we got to 50/50 of "new shit" to "old shit" inside 1 second now!
2.5GB of resident + 2.5GB of streaming pool and it takes XSeX just 1 sec to completely reload the streaming part!
In 1 min we have 60:1 of new/old ratio!
Nice!

What about PS5 then? Is it just 2x faster and that's it?
Not really.
The whole 8GB of the RAM we have "free" can be a "streaming pool" on PS5.

But you said "we cannot load while frame is rendering"?
In XSeX, yes.
But in PS5 we have GPU cache scrubbers.
This is a piece of silicon inside the GPU that will reload our assets on the fly while GPU is rendering the frame.
It has full access to where and what GPU is reading right now (it's all in the GPU cache, hence "cache scrubber")
It will also never invalidate the whole cache (which can still lead to GPU "stall") but reload exactly the data that changed (I hope you've listened to that part of Cerny's talk very closely).

But it's free RAM size doesn't really matter, we still have 2:1 of old/new in one frame, because SSD is only 2x faster?
Yes, and no.
We do have only 2x faster rates (although the max rates are much higher for PS5: 22GB/sec vs 6GB/sec)
But the thing is, GPU can render from 8GB of game data. And XSeX - only from 2.5GB, do you remember that we cannot render from the "streaming" part while it loads?
So in any given scene, potentially, PS5 can have 2x to 3x more details/textures/assets than XSeX.
Yes, XSeX will render it faster, higher FPS or higher frame-buffer resolution (not both, perf difference is too low).
But the scene itself will be less detailed, have less artwork.

OMG, can MSFT do something about it?
Of course they will, and they do!
What are the XSeX advantages? More ALU power (FLOPS) more RT power, more CPU power.
What MSFT will do: rely heavily on this power advantage instead of the artwork: more procedural stuff, more ALU used for physics simulation (remember, RT and lighting is a physics simulation too, after all).
More compute and more complex shaders.

So what will be the end result?
It's pretty simple.
PS5: relies on more artwork and pushing more data through the system. Potentially 2x performance in that.
XSeX: relies more on in-frame calculations, procedural. Potentially 30% performance in that.
Who will win: dunno. There are pros and cons for each.
It will be a fun generation indeed. Much more fun than the previous one, for sure.

Damn bro. That's deep.

 

[SgtCaffran]

On the topic of BW issue, Github had oberon at 448 GB/s theoretical but achieved 512 GB/s ( it's in the DF video about it ) now was it a test with different memory chips, or another typo, or the way the system is built around the SSD allows that.
Any thoughts about it ?

Maybe they did tests with both 448 and 512 and concluded that 448 was enough not to be a bottleneck? Remember PS5 needs less bandwidth due to fewer CUs.

Ultimately, the XSX is better equipped to handle multiplats. All things being equal, it is more capable on the GPU and CPU front.

To be honest, I think both will be so close to being equal om 3rd party games that there won't be any reason to buy one over the other purely for 3rd party quality. Previous console - so game library - ownership is probably the main factor.

 

[ethomaz]

Inside Xbox Series X: the full specs

This is it. After months of teaser trailers, blog posts and even the occasional leak, we can finally reveal firm, hard …

www.eurogamer.net

"Our second component is a high-speed hardware decompression block that can deliver more than 6GB / s," reveals Andrew Goossen. "This is a dedicated silicon block that offloads the decompression job from the CPU and adapts to the SSD so that decompression is never a bottleneck. The decompression hardware supports Zlib for general data and a new (system) of compression called BCPack that adapts to GPU textures that generally comprise the vast majority of a game's pack size. "

You mistook the SSD speeds with the descompression unit.
Descompression unit can do decompression up to 6GB/s.
SSD speed is up to 4.8GB/s in the best compressed scenario.

 

[Coulomb_Barrier]

Ok so I see some laughable statements, "the difference is in Xbox SX favor" when the difference is 16%

But

"SSD is exactly the same you won't see a difference but 2sec in loading times" when the SSD is 2X faster (120+%) on ps5

I mean quit the nonsense lol

Post of the morning lool you're right.

 

[MilkyJoe]

Oh Series X fans are not going to like this tweet:

"Wtf is a "6tflop console", you're just making this up. Teraflop measures one part of the pipeline of one discreet component of a console. You don't call your car a "12 gallon fuel tank saloon" like that means anything"

Matt Phillips, ex-Crytek, Travellers Tales programmer.

Twitter Publish

publish.twitter.com

Dude, you're going to drive yourself insane if you keep up this parallel universe nonsense.

The only thing the PS5 has faster is the hard drive.

If you don't believe it won't make a difference, maybe you'll believe this guy

 

[Lone Wolf]

Oh Series X fans are not going to like this tweet:

"Wtf is a "6tflop console", you're just making this up. Teraflop measures one part of the pipeline of one discreet component of a console. You don't call your car a "12 gallon fuel tank saloon" like that means anything"

Matt Phillips, ex-Crytek, Travellers Tales programmer.

Twitter Publish

publish.twitter.com

I would equate Tflops more to HP than to the fuel tank, and cars are most certainly advertised with HP ratings.

 

[ethomaz]

On the topic of BW issue, Github had oberon at 448 GB/s theoretical but achieved 512 GB/s ( it's in the DF video about it ) now was it a test with different memory chips, or another typo, or the way the system is built around the SSD allows that.
Any thoughts about it ?

GitHub never had 448GB/s memory setup in their tests.

Ariel had 488GB/s and Oberon 512GB/s setups.

 

[VanSiriusGabriel]

Maybe they did tests with both 448 and 512 and concluded that 448 was enough not to be a bottleneck? Remember PS5 needs less bandwidth due to fewer CUs

Which is the opposite strategy compared to PS4, where they have thrown crazy fast GDDR5 memory at a mid level GPU.

 

[Tenka Musou]

Everyone will forget TFs when they see the Demon's Souls glorious ray-traced armor reflections.

Meanwhile Xbox first two years games will be designed around a 1.3TF GPU, Jaguar CPU, and an ancient HDD (Xbone).

Anyone buying XBox Series X for the graphics will be regretting his decision very quickly.

Spoiler

Yes, I am very salty about that 9.2TF number, damn you insiders!!!!!!!

 

[ethomaz]

22 GB\s sounds even more honest.

PS5 can reach 22GB/s speeds the same way Xbox reach 4.8GB/s thought compressed data.

But in most cases you will get only 8-9GB/s compressed with PS5.

