Baseless Next Generation Rumors with no Technical Merits [post E3 2019, pre GDC 2020] [XBSX, PS5]

[mpg1]

It's hard determine the veracity of anything from these insiders because there's very little detail....it's just usually "both systems are close".

Also if this guy is a VR developer I'm a bit skeptical he has been working with a Xbox dev kit but who knows...

 

[PSman1700]

He didn't say they were close; he said people wouldn't be able to see the difference

Lesson one, always read the text yourself, before commenting, thanks for the heads up. Chris1515 do you automaticly translate everything into PS5 being close or more powerfull?

 

[iroboto]

The guy is currently working in the industry and he is a VR developer very different, he is not even videogame journalist but a developer.

Doesn't mean the information he provides is verified by the mods; which is where I think people have an incorrect understanding of what 'verified' means. But yes, he is verified in claiming to be the person he claims to be.

As I noted earlier, Phil Spencer is definitely verified. He claimed there was going to be VR X1X. There's no VR on X1X. He's as good as it gets right?
This is why I'd rather look at data as being more pure than just someone saying something. Words can and are used all the time to describe the technical details. They aren't the technical details however. I don't doubt for a second that what he wrote is probably true; most people will not notice any difference if there is any. But for the sake of our technical discussion purely based around numbers, it's not helpful.

 

[disco_]

And his comments are more close to Kleegamefan's (GPU difference within 10%).

The github data of PS5 is from old chips and their revision (from Ariel A0 to OBR B0). The true PS5 APU is OBR C0 or E0 which is not leaked.

What are you expecting from those later revisions? Even higher clocks and more CU? It's still the same chip.

 

[chris1515]

He didn't say they were close; he said people wouldn't be able to see the difference

Yes, one will be faster than another (but you will not notice it in practice).​

If one renders a game at 1880p upscaled and the other at 4K native, you (the average Joe) won't notice it in practice despite the lower res machine being 3/4 the power, a significant power deficit.

It's giving Sony fans a 'way-out' of the GitHub Doctrinal fears. You can either take it that they are close, and PS5 isn't 9 TFs, or you can now believe that even if PS5 is only 9 TFs to XBSX's 12 - "what you really should care to know is that, in practice, one could have more TF than the other and give less performance." - PS5 will be faster as it's more balanced. Either way, Sony fans win and can quote this insider to support whatever argument they want to hear.

Sorry but for example, the bad checkerboard rendering of PS4 Pro RDR 2 is very noticeable compared to the native 4k RDR2. Here he said precisely it will be difficult for pixel counter to find the difference.

Lesson one, always read the text yourself, before commenting, thanks for the heads up. Chris1515 do you automaticly translate everything into PS5 being close or more powerfull?

Did I say the PS5 is more powerful? Learn to read. I said it means the console will be nearly the same power exactly like Jason Schreier told. The only moment I talk about Xbox power I told I heard the Xbox GPU is more powerful and the PS5 SSD is much faster for optimized games at least double speed.

 

[PSman1700]

Did I say the PS5 is more powerful? I said it means the console will be nearly the same power exactly like Jason Schreier told. The only moment I talk about Xbox power I told I heard the Xbox GPU is more powerful and the PS5 SSD is faster.

You wrote he said they were close, he didn't even state that at all.

 

[mpg1]

This is why I'd rather look at data as being more pure than just someone saying something. Words can and are used all the time to describe the technical details. They aren't the technical details however. I don't doubt for a second that what he wrote is probably true; most people will not notice any difference if there is any. But for the sake of our technical discussion purely based around numbers, it's not helpful.

Which is why the "lack of context" around the github info is not actually a deficit like some argue...if anything its a strength.

It's it's fake...it's the smartest way you can do a hoax because you are not parsing words, only data.

 

[chris1515]

You wrote he said they were close, he didn't even state that at all.

This generation will NOT be the generation of the visual difference that makes you decide what to buy. It will be the generation of the games. Yes, one will be faster than another (but you will not notice it in practice). One will load the screen in 10 seconds and the other in 7 seconds. One will have a better sound than the other. One will run the game at 60fps and the other ... too. What will really matter will be the quality that development studies can achieve with their time and budget.

