Support NeoGAF

[JareBear: Remastered]

That was how it worked in gens past.

Its easier to develop for the most limited platform and port that code to others.

For SSD matters, multiplats can just use downgraded textures where necessary if Xsex should have trouble keeping up. Personally I don’t think it will.

Couldn't it make more sense to develop on XsX, max it out for whatever XsX can do, and then just utilize any extra enhancements when porting to PS5?

 

[NullZ3r0]

Microsoft will try as hard as they can to downplay the PS5 SSD!! The difference is much bigger than the TFlops difference.

The 2 TFLOPS difference will show on the screen while the SSD difference will not. Both systems are orders of magnitude faster in terms of storage than their predecessors. So even if the PS5 storage was 20 times faster, the we're talking Xbox users waiting 5 seconds for something to load which beats 40 seconds to a minute this Gen.

But the PS5 isn't 20x faster storage wise. At best its twice as fast. So in the real world gamers won't notice a difference. But they will notice playing games with reduced framerates and resolution due to the 2 TFLOPS difference.

 

[rntongo]

No I'm not. I've said that that figure is compared to loading the entire texture. So no PRT or SF. What you're doing though is using that figure to link to another official twitter user mentioning something 'custom' and pretending this is some secret sauce that would give it 2x or 3x multiplier compared to the competition. This isn't the case. When that other official twitter user was asked about this figure and it being custom or not they didn't say it was and said they don't know the figure.

You're really confusing me. Microsoft stated that SFS will provide 2-3x multiplier for RAM and SSD. When a Microsoft engineer working on SFS was asked, he declined to comment but gave scenario of 4x improvement by using only 25% of a texture. What more do you want? All the tweets have been posted here.

 

[martino]

You're really confusing me. Microsoft stated that SFS will provide 2-3x multiplier for RAM and SSD. When a Microsoft engineer working on SFS was asked, he declined to comment but gave scenario of 4x improvement by using only 25% of a texture. What more do you want? All the tweets have been posted here.

we want them to show it to us.

 

[quest]

we want them to show it to us.

Can I have white paper and exactly spec sheet on the variable clocks? Not marketing speak like couple, majority ect. And let people make up that variable means fixed around here.

 

[ToadMan]

Like I said the poster was clearly talking about the 12 NAND chips and lanes/channels are synonyms.

NXGamer put up this video today. I think it relevant to this discussion.

Lanes and channels are not synonymous sorry. PCIe has a specific definition of what a lane is.

I only mention this because whichever post it was I first responded to there seemed to be some maths using 12 channels as if they’re Lanes - I didn’t want maths to be screwed up by a simple error.

 

[Thirty7ven]

I’ll be buying the XsX day one. Not even a question. I would get a ps5 as well if I had the time to play them both

Then that’s it papa, only thing that matters.

 

[martino]

Can I have white paper and exactly spec sheet on the variable clocks? Not marketing speak like couple, majority ect. And let people make up that variable means fixed around here.

i want that too.

 

[truth411]

The 2 TFLOPS difference will show on the screen while the SSD difference will not. Both systems are orders of magnitude faster in terms of storage than their predecessors. So even if the PS5 storage was 20 times faster, the we're talking Xbox users waiting 5 seconds for something to load which beats 40 seconds to a minute this Gen.

But the PS5 isn't 20x faster storage wise. At best its twice as fast. So in the real world gamers won't notice a difference. But they will notice playing games with reduced framerates and resolution due to the 2 TFLOPS difference.

This post is disingenuous, The focus of PS5 SSD is not just for load times but for streaming in game Assets. Which is over 2x as fast as the XSX.
The 1.8Tflops advantage the XSX have is 18%, thats not that significant.
In other words with that Unreal Engine 5 example assuming everything else is identical with the PS5 including SSD speeds (which is not).
But hypothetically if everthing is the same except 18% Tflops. It would mean

PS5 unreal 5 demo = 1440p/30fps
XSX unreal 5 demo = ~1504p/30fps

Thats your 18% resolution difference. Not that significant.