 

[VanSiriusGabriel]

I would equate Tflops more to HP than to the fuel tank, and cars are most certainly advertised with HP ratings.

The point of the developer is that for a car engineer a "800 HP car" means relatively shit, aside from the usage in marketing.

 

[Bo_Hazem]

People tend to forget about this I regards to xsex;

I think that until mid gen refresh xbsex will be the better console to play games, it takes time for devs to fully understand new architecture. And then when the xbsex ssd becomes bottleneck we will get a pro version of both consoles.

we need to remember that developers will need time to switch from hdd mindset to ssd and then move towards fully utilising available speeds.

Let's start with saying again XSX is a great piece of hardware. Assuming it can do that 6GB/s then we shouldn't forget what an insane 22GB/s Kraken can do, that means if you turn around extremely fast, you can buy some trickery use some motion blur and download 5GB in 0.25sec, and that's the theoretical number like 6GB/s of Xbox. Reality is both confirm around realistic 4.8GB/s and 9GB/s for XSX and PS5 respectively. Sony is planning to use it as a direct connection between data in the SSD to CPU/GPU/RAM with no replicas or extra processing.

It'll work like that in ALL games for PS5, without an effort.

Both are approaching next gen in different methods, but I guess all roads lead to Rome as they say.

 

[VanSiriusGabriel]

PS5 can reach 22GB/s speeds the same way Xbox reach 4.8GB/s thought compressed data.

But in most cases you will get only 8-9GB/s compressed with PS5.

I got it, but still.

 

[Leskov]

There's a lot of confusion on why SSD is so important for next-gen and how it will change things.
Here I will try to explain the main concepts.
TL;DR fast SSD is a game changing feature, this generation will be fun to watch!

It was working fine before, why do we even need that?
No, it wasn't fine, it was a giant PITA for anything other than small multiplayer maps or fighting games.
Let's talk some numbers. Unfortunately not many games have ever published their RAM pools and asset pools to the public, but some did.
Enter Killzone: Shadowfall Demo presentation.
We have roughly the following:

Type Approx. Size, % Approx. Size, MB
Textures 30% 1400
CPU working set 15% 700
GPU working set 25% 1200
Streaming pool 10% 500
Sounds 10% 450
Meshes 10% 450
Animations/Particles 1% 45

*These numbers are rounded sums of various much more detailed numbers presented in the article above.

We are interested in the "streaming pool" number here (but we will talk about others too)
We have ~500MB of data that is loaded as the demo progresses, on the fly.
The whole chunk of data that the game samples from (for that streaming process) is 1600MB.
The load speed of PS4 drive is (compressed data) <50MB/sec (uncompressed is <20MB/sec), i.e. it will take >30sec to load that at least.

It seems like it's not that big of a problem, and indeed for demo it is. But what about the game?
The game size is ~40GB, you have 6.5GB of usable RAM, you cannot load the whole game, even if you tried.
So what's left? We can either stream things in, or do a loading screen between each new section.
Let's try the easier approach: do a loading screen
We have 6.5GB of RAM, and the resident set is ~2GB from the table above (GPU + CPU working set). We need to load 4.5GB each time. It's 90 seconds, pretty annoying, but it's the best case. Any time you need to load things not sequentially, you will need to seek the drive and the time will increase.
You can't go back, as it will re-load things and - another loading screen.
You can't use more than 4.5GB assets in your whole gaming section, or you will need another loading screen.
It gets even more ridiculous if your levels are dynamic: left an item in previous zone? Load time will increase (item is not built into the gaming world, we load the world, then we seek for each item/item group on disk).
Remember Skyrim? Loading into each house? That's what will happen.
So, loading screens are easy, but if your game is not a linear, static, theme-park style attraction it gets ridiculous pretty fast.

How to we stream then?
We have a chunk of memory (remember 500Mb) that's reserved for streaming things from disk.
With our 50MB/sec speed we fill it up each 10 sec.
So, each 10 sec we can have a totally new data in RAM.
Let's do some metrics, for example: how much new shit we can show to the player in 1 min? Easy: 6*500 = 3GB
How much old shit player sees each minute? Easy again: 1400+450+450+45=~ 2.5GB
So we have a roughly 50/50 old to new shit on screen.
Reused monsters? assets? textures? NPCs? you name it. You have the 50/50 going on.

But PS4 has 6.5GB of RAM, we used only 4.5GB till now, what about other 2GB?
Excellent question!
The answer is: it goes to the old shit. Because if we increase the streaming buffer to 1.5GB it still does nothing to the 50MB/sec speed.
With the full 6.5GB we get to 6GB old vs 3GB new in 1 minute. Which is 2:1 old shit wins.

But what about 10 minutes?
Good, good. Here we go!
In 10 min we can get to 30GB new shit vs 6GB old.
And that's, my friends, how the games worked last gen.
You're as a player were introduced to the new gaming moments very gradually.
Or, there were some tricks they used: open doors animation.
Remember Uncharted with all the "let's open that heavy door for 15sec?" that's because new shit needs to load, players need to get to a new location, but we cannot load it fast.

So, what about SSDs then?
We will answer that later.
Let's ask something else.

What about 4K?
With 4K "GPU working set" will grow 4x, at least.
We are looking at 1200*4 = 4.8GB of GPU data.
CPU working set will also grow (everybody wants these better scripts and physics I presume?) but probably 2x only, to 700*2 = ~1.5GB
So overall the persistent memory will be well over 6GB, let's say 6.5GB.
That leaves us with ~5GB of free RAM in XSeX and ~8GB for PS5.

Stop, stop! Why PS5 has more RAM suddenly?
That's simple.
XSeX RAM is divided into two pools (logically, physically it's the same RAM): 10GB and 3.5GB.
GPU working set must use the 10GB pool (it's the memory set that absolutely needs the fast bandwidth).
So 10 - 4.8 = 5.2 which is ~5GB
CPU working set will use 3.5GB pool and we will have a spare 2GB there for other things.
We may load some low freq data there, like streaming meshes and stuff, but it will hard to use in each frame: accessing that data too frequently will lower the whole system bandwidth to 336Mb/sec.
That's why MSFT calls the 10GB pool "GPU optimal".

