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

[Deleted member 11852]

Obviously he's not expecting robots to be consumer products so could it be stealth Horizon Zero Dawn II hype?

Is Barlog's tweet a coincidence, is he just goimg to CES for the hell of it. Could it be more?

Not sure CES is the place for it, but ...

Doing any form pf PS5 info dump at CES would be unexpected. Like a lot of people I've been expecting something in February because that's what Sony did for PS4. But looking at the facts, Sony have completely changed their comms approach for PS5 so why would it the same? The Wired interviews were just odd but Sony will certainly want to pre-empt any leaks from GDC in March so maybe they will do it as CES.

CES is a logical place for it, the press are all there and let's be honest. A new console is only a big event in the gaming world. Nobody else really cares.

 

[MrFox]

Thank you Richard.
If you’re like me and had issues catching up on a lot of small moving parts; Richard goes through the entirety of it and the recent performance leaks in a single video. Really helpful in clarifying how things became known so quickly and making sense of the results, Without having to read pages on end of concern trolling.

Always great rational analysis. And when Richard starts talking about leaks there's a fair chance they are accurate. It's really cutting through the noise, maybe this shouldn't be in the baseless thread

 

[AbsoluteBeginner]

So we are going from :

"9.2TF???Sony is skimping! This is unnacaptable"

...to

"2.0GHz?? Are they mad? 36CU Navi at that clock will burn your console, should have gone nice wide and slow"

1. There are rules to chip design. Navi CUs are much larger then GCN, therefore they are not straight comparison vs PS4Pro
2. 2.0GHZ Navi, that was product of many years of work and big budget, will almost certainly be better chip then first Navi's that AMD released 1.5yrs before this one is released
3. I assume console manufacturers will use N7P node, as this does not require chip redesign but will give them round about ~10% better TDP at same clocks
4. Wide and slow vs narrower and high clocks is not clear cut. Bigger die sizes are getting more and more expensive. Chips with high transistor counts cost much more then the ones back in 2004, there is a reason why Nvidia and AMD are pushing frequency. Its BETTER.

Really, up until 12TF mini PC premium SKU that MS shown, everyone would have LOVED ~13TF Vega with RT and VRS in next gen console so I don't really understand what is the fuss? Only way Sony will ever match MS is if they go with 2 SKU strat or say "Fuck lower end, Lockhart can eat bottom end, we will only release premium SKU" which, IMO, would be suicidal.

So yea, bottom line. 36CU at 2.0GHz is achievable in 2020, especially on N7P node. It will still result in very advanced console and 399$ is NOT guaranteed. Fact that they went narrow and fast does not mean going wide and slow, like 56CUs at 1500MHz would result in better performance. Only if wide and slow means, 56CUs at 1700Mhz+, which I imagine was never a plan for Sony as this is very much as premium as it gets (and, will have to be unorthodox large compared to every other console released ever).

 

[mpg1]

If every chip has to operate at 2.0GHz how good are the yields going to be...?

 

[PSman1700]

If it's Godfall level then cancel Project PS5 altogether!

PS5 9TF nothing wrong with that. If hellblade 2 is optimized for that 12TF then maybe PS5 won't see those graphics (for a while) but i doubt it is optimized for full 12TF.

Thank you Richard.
If you’re like me and had issues catching up on a lot of small moving parts; Richard goes through the entirety of it and the recent performance leaks in a single video. Really helpful in clarifying how things became known so quickly and making sense of the results, Without having to read pages on end of concern trolling.

And when Richard starts talking about leaks there's a fair chance they are accurate.

Almost certain it is the PS5 DF says, AMD pulling it. Must be the real thing. We are also very close to final hardware, it must be final now one would think, aside from some clocks, as production starts in some months. So 9TF it is then, rdna1/2 hybrid. It is what i expected, ballpark 5700XT, abit lower then AMD's medium card from 2019.
GDDR6 it is then, not the HBM some thought. 3d Audi possibly AMD's TrueAudio Next.

Thankyou DF On a note, 'consoles have delivered way beyond equal radeon parts', maybe some, but not alot, for most multitplat games you get about the same performance, esp when dynamic res is in play, and same settings.

Really, up until 12TF mini PC premium SKU that MS shown, everyone would have LOVED ~13TF Vega with RT and VRS in next gen console so I don't really understand what is the fuss? Only way Sony will ever match MS is if they go with 2 SKU strat or say "Fuck lower end, Lockhart can eat bottom end, we will only release premium SKU" which, IMO, would be suicidal.

