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

[Picao84]

No way, Renior is 150mm^2 and it comes with 8 zen 2 cores and only 8CUs for slightly more than 1TF. You'd need at least to be bigger than that for 4TF.

Wait! Slightly more than 1TF? From my calculations it could potentially reach 2TFlops:

64 SP * 8 CUs * 1,7 GHz * 2 = 1740 GFlops
At 2 GHz it would reach 2 TFlops.

Still makes 56 CUs unlikely, more like 48 maybe. But that makes 12TFlops more unlikely.

 

[AlphaWolf]

In that case that would mean that the XSX having 56 CUs in 380mm2 is not possible. I was basing my calculations on the info we have about it.

Edit - Renoir still uses Vega, instead of Navi, although I doubt the difference would be that big.

And what if the XSX apu was 407mm2? As some have measured it.

 

[iroboto]

Scarlett reveal indicated a 380mm^2

But after the new SOC pics:

Some ambiguity remains, of course. Fundamentally, measurements are being derived from a low quality, low resolution image pushed through untold Photoshop filters. This may explain why other commentators used the distance between the surface-mounted components as a basis for measurement. Personally, I thought I'd try a hybrid approach - using both pitch and component-size, scaling the image until components in all corners of the image would line up when the Scarlett processor was overlaid onto the Xbox One X chip. My result is 405mm2, so just a touch larger than Wccftech's estimate - almost 13 per cent larger than the Scorpio engine. A compelling analysis on Reddit from user _rogame based on the later, much higher quality David Prien image gives an broadly equivalent result - 407mm2.

https://www.eurogamer.net/articles/...crosoft-xbox-series-x-silicon-reveal-analysis

 

[Picao84]

And what if the XSX apu was 407mm2? As some have measured it.

Well, the XBox One X APU managed roughly 50% more TFlops than the PS4 Pro with only an extra 35mm2 of die area. Although they were not equivalent feature wise (e.g. no rapid packed math on X). So I wouldn't say no.

However that Renoir information that I was ignorant to greatly cooled down my expectations. Especially if the PS5 die size is really only around 320 mm2, as rumoured. Didnt know that 8 cores of Zen 2, even on 7Nm still occupied 75mm2. That's much larger than I thought.

 

[Nesh]

Compared to XSX, that's maybe a 3TF difference, i wouldn't call that a nightmare. Historically consoles have been below PC but never has that been a 'nightmare'. Even the PS2 was below the PC in total gflops but it did MGS2 already in 2001 anyway.

But not a simultaneous launch and also had the benefit of a huge load of exclusive third party games.
PS5 will lose the old trend where it was the most powerful console at launch for the first time and will have barely the exclusive titles PS2 was enjoying.

 

[MrFox]

If nothing else it likely explains why real data is hard to come by...both MS and Sony clamping down big time.

What I will say is there's a lot of smoke around one console performing better than the other but minimal difference (probably XB>PS), and one loads faster than the other but both load fast so minimal difference (likely PS>XB)

He's really not saying much, just implies things. The one with less TF in theory is not proportionally less powerful in practice. But that could only happen if they modified RDNA, or if there's a bottleneck on the other.

Supposing they would accept a bit larger CU size (they did so after all with the Pro in order to add FP16 DRM), what can they reasonably customize without altering the architecture too much?

In the past Mark Cerny was obsessed with fine grain compute to improve occupancy. Is there still some margin left to improve the CU occupancy or are we now close to the max possible?

Lisa Su said....
"This has been a really long-term partnership with them. What we have done with Sony is really architect something for their application"

 

[Aaron Elfassy]

One argument I haven't seen made is in regards to what Devs are working towards as a hardware target. Over the last 2 years it's clear that many graphic programmers have been targeting 1080ti performance when creating new rendering techniques. Peruse any papers from the last 2-3 years and you will see that is the target card for testing. Most likely the new consoles will be in the 9-10 tflop territory, if we go by this info.