But what about PS5? It also has some RAM reserved for the system? It should be ~14GB usable!
Nope, sorry.
PS5 has a 5.5GB/sec flash drive. That typically loads 2GB in 0.27 sec. It's write speed is lower, but not less than 5.5GB/sec raw.
What PS5 can do, and I would be pretty surprised if Sony won't do it. Is to save the system image to the disk while the game is playing.
And thus give almost full 16GB of RAM to the game.
2GB system image will load into RAM in <1 sec (save 2GB game data to disk in 0.6 sec + load system from disk 0.3 sec). Why keep it resident?
But I'm on the safe side here. So it's ~14.5GB usable for PS5.

Hmm, essentially MSFT can do that too?
Yep, they can. The speeds will be less sexy but not more than ~3sec, I think.
Why don't they do it? Probably they rely on OS constantly running on the background for all the services it provides.
That's why I gave Sony 14.5GB.
But I have hard time understanding why 2.5GB is needed, all the background services can run on a much smaller RAM footprint just fine, and UI stuff can load on-demand.

Can we talk about SSD for games now?
Yup.
So, let's get to the numbers again.
For XSeX ~5GB of "free" RAM we can divide it into 2 parts: resident and streaming.
Why two? Because typically you cannot load shit into frame while frame is rendering.
GPU is so fast, that each time you ask GPU "what exact memory location are you reading now?" will slow it down to give you an answer.

But can you load things into other part while the first one is rendering?
Absolutely. You can switch "resident" and "streaming" part as much as you like, if it's fast enough.
Anyway, we got to 50/50 of "new shit" to "old shit" inside 1 second now!
2.5GB of resident + 2.5GB of streaming pool and it takes XSeX just 1 sec to completely reload the streaming part!
In 1 min we have 60:1 of new/old ratio!
Nice!

What about PS5 then? Is it just 2x faster and that's it?
Not really.
The whole 8GB of the RAM we have "free" can be a "streaming pool" on PS5.

But you said "we cannot load while frame is rendering"?
In XSeX, yes.
But in PS5 we have GPU cache scrubbers.
This is a piece of silicon inside the GPU that will reload our assets on the fly while GPU is rendering the frame.
It has full access to where and what GPU is reading right now (it's all in the GPU cache, hence "cache scrubber")
It will also never invalidate the whole cache (which can still lead to GPU "stall") but reload exactly the data that changed (I hope you've listened to that part of Cerny's talk very closely).

But it's free RAM size doesn't really matter, we still have 2:1 of old/new in one frame, because SSD is only 2x faster?
Yes, and no.
We do have only 2x faster rates (although the max rates are much higher for PS5: 22GB/sec vs 6GB/sec)
But the thing is, GPU can render from 8GB of game data. And XSeX - only from 2.5GB, do you remember that we cannot render from the "streaming" part while it loads?
So in any given scene, potentially, PS5 can have 2x to 3x more details/textures/assets than XSeX.
Yes, XSeX will render it faster, higher FPS or higher frame-buffer resolution (not both, perf difference is too low).
But the scene itself will be less detailed, have less artwork.

OMG, can MSFT do something about it?
Of course they will, and they do!
What are the XSeX advantages? More ALU power (FLOPS) more RT power, more CPU power.
What MSFT will do: rely heavily on this power advantage instead of the artwork: more procedural stuff, more ALU used for physics simulation (remember, RT and lighting is a physics simulation too, after all).
More compute and more complex shaders.

So what will be the end result?
It's pretty simple.
PS5: relies on more artwork and pushing more data through the system. Potentially 2x performance in that.
XSeX: relies more on in-frame calculations, procedural. Potentially 30% performance in that.
Who will win: dunno. There are pros and cons for each.
It will be a fun generation indeed. Much more fun than the previous one, for sure.

So basically it will be implemented in 30, maybe 40 games per ps5 lifecycle and it doesn't matter for multiplat - which is the wast majority of games on the market. So tradeoff in gpu for better ssd isn't so worthy in that department. I just hope that the difference will be only in resolution part and fps and other things that affect input delay and "feel" of the game won't suffer because of that.

 

[VanSiriusGabriel]

So basically it will be implemented in 30, maybe 40 games per ps5 lifecycle and it doesn't matter for multiplat - which is the wast majority of games on the market. So tradeoff in gpu for better ssd isn't so worthy in that department. I just hope that the difference will be only in resolution part and fps and other things that affect input delay and "feel" of the game won't suffer because of that.

If those are the 30 games you will remember, it's worth it.
Again, one hundred of me doesn't make one Stephen Hawking.

 

[PocoJoe]

You don't say? So Xbox has 120CU's, right? Or let's say PS5 is RDNA1 and Navi 10 and has no RT cores. Go ahead. Mark "Freakin" Cerny been lying to us all this time.

xbox = 52 CU
PS5 = 36CU

It’s pro vs X all over again with Series X coming out on top. That’s fast vs wide design, it is what it is . It’s a lotus sports car philosophy vs a Bugatti Veyron ... everyone knows which one wins in pure power.

But carry on .

It’s clear to me PS5 was designed for 399 in mind with beter performance for 399 and a 2019 release date. And Cerny did a excellent job with these guides in place. To bad component pricing got in the way.

Xbox is best all around.

Says a fanboy?

These are GAMING consoles, not HEY LOOK AT MY TFLOPS! consoles.

So saying one is "best all around" without seeing games, controllers and full reveals = useless trolling.

We already know that PS5 have faster/better SSD, so please explain how xbox is best all around?

and controllers, some like weird xbox stick placement made for mutants, some like symmetrical playstation stick placement made for humans

So there isnt "best all around" console, and specs dont matter much, games will. And we know which system have been more popular among gamers on this gen and which havent.

Pro vs X = cheaper and less powerful system sold more.

Just dont say "best all around" when it is lying and complete bullshit.

Maybe it is just too difficult for xbox fans to admit that even with better specs on some parts = they are still on the losing side because Sony is just much more popular on global scale?

Sony would have to fuck up badly to lose badly, it is about that, not like "if xbox does its best it can win vs Sony at its best"

Xbox can be clear winner probably only in the situation where Playstation does something really horribly wrong, and it doesnt look like it at the moment

 

[Lone Wolf]

Is Stadia more powerful than the PS5?

Google Stadia has a powerful GPU backing its cloud servers, but game streaming is still subject to a host of unforeseen problems. PS5 delivers the entire package as a physical console, ...

www.androidcentral.com

We all know it’s not true, but will Google push this narrative?