 

[Panajev2001a]

You're really confusing me. Microsoft stated that SFS will provide 2-3x multiplier for RAM and SSD. When a Microsoft engineer working on SFS was asked, he declined to comment but gave scenario of 4x improvement by using only 25% of a texture. What more do you want? All the tweets have been posted here.

Yes and in none of those tweets not the tech video posted describing the technique itself or the speaker notes/comments support either SF or SFS delivering a consistent 2-3x bandwidth or storage multiplier over PRT usage which is a mind boggling performance increase. You are underselling how difficult it is to get 15-20% performance improvements by hyping up 200% improvements out of thin air.

 

[martino]

This post is disingenuous, The focus of PS5 SSD is not just for load times but for streaming in game Assets. Which over 2x as fast as the XSX.
The 1.8Tflops advantage the XSX have is 18%, thats not that significant.
In other words with that Unreal Engine 5 example assuming everything else is identical with the PS5 including SSD speeds (which is not).
But hypothetically if everthing is the same except 18% Tflops. It would mean

PS5 unreal 5demo = 1440p/30fps
XSX unreal 5 demo = ~1504p/30fps

Thats your 18% resolution difference. Not that significant.

in your example it's also 18% less pixels to display smaller details.

 

[THE:MILKMAN]

Lanes and channels are not synonymous sorry. PCIe has a specific definition of what a lane is.

I only mention this because whichever post it was I first responded to there seemed to be some maths using 12 channels as if they’re Lanes - I didn’t want maths to be screwed up by a simple error.

Fine but understand this is just being pedantic. We both understood what Oldgamer was saying even if they were using the wrong technical terms or is wrong with the speculation.

Maybe you could quote Oldgamer's post and explain to both of us how the 12 channels/chips/4 PCIe lanes adds up in speeds (5.5GB/s) and data (825GB)?

 

[Oddvintagechap]

The 2 TFLOPS difference will show on the screen while the SSD difference will not. Both systems are orders of magnitude faster in terms of storage than their predecessors. So even if the PS5 storage was 20 times faster, the we're talking Xbox users waiting 5 seconds for something to load which beats 40 seconds to a minute this Gen.

But the PS5 isn't 20x faster storage wise. At best its twice as fast. So in the real world gamers won't notice a difference. But they will notice playing games with reduced framerates and resolution due to the 2 TFLOPS difference.

PS4
Ram used to load things you might see. Such as things behind you. Ram is packed with data you might never interact with but it has to be on the RAM in case you do.
PS5
Everything can be loaded from the SSD into RAM under a second. So everything in RAM is things you can see. The moment you dont see it, its dumped.

This isnt just loading. This is fundamental different game design.

'playing a PS5 game should be as easy as Netflix.'

PS5 wont even have a suspend mode. Games can have save states and be immediately loaded when you open them. Completely dumped from the Ram when youre not playing them. So you can switch from any game on your SSD from where you left off. The moment your cursor is on a game’s icon the game is already loaded.

DYNAMIC INTERFACES FOR LAUNCHING DIRECT GAMEPLAY

Complete Patent Searching Database and Patent Data Analytics Services.

www.freepatentsonline.com

"Multiplayer game servers will provide the console with the set of joinable activities in real time. Single-player games will provide information like what missions you could do and what rewards you might receive for completing them – and all of those choices will be visible in the UI. As a player you just jump right into whatever you like."

will Sony have any games on launch to display these advantages? Nah. Probably Knack 3.

 

[Dory16]

This post is disingenuous, The focus of PS5 SSD is not just for load times but for streaming in game Assets. Which over 2x as fast as the XSX.
The 1.8Tflops advantage the XSX have is 18%, thats not that significant.
In other words with that Unreal Engine 5 example assuming everything else is identical with the PS5 including SSD speeds (which is not).
But hypothetically if everthing is the same except 18% Tflops. It would mean

PS5 unreal 5demo = 1440p/30fps
XSX unreal 5 demo = ~1504p/30fps

Thats your 18% resolution difference. Not that significant.