And he said it will be difficult for pixel counter and website like Digitalfoundry to find the difference. I don't know what he can tell more. Maybe one will be full native 4k and the other only 10% less in res.

 

[RDGoodla]

He didn't say they were close; he said people wouldn't be able to see the difference

Yes, one will be faster than another (but you will not notice it in practice).​

If one renders a game at 1880p upscaled and the other at 4K native, you (the average Joe) won't notice it in practice despite the lower res machine being 3/4 the power, a significant power deficit.

It's giving Sony fans a 'way-out' of the GitHub Doctrinal fears. You can either take it that they are close, and PS5 isn't 9 TFs, or you can now believe that even if PS5 is only 9 TFs to XBSX's 12 - "what you really should care to know is that, in practice, one could have more TF than the other and give less performance." - PS5 will be faster as it's more balanced. Either way, Sony fans win and can quote this insider to support whatever argument they want to hear.

"What I can say is that you are making films of biblical proportions and based on data that, if true, were 1 year ago in a given circumstance. But do you really believe that real data is filtered? "

What are you expecting from those later revisions? Even higher clocks and more CU? It's still the same chip.

OBR C0~E0 doesn't necessarily mean the same chip (although havng the same code name).

 

[Proelite]

1800p and 4k is close.

I wonder what the narrative will be when physical Xsx devkits is shipped to 3rd parties. AFAIK, they're still working with paper target specs (PC).

 

[iroboto]

And he said it will be difficult for pixel counter and website like Digitalfoundry to find the difference. I don't know what he can tell more.

I believe him.
Honestly I do. It's a reasonable take, and its more or less what I've been thinking as well.
And I also believe when fully utilized, 4Pro competes extremely well against X1X. It is very hard for people to notice the difference.

But when it's not fully utilized, or utilized well, the difference becomes noticeable.
So I also think this type of pattern will show up if it is 9.2 to 12 TF.
Another replay of 4Pro vs X1X.
MS would really need to screw the pooch on hardware balance for XSX for it to be close in final performance but still have a 30% TF advantage.

 

[Nisaaru]

And his comments are more close to Kleegamefan's (GPU difference within 10%).

The github data of PS5 is from old chips and their revision (from Ariel A0 to OBR B0). The true PS5 APU is OBR C0 or E0 which is not leaked.

Sorry but a chip "revision" doesn't just add 20 more CUs. That's a new design.

 

[chris1515]

I believe him.
Honestly I do. It's a reasonable take, and its more or less what I've been thinking as well.
And I also believe when fully utilized, 4Pro competes extremely well against X1X. It is very hard for people to notice the difference.

But when it's not fully utilized, or utilized well, the difference becomes noticeable.
So I also think this type of pattern will show up if it is 9.2 to 12 TF.
Another replay of 4Pro vs X1X.
MS would really need to screw the pooch on hardware balance for XSX for it to be close in final performance but still have a 30% TF advantage.

I think the PS5 will not be 9 Tflops at 2 Ghz. It will be 8 Tflops at 36 CUs or with a GPU with more CUs even if it is 9 Tflops not 56 CUs but maybe 44 Cus and 4 disabled for example.

1800p and 4k is close.

I wonder what the narrative will be when physical Xsx devkits is shipped to 3rd parties. AFAIK, they're still working with paper target specs (PC).

Some third-party have the two devkits. The big 5 publisher studios working for title releasing in 2020/2021 at least.

 

[Proelite]

I think the PS5 will not be 9 Tflops at 2 Ghz. It will be 8 Tflops at 36 CUs or with a GPU with more CUs even if it is 9 Tflops not 56 CUs but maybe 44 Cus and 4 disabled for example.

Some third-party have the two devkits. The big 5 publisher studios working for title releasing in 2020/2021 at least.

Consider the timing of when the chips shots for Xsx were released, Jan 8th. Add at least a month to that.

The devkits for the third-parties have are all PCs.

Scarlett is about a month ahead of Scorpio. Scorpio kits went out for studios in March 2017.

 

[N2O]

If they have a "re-design" in 2019, expect PS5 in 2023 then.
Oh wait there will be, it's called PS5 Pro.

 

[RDGoodla]

Sorry but for example, the bad checkerboard rendering of PS4 Pro RDR 2 is very noticeable compared to the native 4k RDR2. Here he said precisely it will be difficult for pixel counter to find the difference.