 

[PocoJoe]

So basically it will be implemented in 30, maybe 40 games per ps5 lifecycle and it doesn't matter for multiplat - which is the wast majority of games on the market. So tradeoff in gpu for better ssd isn't so worthy in that department. I just hope that the difference will be only in resolution part and fps and other things that affect input delay and "feel" of the game won't suffer because of that.

If it is as easy as Cerny said (to use the speed of SSD), that isnt the case. Even 3rd party games can benefit from it.

Of course if game is designed the shitty xbox one and/or PCs with HDD in mind = they would have to make two versions.

it all depends about the devs, so nobody knows the numbers.

Dont downplay this SSD tech.

And anyway if most popular multiplatform games are fifa/nhl/cod = there isnt even much of a need to have all next gen experiences.

But games that use the full potential of next gen in the design = maybe they are done for PS5 and downsized for pc/xbox, who knows

 

[Moses85]

Here is a post from activision, maybe they are not so far apart...., the plot thickens.

Oh, wow ! Look at the Avatar this employee from Activision is using!
The Guy is MARK CERNY!!

 

[ethomaz]

Is Stadia more powerful than the PS5?

Google Stadia has a powerful GPU backing its cloud servers, but game streaming is still subject to a host of unforeseen problems. PS5 delivers the entire package as a physical console, ...

www.androidcentral.com

We all know it’s not true, but will Google push this narrative?

Games on Stadia tell themselves

 

[VanSiriusGabriel]

Is Stadia more powerful than the PS5?

Google Stadia has a powerful GPU backing its cloud servers, but game streaming is still subject to a host of unforeseen problems. PS5 delivers the entire package as a physical console, ...

www.androidcentral.com

We all know it’s not true, but will Google push this narrative?

Stadia already pushed itself from a bridge, there isn't crap to push anymore.

 

[geordiemp]

There's a lot of confusion on why SSD is so important for next-gen and how it will change things.
Here I will try to explain the main concepts.
TL;DR fast SSD is a game changing feature, this generation will be fun to watch!

It was working fine before, why do we even need that?
No, it wasn't fine, it was a giant PITA for anything other than small multiplayer maps or fighting games.
Let's talk some numbers. Unfortunately not many games have ever published their RAM pools and asset pools to the public, but some did.
Enter Killzone: Shadowfall Demo presentation.
We have roughly the following:

Type Approx. Size, % Approx. Size, MB
Textures 30% 1400
CPU working set 15% 700
GPU working set 25% 1200
Streaming pool 10% 500
Sounds 10% 450
Meshes 10% 450
Animations/Particles 1% 45

*These numbers are rounded sums of various much more detailed numbers presented in the article above.

We are interested in the "streaming pool" number here (but we will talk about others too)
We have ~500MB of data that is loaded as the demo progresses, on the fly.
The whole chunk of data that the game samples from (for that streaming process) is 1600MB.
The load speed of PS4 drive is (compressed data) <50MB/sec (uncompressed is <20MB/sec), i.e. it will take >30sec to load that at least.

It seems like it's not that big of a problem, and indeed for demo it is. But what about the game?
The game size is ~40GB, you have 6.5GB of usable RAM, you cannot load the whole game, even if you tried.
So what's left? We can either stream things in, or do a loading screen between each new section.
Let's try the easier approach: do a loading screen
We have 6.5GB of RAM, and the resident set is ~2GB from the table above (GPU + CPU working set). We need to load 4.5GB each time. It's 90 seconds, pretty annoying, but it's the best case. Any time you need to load things not sequentially, you will need to seek the drive and the time will increase.
You can't go back, as it will re-load things and - another loading screen.
You can't use more than 4.5GB assets in your whole gaming section, or you will need another loading screen.
It gets even more ridiculous if your levels are dynamic: left an item in previous zone? Load time will increase (item is not built into the gaming world, we load the world, then we seek for each item/item group on disk).
Remember Skyrim? Loading into each house? That's what will happen.
So, loading screens are easy, but if your game is not a linear, static, theme-park style attraction it gets ridiculous pretty fast.

How to we stream then?
We have a chunk of memory (remember 500Mb) that's reserved for streaming things from disk.
With our 50MB/sec speed we fill it up each 10 sec.
So, each 10 sec we can have a totally new data in RAM.
Let's do some metrics, for example: how much new shit we can show to the player in 1 min? Easy: 6*500 = 3GB
How much old shit player sees each minute? Easy again: 1400+450+450+45=~ 2.5GB
So we have a roughly 50/50 old to new shit on screen.
Reused monsters? assets? textures? NPCs? you name it. You have the 50/50 going on.

But PS4 has 6.5GB of RAM, we used only 4.5GB till now, what about other 2GB?
Excellent question!
The answer is: it goes to the old shit. Because if we increase the streaming buffer to 1.5GB it still does nothing to the 50MB/sec speed.
With the full 6.5GB we get to 6GB old vs 3GB new in 1 minute. Which is 2:1 old shit wins.

But what about 10 minutes?
Good, good. Here we go!
In 10 min we can get to 30GB new shit vs 6GB old.
And that's, my friends, how the games worked last gen.
You're as a player were introduced to the new gaming moments very gradually.
Or, there were some tricks they used: open doors animation.
Remember Uncharted with all the "let's open that heavy door for 15sec?" that's because new shit needs to load, players need to get to a new location, but we cannot load it fast.

So, what about SSDs then?
We will answer that later.
Let's ask something else.

What about 4K?
With 4K "GPU working set" will grow 4x, at least.
We are looking at 1200*4 = 4.8GB of GPU data.
CPU working set will also grow (everybody wants these better scripts and physics I presume?) but probably 2x only, to 700*2 = ~1.5GB
So overall the persistent memory will be well over 6GB, let's say 6.5GB.
That leaves us with ~5GB of free RAM in XSeX and ~8GB for PS5.

Stop, stop! Why PS5 has more RAM suddenly?
That's simple.
XSeX RAM is divided into two pools (logically, physically it's the same RAM): 10GB and 3.5GB.
GPU working set must use the 10GB pool (it's the memory set that absolutely needs the fast bandwidth).
So 10 - 4.8 = 5.2 which is ~5GB
CPU working set will use 3.5GB pool and we will have a spare 2GB there for other things.
We may load some low freq data there, like streaming meshes and stuff, but it will hard to use in each frame: accessing that data too frequently will lower the whole system bandwidth to 336Mb/sec.
That's why MSFT calls the 10GB pool "GPU optimal".