We’re speaking about 16 RDNA2 CUs here. There’s no real world data to measure the pixel pushing improvement but it’s not as simple as difference percentage = difference in number of pixels. It all depends on what the devs want to do with it. Asymetric computing means they can use it as additional cpu power and bolster AI or frame rates. And we’re speaking about more than the power of a base PS4 so believe me, it is significant although it won’t appear as such in every game. In games where the devs want it to be, it will be quite noticeable.

 

[NickFire]

PS4
Ram used to load things you might see. Such as things behind you. Ram is packed with data you might never interact with but it has to be on the RAM in case you do.
PS5
Everything can be loaded from the SSD into RAM under a second. So everything in RAM is things you can see. The moment you dont see it, its dumped.

This isnt just loading. This is fundamental different game design.

'playing a PS5 game should be as easy as Netflix.'

PS5 wont even have a suspend mode. Games can have save states and be immediately loaded when you open them. Completely dumped from the Ram when youre not playing them. So you can switch from any game on your SSD from where you left off. The moment your cursor is on a game’s icon the game is already loaded.

DYNAMIC INTERFACES FOR LAUNCHING DIRECT GAMEPLAY

Complete Patent Searching Database and Patent Data Analytics Services.

www.freepatentsonline.com

"Multiplayer game servers will provide the console with the set of joinable activities in real time. Single-player games will provide information like what missions you could do and what rewards you might receive for completing them – and all of those choices will be visible in the UI. As a player you just jump right into whatever you like."

will Sony have any games on launch to display these advantages? Nah. Probably Knack 3.

Pardon me - coming in late, and this stuff is confusing.

Question - my layman brain says what you wrote sounds awesome. I'm not following why people are arguing over the SSD though. Maybe I need to digest the information more, but isn't the point of this thread that Series X also has at least part of ssd available to use like Sony is doing? If so, what is the actual difference that people are debating?

 

[ToadMan]

Couldn't it make more sense to develop on XsX, max it out for whatever XsX can do, and then just utilize any extra enhancements when porting to PS5?

No for a few reasons

1) the PS5 has a lower cpu clock and less tflop gpu. Develop games for PS5 at acceptable performance and they should maintain that performance on Xsex with little or no additional work.

2) PS3. That was when PS target and port to xbox became popular. The ps3 was a dog to develop on so that was the target. Get it running ok there and then moving to 360 was easy. Those teams kept that approach through the PS4/xb1 gen and here we are. There’s no compelling reason to change that approach and indeed there are reasons to stick with it.

3) The most important multiplat engine - UE5 - seems to be targeting PS5 first. So that will be the way developers using UE will go if they’re producing a multiplat.

4) XVA. MS have done a wonderful job of keeping the relatively limited transfer speed of their storage solution from inhibiting their console. While PS5 has the raw throughput to throw around large amounts of data, MS have given the Xsex a set of ways to transfer that same data without compromise. Xsex will keep up with PS5 - in the unlikely scenario it can’t, a reduction in data fidelity will be enough. It’s simple work to reduce texture resolution compared to trying to optimise code already written for a slower platform.


This gen, the differences between the machines is minimal. There’s not gonna be a whole bunch of multiplat studios spending extra 100s of man hours of effort to get a few extra frames out of Xsex. They’ll build for PS5, dump it on Xsex and move on to the next project. There just isn’t enough headroom to make the extra work worthwhile.

 

[dxdt]

I think the same exact way. What Sony did with the SSD is smart by really pushing the I/O throughput to 9GB/s. On the other hand what MSFT did with the GPU/CPU and SFS is impressive as well. I look forward to seeing the results of both machines and how they utilize RAM.

I think MS has a more generalized solution for PC and Xbox.