Did I say the PS5 is more powerful? Learn to read. I said it means the console will be nearly the same power exactly like Jason Schreier told. The only moment I talk about Xbox power I told I heard the Xbox GPU is more powerful and the PS5 SSD is much faster for optimized games at least double speed.

https://www.neogaf.com/threads/next...-analysis-leaks-thread.1480978/post-256731237
BGs said "At the beginning of the year PS5 was still the one that offered the best performance."

Sorry but a chip "revision" doesn't just add 20 more CUs. That's a new design.

Yes previous discussion already mentioned a example of new design with the same code name (sorry I forget where the post is).

 

[Shifty Geezer]

Sorry but for example, the bad checkerboard rendering of PS4 Pro RDR 2 is very noticeable compared to the native 4k RDR2. Here he said precisely it will be difficult for pixel counter to find the difference.

And if upscaling tech doesn't improve, that'd remain the case. You think upscaling tech is going to stand still? You think, for example, Sony's acquisition of Insomniac and their superb Temporal Injection won't find its way into the PS5 libraries for other games to use? What if, for example, Sony's customised solution contains the ID buffer of PS4Pro but now much better targeted by all games and some good experience behind it? We'd have a console that could be less powerful but better 'balanced' and the results hard for people to tell the difference.

You need a leap of faith to get to the conclusion that both consoles are close in power. The statement is open ended and equally supports the possibilities of both being similar in power, or one being a fair bit behind in TFs but looking (subjective) similar on the TVs in the games people will be playing.

 

[N2O]

I believe him.
Honestly I do. It's a reasonable take, and its more or less what I've been thinking as well.
And I also believe when fully utilized, 4Pro competes extremely well against X1X. It is very hard for people to notice the difference.

But when it's not fully utilized, or utilized well, the difference becomes noticeable.
So I also think this type of pattern will show up if it is 9.2 to 12 TF.
Another replay of 4Pro vs X1X.
MS would really need to screw the pooch on hardware balance for XSX for it to be close in final performance but still have a 30% TF advantage.

Well i know the advantage of XSX for other parts are not as big as GPU TFs(let's say the rumor is real), hell the CPU part could very well be same speed.
But doesn't mean it's unbalanced, i mean come on PS5 isn't Anakin Skywalker, in fact who knows which one is more "balance" til the console released.

 

[iroboto]

Well i know the advantage of XSX for other parts are not as big as GPU TFs(let's say the rumor is real), hell the CPU part could very well be same speed.
But doesn't mean it's unbalanced, i mean come on PS5 isn't Anakin Skywalker, in fact who knows which one is more "balance" til the console released.

Well there's some bottlenecks. If we are fair to the idea that Github leaks apply to Oberon, we must apply the same concepts to Arden.
There is a serious lack of memory bandwidth for double the power of X1X.
Double X1X we should be looking at 650 GB/s of bandwidth. The GitHub leaks are showing ~520/540 IIRC.

Things like that are an issue, and there could be others.

 

[MrFox]

I believe him.
Honestly I do. It's a reasonable take, and its more or less what I've been thinking as well.
And I also believe when fully utilized, 4Pro competes extremely well against X1X. It is very hard for people to notice the difference.

But when it's not fully utilized, or utilized well, the difference becomes noticeable.
So I also think this type of pattern will show up if it is 9.2 to 12 TF.
Another replay of 4Pro vs X1X.
MS would really need to screw the pooch on hardware balance for XSX for it to be close in final performance but still have a 30% TF advantage.

Pro vs 1X was 50% more memory, 50% more bandwidth, and 50% more TF. The studios using ID buffer and RPM have reduced some of that difference in practice but only for the TF aspect.

Here it's nowhere near that gap in TF, and memory amount/bw are not expected to be very different either. So it's a reasonable guess it won't be noticeable considering we add VRS which will mess DF attempt at pixel counting.

Also a higher clock means the other sections of the GPU are faster, depending how those can or cannot also be scaled wider like the CU count. Other factors to consider includes their RT approach, CPU clock, caches, and whether the new audio section can seriously free up some CPU processing power.

It remains a warning against following the TF number as a complete metric. The SoC has a lot more than just CUs in it, and we don't even know how different these CUs will be.