But what about PS5? It also has some RAM reserved for the system? It should be ~14GB usable!
Nope, sorry.
PS5 has a 5.5GB/sec flash drive. That typically loads 2GB in 0.27 sec. It's write speed is lower, but not less than 5.5GB/sec raw.
What PS5 can do, and I would be pretty surprised if Sony won't do it. Is to save the system image to the disk while the game is playing.
And thus give almost full 16GB of RAM to the game.
2GB system image will load into RAM in <1 sec (save 2GB game data to disk in 0.6 sec + load system from disk 0.3 sec). Why keep it resident?
But I'm on the safe side here. So it's ~14.5GB usable for PS5.

Hmm, essentially MSFT can do that too?
Yep, they can. The speeds will be less sexy but not more than ~3sec, I think.
Why don't they do it? Probably they rely on OS constantly running on the background for all the services it provides.
That's why I gave Sony 14.5GB.
But I have hard time understanding why 2.5GB is needed, all the background services can run on a much smaller RAM footprint just fine, and UI stuff can load on-demand.

Can we talk about SSD for games now?
Yup.
So, let's get to the numbers again.
For XSeX ~5GB of "free" RAM we can divide it into 2 parts: resident and streaming.
Why two? Because typically you cannot load shit into frame while frame is rendering.
GPU is so fast, that each time you ask GPU "what exact memory location are you reading now?" will slow it down to give you an answer.

But can you load things into other part while the first one is rendering?
Absolutely. You can switch "resident" and "streaming" part as much as you like, if it's fast enough.
Anyway, we got to 50/50 of "new shit" to "old shit" inside 1 second now!
2.5GB of resident + 2.5GB of streaming pool and it takes XSeX just 1 sec to completely reload the streaming part!
In 1 min we have 60:1 of new/old ratio!
Nice!

What about PS5 then? Is it just 2x faster and that's it?
Not really.
The whole 8GB of the RAM we have "free" can be a "streaming pool" on PS5.

But you said "we cannot load while frame is rendering"?
In XSeX, yes.
But in PS5 we have GPU cache scrubbers.
This is a piece of silicon inside the GPU that will reload our assets on the fly while GPU is rendering the frame.
It has full access to where and what GPU is reading right now (it's all in the GPU cache, hence "cache scrubber")
It will also never invalidate the whole cache (which can still lead to GPU "stall") but reload exactly the data that changed (I hope you've listened to that part of Cerny's talk very closely).

But it's free RAM size doesn't really matter, we still have 2:1 of old/new in one frame, because SSD is only 2x faster?
Yes, and no.
We do have only 2x faster rates (although the max rates are much higher for PS5: 22GB/sec vs 6GB/sec)
But the thing is, GPU can render from 8GB of game data. And XSeX - only from 2.5GB, do you remember that we cannot render from the "streaming" part while it loads?
So in any given scene, potentially, PS5 can have 2x to 3x more details/textures/assets than XSeX.
Yes, XSeX will render it faster, higher FPS or higher frame-buffer resolution (not both, perf difference is too low).
But the scene itself will be less detailed, have less artwork.

OMG, can MSFT do something about it?
Of course they will, and they do!
What are the XSeX advantages? More ALU power (FLOPS) more RT power, more CPU power.
What MSFT will do: rely heavily on this power advantage instead of the artwork: more procedural stuff, more ALU used for physics simulation (remember, RT and lighting is a physics simulation too, after all).
More compute and more complex shaders.

So what will be the end result?
It's pretty simple.
PS5: relies on more artwork and pushing more data through the system. Potentially 2x performance in that.
XSeX: relies more on in-frame calculations, procedural. Potentially 30% performance in that.
Who will win: dunno. There are pros and cons for each.
It will be a fun generation indeed. Much more fun than the previous one, for sure.

You got that from beyond3d lol, nice cut and paste. Naughty boy...

Its good stuff though

 

[scie]

Guys do you wanna hear something shocking? Both consoles will play games to your specific liking!
90 % of the customers don´t give a damn shit how many Tflops, how fast the SSD in the consoles are. They are going to buy them online or in a local store, hook them up to a crappy uncalibrated TV and maybe they have some sort of sound system and don´t use the internal speakers of their TVs. And guess what? They will have a blast of their time! Regardless of having the "weaker" console. My thought for the day Enjoy playing games

 

[TBiddy]

You got that from beyond3d lol, nice cut and paste. Naughty boy...

Its good stuff though

Wasn't it copy/pasted from another thread here on gaf, by P psorcerer ?

 

[B_Boss]

In that case, I got work to do...

Wow, there's a lot of people doing a lot of work to deny what's really undeniable. The Series X has a spec advantage this time around and it'll almost certainly be the best place to play third party titles.

The sudden, almost religious, fervour around SSDs is exactly the same kind of speculative claims that we saw at the start of last gen but with the roles reversed. "...but the cloud!" has become ...but the SSD!".

More power is just that and more power means better graphics and performance.

It never crashed Haz. Peoples expectations did . Cerny delivered in a way totally unexpected only by the tflop wielding masses.

 

[VanSiriusGabriel]

Guys do you wanna hear something shocking? Both consoles will play games to your specific liking!
90 % of the customers don´t give a damn shit how many Tflops, how fast the SSD in the consoles are. They are going to buy them online or in a local store, hook them up to a crappy uncalibrated TV and maybe they have some sort of sound system and don´t use the internal speakers of their TVs. And guess what? They will have a blast of their time! Regardless of having the "weaker" console. My thought for the day Enjoy playing games

I agree, but this is a forum for more dedicated gamers on average, discussions about 5% this or 12% that and software applications will never stop.

 

[Bo_Hazem]

There's a lot of confusion on why SSD is so important for next-gen and how it will change things.
Here I will try to explain the main concepts.
TL;DR fast SSD is a game changing feature, this generation will be fun to watch!

It was working fine before, why do we even need that?
No, it wasn't fine, it was a giant PITA for anything other than small multiplayer maps or fighting games.
Let's talk some numbers. Unfortunately not many games have ever published their RAM pools and asset pools to the public, but some did.
Enter Killzone: Shadowfall Demo presentation.
We have roughly the following:

Type Approx. Size, % Approx. Size, MB
Textures 30% 1400
CPU working set 15% 700
GPU working set 25% 1200
Streaming pool 10% 500
Sounds 10% 450
Meshes 10% 450
Animations/Particles 1% 45

*These numbers are rounded sums of various much more detailed numbers presented in the article above.