 

[DarkMage619]

The 2 TFLOPS difference will show on the screen while the SSD difference will not. Both systems are orders of magnitude faster in terms of storage than their predecessors. So even if the PS5 storage was 20 times faster, the we're talking Xbox users waiting 5 seconds for something to load which beats 40 seconds to a minute this Gen.

But the PS5 isn't 20x faster storage wise. At best its twice as fast. So in the real world gamers won't notice a difference. But they will notice playing games with reduced framerates and resolution due to the 2 TFLOPS difference.

I still don't understand why there is always an assumption that the PS5 will have only a 2 TF performance delta with regards to the GPU. Was there an update stating that PS5 will now have a fixed clock rate so that is will never dip below 2 TF delta in graphics performance? Only multiplatform titles will show just which design choices made the difference in game performance. This will be even more interesting if PS5 is the target platform for game development. I'd expect the PS5 to have the best version of multiplatform games since the games would be developed with that platform in mind.

 

[truth411]

We’re speaking about 16 RDNA2 CUs here. There’s no real world data to measure the pixel pushing improvement but it’s not as simple as difference percentage = difference in number of pixels. It all depends on what the devs want to do with it. Asymetric computing means they can use it as additional cpu power and bolster AI or frame rates. And we’re speaking about more than the power of a base PS4 so believe me, it is significant although it won’t appear as such in every game. In games where the devs want it to be, it will be quite noticeable.

Sort of but its not that simplistic, M.S. and Sony CUs are not running at the same speeds.
M.S. chose a wider, slower ( no its not slow) design.
Sony chose a narrower, faster design, with much faster streaming capability.
Both have advantages/disadvantages.

I'm just saying you guys are blowing the XSX 18% Tflop advantage out of proportion.

And underestimating the ramifications of Sony console design. Have to wait for the 1st, 2nd party exclusives, but the throughput is so fast it could fundamentally change Level Design.

 

[rntongo]

I think MS has a more generalized solution for PC and Xbox.

They do and there are benefits and positives to that. But they built custom hardware not available in RDNA2 and will not be available for PC. The tweets from a Microsoft engineer have been posted here.

 

[rntongo]

Yes and in none of those tweets not the tech video posted describing the technique itself or the speaker notes/comments support either SF or SFS delivering a consistent 2-3x bandwidth or storage multiplier over PRT usage which is a mind boggling performance increase. You are underselling how difficult it is to get 15-20% performance improvements by hyping up 200% improvements out of thin air.

Your agenda clear. you keep on bringing up demos that were not using SFS nor direct storage or even the hw decomp block.

 

[JareBear: Remastered]

I still don't understand why there is always an assumption that the PS5 will have only a 2 TF performance delta with regards to the GPU. Was there an update stating that PS5 will now have a fixed clock rate so that is will never dip below 2 TF delta in graphics performance? Only multiplatform titles will show just which design choices made the difference in game performance. This will be even more interesting if PS5 is the target platform for game development. I'd expect the PS5 to have the best version of multiplatform games since the games would be developed with that platform in mind.

Well, a lot of people believe that the PS5 can operate at 10.28TF basically all the time, and the quotes regarding that seem to indicate it could or couldn't, depending on how you interpret the wordings.

My question is this:

If the CPU has to be throttled back to sustain max GPU output, and the 10.28TF threshold, would the slower cpu possibly have a negative impact on optimally accessing the SSD in any way whatsoever?

 

[rntongo]

we want them to show it to us.

Me too.

 

[Ascend]

I did my best to explain SFS. But for some reason, that post is largely ignored. I wonder why....

 

[Panajev2001a]

Your agenda clear. you keep on bringing up demos that were not using SFS nor direct storage or even the hw decomp block.