 

[Picao84]

One argument I haven't seen made is in regards to what Devs are working towards as a hardware target. Over the last 2 years it's clear that many graphic programmers have been targeting 1080ti performance when creating new rendering techniques. Peruse any papers from the last 2-3 years and you will see that is the target card for testing. Most likely the new consoles will be in the 9-10 tflop territory, if we go by this info.

Yup, I'm suddenly convinced that neither one will exceed 10 TFlops, unless they would have discrete GPUs. Not with Zen 2 occupying a massive 75mm2 of the die.

Edit - Wait a second. The 75mm2 includes a large amount of cache that is rumoured to be cut substantially for the consoles. The cores alone are less than half of that, at 31.3 mm2! So maybe they could only occupy 50 mm2 or less, freeing 25 mm2 for other stuff.

 

[N2O]

But not a simultaneous launch and also had the benefit of a huge load of exclusive third party games.
PS5 will lose the old trend where it was the most powerful console at launch for the first time and will have barely the exclusive titles PS2 was enjoying.

Well Xbox and Playstation only released together once, which is this gen.
This "first time" really means nothing.

 

[Picao84]

No way, Renior is 150mm^2 and it comes with 8 zen 2 cores and only 8CUs for slightly more than 1TF. You'd need at least to be bigger than that for 4TF.

Sorry for quoting you for the third time, but apparently over 43 mm2 of those 150 mm2 are CPU caches, including a massive 16MB of Lvl 3 cache. It's been speculated that caches will be greatly axed for console APUs and, as I typed earlier Xbox One X fit 2TFlops extra of compute compared to PS4 Pro in a die that was only overall 35 mm2 larger. In this case I still think it's possible to have a 4TFlop APU on roughly 150 mm2, given that Renoir GPU is at 1,74 TFlops already @ 1,7 GHz.

 

[AzBat]

Well, that was a whole lot of nothing.

Well, I saw the word "balance".

Tommy McClain

 

[PSman1700]

Doom is not a great example, it runs well on just about everything - even the Switch!

That is why it is a great example, games tend to scale and are optimized better then ever before. My 2012 pc has no trouble outperforming the 2013 PS4 even today.

I'm not totally disagreeing about the gap shrinking, but have you guys seen the minimum specs for RDR2? Any idea how much that would cost to build compared to buying a PS4?
Anything on a PC built to PS4 spec that looks as good as God of War or Days Gone (etc)?
Also, this time during last gen games on PCs and consoles were running really badly on the consoles...not so much this gen from what I've seem.

RDR2 on PS4 is running @ 30fps (with drops, sometimes in the low 20's), at low to med settings.
Doom arguably looks (art aside) as impressive as GoW does to me. No idea why days gone is being mentioned. Death Straning and HZD will be nice proving grounds on different pc configurations. Detroit could be a test-case too.

New rumor from reddit:

Another troll knowing specs of both consoles, that won't reveal his/her identity. And that makes it to B3D.

the guy is verified

Wasn't Klee also verified? Very troll-like text.

No he just said like Jason Schreier told before the two consoles are powerful and not far from each other. He can't give numerical data because he is under NDA and he thinks digitalfoundry or NxGamer will have difficulty to find difference this time. This is something precise for someone under NDA and he said too every leak did not tell the full truth.

What is close in power to this troll? 9 vs 12TF could seem close, without numbers one can say whatever he or she wants to, to gain attention.

But not a simultaneous launch and also had the benefit of a huge load of exclusive third party games.
PS5 will lose the old trend where it was the most powerful console at launch for the first time and will have barely the exclusive titles PS2 was enjoying.

True, and yes it is easy to be the most powerfull if your the only one launching at that date

 

[Rangers]

I'm not sure that having Lockhart with 7 TFlops makes much business sense within a two tier system, unless it's sold for really cheap like 199. This would most likely barely cover the BOM, so it would need to be more expensive and with that comes the risk of osbourning Anaconda, negating any advantage they might have on Sony if PS5 is indeed 9.2 TFlops.