We are interested in the "streaming pool" number here (but we will talk about others too)
We have ~500MB of data that is loaded as the demo progresses, on the fly.
The whole chunk of data that the game samples from (for that streaming process) is 1600MB.
The load speed of PS4 drive is (compressed data) <50MB/sec (uncompressed is <20MB/sec), i.e. it will take >30sec to load that at least.

It seems like it's not that big of a problem, and indeed for demo it is. But what about the game?
The game size is ~40GB, you have 6.5GB of usable RAM, you cannot load the whole game, even if you tried.
So what's left? We can either stream things in, or do a loading screen between each new section.
Let's try the easier approach: do a loading screen
We have 6.5GB of RAM, and the resident set is ~2GB from the table above (GPU + CPU working set). We need to load 4.5GB each time. It's 90 seconds, pretty annoying, but it's the best case. Any time you need to load things not sequentially, you will need to seek the drive and the time will increase.
You can't go back, as it will re-load things and - another loading screen.
You can't use more than 4.5GB assets in your whole gaming section, or you will need another loading screen.
It gets even more ridiculous if your levels are dynamic: left an item in previous zone? Load time will increase (item is not built into the gaming world, we load the world, then we seek for each item/item group on disk).
Remember Skyrim? Loading into each house? That's what will happen.
So, loading screens are easy, but if your game is not a linear, static, theme-park style attraction it gets ridiculous pretty fast.

How to we stream then?
We have a chunk of memory (remember 500Mb) that's reserved for streaming things from disk.
With our 50MB/sec speed we fill it up each 10 sec.
So, each 10 sec we can have a totally new data in RAM.
Let's do some metrics, for example: how much new shit we can show to the player in 1 min? Easy: 6*500 = 3GB
How much old shit player sees each minute? Easy again: 1400+450+450+45=~ 2.5GB
So we have a roughly 50/50 old to new shit on screen.
Reused monsters? assets? textures? NPCs? you name it. You have the 50/50 going on.

But PS4 has 6.5GB of RAM, we used only 4.5GB till now, what about other 2GB?
Excellent question!
The answer is: it goes to the old shit. Because if we increase the streaming buffer to 1.5GB it still does nothing to the 50MB/sec speed.
With the full 6.5GB we get to 6GB old vs 3GB new in 1 minute. Which is 2:1 old shit wins.

But what about 10 minutes?
Good, good. Here we go!
In 10 min we can get to 30GB new shit vs 6GB old.
And that's, my friends, how the games worked last gen.
You're as a player were introduced to the new gaming moments very gradually.
Or, there were some tricks they used: open doors animation.
Remember Uncharted with all the "let's open that heavy door for 15sec?" that's because new shit needs to load, players need to get to a new location, but we cannot load it fast.

So, what about SSDs then?
We will answer that later.
Let's ask something else.

What about 4K?
With 4K "GPU working set" will grow 4x, at least.
We are looking at 1200*4 = 4.8GB of GPU data.
CPU working set will also grow (everybody wants these better scripts and physics I presume?) but probably 2x only, to 700*2 = ~1.5GB
So overall the persistent memory will be well over 6GB, let's say 6.5GB.
That leaves us with ~5GB of free RAM in XSeX and ~8GB for PS5.

Stop, stop! Why PS5 has more RAM suddenly?
That's simple.
XSeX RAM is divided into two pools (logically, physically it's the same RAM): 10GB and 3.5GB.
GPU working set must use the 10GB pool (it's the memory set that absolutely needs the fast bandwidth).
So 10 - 4.8 = 5.2 which is ~5GB
CPU working set will use 3.5GB pool and we will have a spare 2GB there for other things.
We may load some low freq data there, like streaming meshes and stuff, but it will hard to use in each frame: accessing that data too frequently will lower the whole system bandwidth to 336Mb/sec.
That's why MSFT calls the 10GB pool "GPU optimal".

But what about PS5? It also has some RAM reserved for the system? It should be ~14GB usable!
Nope, sorry.
PS5 has a 5.5GB/sec flash drive. That typically loads 2GB in 0.27 sec. It's write speed is lower, but not less than 5.5GB/sec raw.
What PS5 can do, and I would be pretty surprised if Sony won't do it. Is to save the system image to the disk while the game is playing.
And thus give almost full 16GB of RAM to the game.
2GB system image will load into RAM in <1 sec (save 2GB game data to disk in 0.6 sec + load system from disk 0.3 sec). Why keep it resident?
But I'm on the safe side here. So it's ~14.5GB usable for PS5.

Hmm, essentially MSFT can do that too?
Yep, they can. The speeds will be less sexy but not more than ~3sec, I think.
Why don't they do it? Probably they rely on OS constantly running on the background for all the services it provides.
That's why I gave Sony 14.5GB.
But I have hard time understanding why 2.5GB is needed, all the background services can run on a much smaller RAM footprint just fine, and UI stuff can load on-demand.

Can we talk about SSD for games now?
Yup.
So, let's get to the numbers again.
For XSeX ~5GB of "free" RAM we can divide it into 2 parts: resident and streaming.
Why two? Because typically you cannot load shit into frame while frame is rendering.
GPU is so fast, that each time you ask GPU "what exact memory location are you reading now?" will slow it down to give you an answer.

But can you load things into other part while the first one is rendering?
Absolutely. You can switch "resident" and "streaming" part as much as you like, if it's fast enough.
Anyway, we got to 50/50 of "new shit" to "old shit" inside 1 second now!
2.5GB of resident + 2.5GB of streaming pool and it takes XSeX just 1 sec to completely reload the streaming part!
In 1 min we have 60:1 of new/old ratio!
Nice!

What about PS5 then? Is it just 2x faster and that's it?
Not really.
The whole 8GB of the RAM we have "free" can be a "streaming pool" on PS5.

But you said "we cannot load while frame is rendering"?
In XSeX, yes.
But in PS5 we have GPU cache scrubbers.
This is a piece of silicon inside the GPU that will reload our assets on the fly while GPU is rendering the frame.
It has full access to where and what GPU is reading right now (it's all in the GPU cache, hence "cache scrubber")
It will also never invalidate the whole cache (which can still lead to GPU "stall") but reload exactly the data that changed (I hope you've listened to that part of Cerny's talk very closely).