The extra efficiency of BCPack is known and estimated by MS: XSX advertises 2.4 GB/s uncompressed and ~4.8 GB/s compressed, PS5 advertises 5.5 GB/s uncompressed and ~8-9 GB/s compressed (so you can see the relative improvement thanks to BCPack is not in the orders of magnitude you were talking about earlier).
No magic brought by BCPack or Kraken (on XSX you have BCPack optimised for texture data + lmza for general data while PS5 uses Kraken globally which is a middle ground in terms of efficiency between those other two choices).

Interesting how we are moving the goalposts now: first you are talking about SFS bringing in a 2-3x multiplier in storage (edit: typed memory instead of storage) bandwidth and memory storage and without batting an eyelid bring in direct storage in the picture and HW decompression. Let’s muddle the waters some more by bringing in other factors and make the baseline of the comparison more arbitrary and handwavey, won’t we?

 

[rntongo]

The extra efficiency of BCPack is known and estimated by MS: XSX advertises 2.4 GB/s uncompressed and ~4.8 GB/s compressed, PS5 advertises 5.5 GB/s uncompressed and ~8-9 GB/s compressed (so you can see the relative improvement thanks to BCPack is not in the orders of magnitude you were talking about earlier).
No magic brought by BCPack or Kraken (on XSX you have BCPack optimised for texture data + lmza for general data while PS5 uses Kraken globally which is a middle ground in terms of efficiency between those other two choices).

Interesting how we are moving the goalposts now: first you are talking about SFS bringing in a 2-3x multiplier in memory bandwidth and memory storage and without batting an eyelid bring in direct storage in the picture and HW decompression. Let’s muddle the waters some more by bringing in other factors and make the baseline of the comparison more arbitrary and handwavey, won’t we?

Now you’re confusing memory bandwidth for I/O bandwidth. Why am I still even talking to you?

 

[Three]

You're really confusing me. Microsoft stated that SFS will provide 2-3x multiplier for RAM and SSD. When a Microsoft engineer working on SFS was asked, he declined to comment but gave scenario of 4x improvement by using only 25% of a texture. What more do you want? All the tweets have been posted here.

MS stated 2x-3x (or more) multplier when transferring only part of a texture (no SF or PRT) as SF is part of SFS. This sort of thing is also possible on old GPUs. The other tweet mentioned some other features of SFS for texture filters. You took that to mean SFS is custom and is providing 2-3x the performance over other GPUs.
What is confusing, you were using the 2x-3x figure as some new secret sauce feature when in actual fact it's in comparison to just transferring the whole texture which doesn't happen with SF or PRT anyway.


The person who made that comment was asked is the 2x-3x figure correct and is it only on xbox they answered that they don't know the figure and their reply did not answer the question about xbox only.

 

[BrentonB]

Well, a lot of people believe that the PS5 can operate at 10.28TF basically all the time, and the quotes regarding that seem to indicate it could or couldn't, depending on how you interpret the wordings.

My question is this:

If the CPU has to be throttled back to sustain max GPU output, and the 10.28TF threshold, would the slower cpu possibly have a negative impact on optimally accessing the SSD in any way whatsoever?

Sony has completely offloaded the decompression of data coming from the SSD to the I/O complex. CPU speed shouldn't have an impact on this although there may be fewer data requests per second depending on the processor's speed. I think that's how it works, anyway.

 

[rntongo]

MS stated 2x-3x (or more) multplier when transferring only part of a texture (no SF or PRT) as SF is part of SFS. This sort of thing is also possible on old GPUs. The other tweet mentioned some other features of SFS for texture filters. You took that to mean SFS is custom and is providing 2-3x the performance over other GPUs.
What is confusing, you were using the 2x-3x figure as some new secret sauce feature when in actual fact it's in comparison to just transferring the whole texture which doesn't happen with SF or PRT anyway.


The person who made that comment was asked is the 2x-3x figure correct and is it only on xbox they answered that they don't know the figure and their reply did not answer the question about xbox only.

What you just said below here is a lie:
MS stated 2x-3x (or more) multplier when transferring only part of a texture (no SF or PRT)


This is from Microsoft's official page you can read about it HERE:

Because it avoids the wastage of loading into memory the portions of textures that are never needed, it is an effective 2x or 3x (or higher) multiplier on both amount of physical memory and SSD performance.