Like yes, Anaconda is more powerful but also likely more expensive. If Microsoft launches Lockhart with a price point that is barely cheaper than the PS5, they are changing the value proposition and making Lockhart look more appealing since it's closer to PS5. What's the point of Anaconda then?

If Lockhart was indeed upgraded to 7 TFlops is it because Microsoft started to doubt their 2 tier approach (at some point Lockhart had apparently been cancelled) as PS5 would have been a perfect middle ground? Triggering them to update Lockhart specs, sacrificing Anaconda sales?

Yeah, 7TF with 16 GB GDDR doesn't make any sense to me. The BOM cant be all that much lower for that, and they start encroaching on each others territory, Lockhart and XSX. At that point it becomes a strong case of why bother IMO.

 

[chris1515]

That is why it is a great example, games tend to scale and are optimized better then ever before. My 2012 pc has no trouble outperforming the 2013 PS4 even today.



RDR2 on PS4 is running @ 30fps (with drops, sometimes in the low 20's), at low to med settings.
Doom arguably looks (art aside) as impressive as GoW does to me. No idea why days gone is being mentioned. Death Straning and HZD will be nice proving grounds on different pc configurations. Detroit could be a test-case too.



Another troll knowing specs of both consoles, that won't reveal his/her identity. And that makes it to B3D.



Wasn't Klee also verified? Very troll-like text.



What is close in power to this troll? 9 vs 12TF could seem close, without numbers one can say whatever he or she wants to, to gain attention.



True, and yes it is easy to be the most powerfull if your the only one launching at that date

No Kleegamefan identity was verified but not his information. When you have the tag verified your identity and the fact you currently work in the industry and your information are verified.

 

[RDGoodla]

No he just said like Jason Schreier told before the two consoles are powerful and not far from each other. He can't give numerical data because he is under NDA and he thinks digitalfoundry or NxGamer will have difficulty to find difference this time. This is something precise for someone under NDA and he said too every leak did not tell the full truth.

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.

 

[PSman1700]

No Kleegamefan identity was verified but not his information. When you have the tag verified your identity and the fact you currently work in the industry and you information are verified.

So, any specs verified then?

 

[iroboto]

When you have the tag verified your identity and the fact you currently work in the industry and your information are verified.

When you have the tag verified your identity and the fact you currently work/have worked in the industry. __ it does not imply ___ your information is verified. Mods have no method of verifying your information.

@DSoup set me straight on this one. The last piece, I believe, is what most readers will forget.

 

[MrFox]

When you have the tag verified your identity and the fact you currently work/have worked in the industry. __ it does not imply ___ your information is verified. Mods have no method of verifying your information.

@DSoup set me straight on this one. The last piece, I believe, is what most readers will forget.

Klee was verified to be an ex-journalist, a blogger, he does not work in the industry. His friend was not verified to even exist, let alone the information he claimed to have. He's the reason a lot are questionning why such a low bar was set in his case.

This guy here was verified as actually working in the industry and a VR developper. They verifiy the persons are who they claim to be.

 

[Shifty Geezer]

No he just said like Jason Schreier told before the two consoles are powerful and not far from each other.

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.

 

[chris1515]

When you have the tag verified your identity and the fact you currently work/have worked in the industry. __ it does not imply ___ your information is verified. Mods have no method of verifying your information.

@DSoup set me straight on this one. The last piece, I believe, is what most readers will forget.

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.

And if it is for multiplatform games, buy the cheapest one or where you have more friends, because you will not notice the difference (yes, there are people who are able to perceive a laser pointer on the moon, and others like to play 30cm from the TV trying to find out if the 4K is native or scaled by switching quickly between different video inputs to find out the difference to justify their purchase, but those people are not happy. Fortunately, those people will have it very difficult in the next generation... in multiplatform games).

I think he talks precisely of pixel counter and said it will be very difficult to find the difference next gen.