But it's free RAM size doesn't really matter, we still have 2:1 of old/new in one frame, because SSD is only 2x faster?
Yes, and no.
We do have only 2x faster rates (although the max rates are much higher for PS5: 22GB/sec vs 6GB/sec)
But the thing is, GPU can render from 8GB of game data. And XSeX - only from 2.5GB, do you remember that we cannot render from the "streaming" part while it loads?
So in any given scene, potentially, PS5 can have 2x to 3x more details/textures/assets than XSeX.
Yes, XSeX will render it faster, higher FPS or higher frame-buffer resolution (not both, perf difference is too low).
But the scene itself will be less detailed, have less artwork.

OMG, can MSFT do something about it?
Of course they will, and they do!
What are the XSeX advantages? More ALU power (FLOPS) more RT power, more CPU power.
What MSFT will do: rely heavily on this power advantage instead of the artwork: more procedural stuff, more ALU used for physics simulation (remember, RT and lighting is a physics simulation too, after all).
More compute and more complex shaders.

So what will be the end result?
It's pretty simple.
PS5: relies on more artwork and pushing more data through the system. Potentially 2x performance in that.
XSeX: relies more on in-frame calculations, procedural. Potentially 30% performance in that.
Who will win: dunno. There are pros and cons for each.
It will be a fun generation indeed. Much more fun than the previous one, for sure.

There's a lot of confusion on why SSD is so important for next-gen and how it will change things.
Here I will try to explain the main concepts.
TL;DR fast SSD is a game changing feature, this generation will be fun to watch!

It was working fine before, why do we even need that?
No, it wasn't fine, it was a giant PITA for anything other than small multiplayer maps or fighting games.
Let's talk some numbers. Unfortunately not many games have ever published their RAM pools and asset pools to the public, but some did.
Enter Killzone: Shadowfall Demo presentation.
We have roughly the following:

Type Approx. Size, % Approx. Size, MB
Textures 30% 1400
CPU working set 15% 700
GPU working set 25% 1200
Streaming pool 10% 500
Sounds 10% 450
Meshes 10% 450
Animations/Particles 1% 45

*These numbers are rounded sums of various much more detailed numbers presented in the article above.

We are interested in the "streaming pool" number here (but we will talk about others too)
We have ~500MB of data that is loaded as the demo progresses, on the fly.
The whole chunk of data that the game samples from (for that streaming process) is 1600MB.
The load speed of PS4 drive is (compressed data) <50MB/sec (uncompressed is <20MB/sec), i.e. it will take >30sec to load that at least.

It seems like it's not that big of a problem, and indeed for demo it is. But what about the game?
The game size is ~40GB, you have 6.5GB of usable RAM, you cannot load the whole game, even if you tried.
So what's left? We can either stream things in, or do a loading screen between each new section.
Let's try the easier approach: do a loading screen
We have 6.5GB of RAM, and the resident set is ~2GB from the table above (GPU + CPU working set). We need to load 4.5GB each time. It's 90 seconds, pretty annoying, but it's the best case. Any time you need to load things not sequentially, you will need to seek the drive and the time will increase.
You can't go back, as it will re-load things and - another loading screen.
You can't use more than 4.5GB assets in your whole gaming section, or you will need another loading screen.
It gets even more ridiculous if your levels are dynamic: left an item in previous zone? Load time will increase (item is not built into the gaming world, we load the world, then we seek for each item/item group on disk).
Remember Skyrim? Loading into each house? That's what will happen.
So, loading screens are easy, but if your game is not a linear, static, theme-park style attraction it gets ridiculous pretty fast.

How to we stream then?
We have a chunk of memory (remember 500Mb) that's reserved for streaming things from disk.
With our 50MB/sec speed we fill it up each 10 sec.
So, each 10 sec we can have a totally new data in RAM.
Let's do some metrics, for example: how much new shit we can show to the player in 1 min? Easy: 6*500 = 3GB
How much old shit player sees each minute? Easy again: 1400+450+450+45=~ 2.5GB
So we have a roughly 50/50 old to new shit on screen.
Reused monsters? assets? textures? NPCs? you name it. You have the 50/50 going on.

But PS4 has 6.5GB of RAM, we used only 4.5GB till now, what about other 2GB?
Excellent question!
The answer is: it goes to the old shit. Because if we increase the streaming buffer to 1.5GB it still does nothing to the 50MB/sec speed.
With the full 6.5GB we get to 6GB old vs 3GB new in 1 minute. Which is 2:1 old shit wins.

But what about 10 minutes?
Good, good. Here we go!
In 10 min we can get to 30GB new shit vs 6GB old.
And that's, my friends, how the games worked last gen.
You're as a player were introduced to the new gaming moments very gradually.
Or, there were some tricks they used: open doors animation.
Remember Uncharted with all the "let's open that heavy door for 15sec?" that's because new shit needs to load, players need to get to a new location, but we cannot load it fast.

So, what about SSDs then?
We will answer that later.
Let's ask something else.

What about 4K?
With 4K "GPU working set" will grow 4x, at least.
We are looking at 1200*4 = 4.8GB of GPU data.
CPU working set will also grow (everybody wants these better scripts and physics I presume?) but probably 2x only, to 700*2 = ~1.5GB
So overall the persistent memory will be well over 6GB, let's say 6.5GB.
That leaves us with ~5GB of free RAM in XSeX and ~8GB for PS5.

Stop, stop! Why PS5 has more RAM suddenly?
That's simple.
XSeX RAM is divided into two pools (logically, physically it's the same RAM): 10GB and 3.5GB.
GPU working set must use the 10GB pool (it's the memory set that absolutely needs the fast bandwidth).
So 10 - 4.8 = 5.2 which is ~5GB
CPU working set will use 3.5GB pool and we will have a spare 2GB there for other things.
We may load some low freq data there, like streaming meshes and stuff, but it will hard to use in each frame: accessing that data too frequently will lower the whole system bandwidth to 336Mb/sec.
That's why MSFT calls the 10GB pool "GPU optimal".

But what about PS5? It also has some RAM reserved for the system? It should be ~14GB usable!
Nope, sorry.
PS5 has a 5.5GB/sec flash drive. That typically loads 2GB in 0.27 sec. It's write speed is lower, but not less than 5.5GB/sec raw.
What PS5 can do, and I would be pretty surprised if Sony won't do it. Is to save the system image to the disk while the game is playing.
And thus give almost full 16GB of RAM to the game.
2GB system image will load into RAM in <1 sec (save 2GB game data to disk in 0.6 sec + load system from disk 0.3 sec). Why keep it resident?
But I'm on the safe side here. So it's ~14.5GB usable for PS5.