Please stop the lies.

 

[funking giblet]

MS stated 2x-3x (or more) multplier when transferring only part of a texture (no SF or PRT) as SF is part of SFS. This sort of thing is also possible on old GPUs. The other tweet mentioned some other features of SFS for texture filters. You took that to mean SFS is custom and is providing 2-3x the performance over other GPUs.
What is confusing, you were using the 2x-3x figure as some new secret sauce feature when in actual fact it's in comparison to just transferring the whole texture which doesn't happen with SF or PRT anyway.


The person who made that comment was asked is the 2x-3x figure correct and is it only on xbox they answered that they don't know the figure and their reply did not answer the question about xbox only.

SFS is a fancy name for PRT or Virtual Textures. UE5 implements their own version.

 

[THE:MILKMAN]

Please stop the lies.

What numbers for physical RAM and SSD performance do you think apply, rntongo? Just so we're clear here.

 

[rntongo]

What numbers for physical RAM and SSD performance do you think apply, rntongo? Just so we're clear here.

I honestly have nothing to say besides what's on the official page. It is up to MSFT to prove the figure they claim.

 

[rntongo]

I didn't actually suggest that, I said 'downplaying good Xbox news' in an Xbox thread in the way you pointedly accused others of doing to Sony. But maybe you're right, how could I possibly get that impression?

Honestly I don't understand what his aim is.

 

[TBiddy]

Honestly I don't understand what his aim is.

It’s obvious. He’s hiding it behind smileys, but he’s no better than every other fanboy on here.

What I don’t understand is the motive.

 

[Three]

What you just said below here is a lie:
MS stated 2x-3x (or more) multplier when transferring only part of a texture (no SF or PRT)


This is from Microsoft's official page you can read about it HERE:


Please stop the lies.

So your'e not reading the text or what?

Because it avoids the wastage of loading into memory the portions of textures that are never needed, it is an effective 2x or 3x (or higher) multiplier

I wonder what else does that? Oh that's right, PRT and SF. Is that custom or new? No.

So why did you say I lied here:

You know this is a lie? I didn't read the rest of what you were saying when I saw this:
What you're not getting is that the 2x efficiency (not performance in the spec) exists already on other GPUs in SF and PRT.

When I ask you OK so where is this 2x coming from do you think? You essentially wave your hands around shouting "I don't know secret sauce." so tell me where do YOU think the 2x comes from to be suggesting I'm being dishonest?
Matt already answered your bull on the other forum.

 

[THE:MILKMAN]

I honestly have nothing to say besides what's on the official page. It is up to MSFT to prove the figure they claim.

I think that is a little unfair given your stance so far but OK. If I'm honest I think you'll end up disappointed about what SFS brings (I'm sure they'll be improvements).

 

[MasterCornholio]

I honestly have nothing to say besides what's on the official page. It is up to MSFT to prove the figure they claim.

I'm guessing it's what their spec sheet says if everything works out. If it doesn't the results will be worse.

Right now it seems they are having some sort of issues with the software which is stopping them from reaching the potential of their I/O system.

 

[oldergamer]

Sheesh! People here act like they have never before saw a hardware feature that can both reduce bandwidth and memory usage and bring an effective multiplier to system resources. It hasn't been that long since power vr was a thing.

Not saying that is what this is, but calling any tech that can help maximizes performance "magical" is pathetic.

 

[Trueblakjedi]

Microsoft isnt using any server technology in the XSX. Those decompression blocks in a server environment are feeding a raid array made of up

Actually, XSX uses 16 GB GDDR6 ECC memory. ECC has never been used in a console before. Why would they do that?


Although i like your deadpan style of posting... some of the assertions you make arent true. We can start with this but im just now eading through the thread.