Hmm, essentially MSFT can do that too?
Yep, they can. The speeds will be less sexy but not more than ~3sec, I think.
Why don't they do it? Probably they rely on OS constantly running on the background for all the services it provides.
That's why I gave Sony 14.5GB.
But I have hard time understanding why 2.5GB is needed, all the background services can run on a much smaller RAM footprint just fine, and UI stuff can load on-demand.

Can we talk about SSD for games now?
Yup.
So, let's get to the numbers again.
For XSeX ~5GB of "free" RAM we can divide it into 2 parts: resident and streaming.
Why two? Because typically you cannot load shit into frame while frame is rendering.
GPU is so fast, that each time you ask GPU "what exact memory location are you reading now?" will slow it down to give you an answer.

But can you load things into other part while the first one is rendering?
Absolutely. You can switch "resident" and "streaming" part as much as you like, if it's fast enough.
Anyway, we got to 50/50 of "new shit" to "old shit" inside 1 second now!
2.5GB of resident + 2.5GB of streaming pool and it takes XSeX just 1 sec to completely reload the streaming part!
In 1 min we have 60:1 of new/old ratio!
Nice!

What about PS5 then? Is it just 2x faster and that's it?
Not really.
The whole 8GB of the RAM we have "free" can be a "streaming pool" on PS5.

But you said "we cannot load while frame is rendering"?
In XSeX, yes.
But in PS5 we have GPU cache scrubbers.
This is a piece of silicon inside the GPU that will reload our assets on the fly while GPU is rendering the frame.
It has full access to where and what GPU is reading right now (it's all in the GPU cache, hence "cache scrubber")
It will also never invalidate the whole cache (which can still lead to GPU "stall") but reload exactly the data that changed (I hope you've listened to that part of Cerny's talk very closely).

But it's free RAM size doesn't really matter, we still have 2:1 of old/new in one frame, because SSD is only 2x faster?
Yes, and no.
We do have only 2x faster rates (although the max rates are much higher for PS5: 22GB/sec vs 6GB/sec)
But the thing is, GPU can render from 8GB of game data. And XSeX - only from 2.5GB, do you remember that we cannot render from the "streaming" part while it loads?
So in any given scene, potentially, PS5 can have 2x to 3x more details/textures/assets than XSeX.
Yes, XSeX will render it faster, higher FPS or higher frame-buffer resolution (not both, perf difference is too low).
But the scene itself will be less detailed, have less artwork.

OMG, can MSFT do something about it?
Of course they will, and they do!
What are the XSeX advantages? More ALU power (FLOPS) more RT power, more CPU power.
What MSFT will do: rely heavily on this power advantage instead of the artwork: more procedural stuff, more ALU used for physics simulation (remember, RT and lighting is a physics simulation too, after all).
More compute and more complex shaders.

So what will be the end result?
It's pretty simple.
PS5: relies on more artwork and pushing more data through the system. Potentially 2x performance in that.
XSeX: relies more on in-frame calculations, procedural. Potentially 30% performance in that.
Who will win: dunno. There are pros and cons for each.
It will be a fun generation indeed. Much more fun than the previous one, for sure.

What kind of mind-fuckery is that! That was insane! You deserve a golden medal gift for that insanely-detailed post!

Man, I've just got a brain enlargement treat only by reading all of that.

 

[lepermessiah]

Oh Series X fans are not going to like this tweet:

"Wtf is a "6tflop console", you're just making this up. Teraflop measures one part of the pipeline of one discreet component of a console. You don't call your car a "12 gallon fuel tank saloon" like that means anything"

Matt Phillips, ex-Crytek, Travellers Tales programmer.

Twitter Publish

publish.twitter.com

Yep, boy I've learnt so much about how architecture actually works and what is actually subtracted when everything is actually moving around, the importance of the I/O and the controllers, etc, than just assuming "Oh mORe ShadERs mEanS BeTteR!" Sony made some really smart decisions with their silicon here, I expect a lot of crow will be eaten once we see what it will do.

 

[pasterpl]

xbox = 52 CU
PS5 = 36CU


Says a fanboy?

These are GAMING consoles, not HEY LOOK AT MY TFLOPS! consoles.

So saying one is "best all around" without seeing games, controllers and full reveals = useless trolling.

We already know that PS5 have faster/better SSD, so please explain how xbox is best all around?

and controllers, some like weird xbox stick placement made for mutants, some like symmetrical playstation stick placement made for humans

So there isnt "best all around" console, and specs dont matter much, games will. And we know which system have been more popular among gamers on this gen and which havent.

Pro vs X = cheaper and less powerful system sold more.

Just dont say "best all around" when it is lying and complete bullshit.

Maybe it is just too difficult for xbox fans to admit that even with better specs on some parts = they are still on the losing side because Sony is just much more popular on global scale?

Sony would have to fuck up badly to lose badly, it is about that, not like "if xbox does its best it can win vs Sony at its best"

Xbox can be clear winner probably only in the situation where Playstation does something really horribly wrong, and it doesnt look like it at the moment

fanboy much?

 

[TBiddy]

What kind of mind-fuckery is that! That was insane! You deserve a golden medal gift for that insanely-detailed post!

Man, I've just got a brain enlargement treat only by reading all of that.

He stole it.

(*)SSD and loading times demystified

There's a lot of confusion on why SSD is so important for next-gen and how it will change things. Here I will try to explain the main concepts. TL;DR fast SSD is a game changing feature, this generation will be fun to watch! It was working fine before, why do we even need that? No, it wasn't...

www.neogaf.com

 

[Bo_Hazem]

He stole it.

(*)SSD and loading times demystified

There's a lot of confusion on why SSD is so important for next-gen and how it will change things. Here I will try to explain the main concepts. TL;DR fast SSD is a game changing feature, this generation will be fun to watch! It was working fine before, why do we even need that? No, it wasn't...

www.neogaf.com

And that's one helluva steal then, well deserved

 

[TBiddy]

And that's one helluva steal then, well deserved

True dat. Or could just be he wanted to spread the word, I don't know. He's still new, so let's not be too harsh on him

 

[Bo_Hazem]

True dat. Or could just be he wanted to spread the word, I don't know. He's still new, so let's not be too harsh on him

Not anymore, he's Gold Member now