 

[ToadMan]

Fine but understand this is just being pedantic. We both understood what Oldgamer was saying even if they were using the wrong technical terms or is wrong with the speculation.

Maybe you could quote Oldgamer's post and explain to both of us how the 12 channels/chips/4 PCIe lanes adds up in speeds (5.5GB/s) and data (825GB)?

I had a quick look but I don’t really understand what question you’re even trying to answer.

The PS5 SSD is using pcie 4x lanes the same as the Xsex. That has a theoretical limit of 7.8gb/s - so Sony’s 5.5GB/s is well within the specification.

 

[Three]

Sheesh! People here act like they have never before saw a hardware feature that can both reduce bandwidth and memory usage and bring an effective multiplier to system resources. It hasn't been that long since power vr was a thing.

Not saying that is what this is, but calling any tech that can help maximizes performance "magical" is pathetic.

We know where it comes from the issue is that people are trying to make it seem like it's from something secret and calling others liars when they have no additional information at all.

 

[THE:MILKMAN]

I had a quick look but I don’t really understand what question you’re even trying to answer.

The PS5 SSD is using pcie 4x lanes the same as the Xsex. That has a theoretical limit of 7.8gb/s - so Sony’s 5.5GB/s is well within the specification.

Just trying to figure out what specs each of the 12 chips will have (MT/s) and what current NVMe drives have the same or similar spec chips. I'm bored and a nerd.....

 

[MasterCornholio]

We know where it comes from the issue is that people are trying to make it seem like it's from something secret and calling others liars when they have no additional information at all.

That's something that always bothers me about hardware speculation. The typical we get specs and it turns out that one system is weaker in one point than the other. From there there's speculation that sounds like made up Mr X stuff that tries to make the weakness a strength.

I love speculating on hardware but we should always set the limit at whatever the manufacturer gives us. If it gets better in the future they Will give us updated information.

 

[oldergamer]

We know where it comes from the issue is that people are trying to make it seem like it's from something secret and calling others liars when they have no additional information at all.

That's pure BS, nobody is trying to make any feature "seem like it's from something secret". People in an xbox thread wan to talk about something that we know nothing about. leave it at that. Also, what reaction do you expect when people quote someone connected to or works at MS, and then "others" try to call them out like it was information thy were making up.

 

[MasterCornholio]

Also, what reaction do you expect when people quote someone connected to or works at MS, and then "others" try to call them out like it was information thy were making up.

From what I've seen here it's not the quoting that's the problem it's the interpretation of the information that's given. Goes for both sides BTW.

 

[DForce]

the thing is, we all know that the PS5 is 10.2 tflops at its absolute max. in the real world it will hover around 9.2, which is why it was tested at that on the github leaks. the tflop difference is actually more than 18%.

We know this is a narrative people keep pushing with no facts.

 

[truth411]

the thing is, we all know that the PS5 is 10.2 tflops at its absolute max. in the real world it will hover around 9.2, which is why it was tested at that on the github leaks. the tflop difference is actually more than 18%.

This is false... Stop with the 9.2Tflop nonsense.

 

[rntongo]

Actually, XSX uses 16 GB GDDR6 ECC memory. ECC has never been used in a console before. Why would they do that?


Although i like your deadpan style of posting... some of the assertions you make arent true. We can start with this but im just now eading through the thread.

Thats interesting! I had no idea. Do you have a source I can read about it

 

[THE:MILKMAN]

Thats interesting! I had no idea

The XSX SoC is going to be used for xCloud servers. That surprised me.

 

[JareBear: Remastered]

the thing is, we all know that the PS5 is 10.2 tflops at its absolute max. in the real world it will hover around 9.2, which is why it was tested at that on the github leaks. the tflop difference is actually more than 18%.

We know this is a narrative people keep pushing with no facts.

Maybe this is a dumb question but is there even a way to ever know for sure? Can the guys at DF or any expert analysis person like NXGamer actually measure a consoles TFLOPs in use? Or will we always just kinda be guessing.