MS can do it, since they have a 2 SKU plan, also a very odd pc design for a console which the japanese sony not wanted to do. DF's take on this is very realistic and what i expected. Not sure what MS wants to do either, chasing hardcore pc gamers not really, as that 12TF will be rather low compared to 2020/2021 big navi2 GPU's.

 

[McHuj]

Doing any form pf PS5 info dump at CES would be unexpected.

It would be, unless it's related to HDMI 2.1 features that the PS5 will support. I think Sony could focus on those features of the PS5 (assuming it does support them) and couple that in conjunction with new TV announcements that support those as well.

I expect Sony to announce TV's with HDMI 2.1 next week and it wouldn't be that much of a leap for them to say PS5 supports these features as well "and if you want the best TV for PS5 gaming, it's a Sony Bravia" (not a Samsung/LG). For as big of a gaming company as Sony is, I've always found their TV's to be lacking (or behind) for gaming (input latency).

 

[JoeJ]

Really, up until 12TF mini PC premium SKU that MS shown, everyone would have LOVED ~13TF Vega with RT and VRS in next gen console so I don't really understand what is the fuss?

Well, i can answer this from my perspective, being someone who depends on the TF number.

1 Navi TF > 1 GCN TF? Bullshit! 1==1, period. AFAIK Navi has improved on rasterization, geometry throughput... things like that. But compute does not benefit from this directly.
PS more pixels than XSX? Great for those who think pushing pixels is the bottleneck. I certainly do not. And again no direct benefit for me from this.

Said this only to illustrate that some complaints are justified if put in proper context. If the very most complaints are not, then just don't bother?
I'm fine with all this - expected those movements. I'm also fine with 4TF Lockhart, which is really low and i did not expect that. It's a challenge.

So if somebody is disappointed from 9TF, go buy Anaconda and done.
If we get a 4,9,12 TF landscape, what could be more interesting than that?
I'ts cool having differences. Otherwise one platform would just destroy the other, and nobody would benefit from this.

 

[snc]

So we are going from :
Really, up until 12TF mini PC premium SKU that MS shown, everyone would have LOVED ~13TF Vega with RT and VRS in next gen console so I don't really understand what is the fuss?

Yep, but people like big numbers

 

[BRiT]

The numbers are fine ... in isolation. Some camps hate their numbers only when compared to the other camps. If only those rooted deeply in a camp could be wiped of any knowledge of the other camp, online discussions would be far better for the non-campers.

 

[Shortbread]

Quite odd to randomly post this...

https://twitter.com/i/web/status/1211719104681463808

 

[AbsoluteBeginner]

Oberon

CU 36
GHZ 2.0
TF 9.2
Bus width 256bit (probably 544GB/s - 576GB/s)
Theoretical Pixel Fill Rate 64ROPs * 2 = 128 GPix/s
Theoretical Texture Fill Rate 36 * 4 * 2 = 288 GT/s

Arden

CU 56
GHZ 1.7
TF 12TF
Bus width 320bit (probably 560GB/s - 640GB/s)
Theoretical Pixel Fill Rate 64ROPs * 2 = 108 GPix/s
Theoretical Texture Fill Rate 56 * 4 * 1.7 = 380.800 GT/s

Quite odd to randomly post this...

https://twitter.com/i/web/status/1211719104681463808

They post these game quotes every few days I think.

 

[Rootax]

Have we confirmation that Sony is using rdna >1 btw ? What not a normal rnda arch+customs modifications but not linked to rdna2 ?

 

[PSman1700]

They post these game quotes every few days I think.

True, it has nothing to do with PS5 specs.

 

[MrFox]

Oberon

CU 36
GHZ 2.0
TF 9.2
Bus width 256bit (probably 544GB/s - 576GB/s)
Theoretical Pixel Fill Rate 64ROPs * 2 = 128 GPix/s
Theoretical Texture Fill Rate 36 * 4 * 2 = 288 GT/s

Arden

CU 56
GHZ 1.7
TF 12TF
Bus width 320bit (probably 560GB/s - 640GB/s)
Theoretical Pixel Fill Rate 64ROPs * 2 = 108 GPix/s
Theoretical Texture Fill Rate 56 * 4 * 1.7 = 380.800 GT/s


They post these game quotes every few days I think.

Where is the 544-576GB/s memory rumor coming from? Wasn't it 448 in that leak?

Seems very unlikely they would use the top bin, it would be wreckless while everything else in the ps5 rumor is precisely balanced to reach a good price/performance.

512 is quite possible though.

 

[AbsoluteBeginner]

Where is the 544-576GB/s memory rumor coming from? Wasn't it 448 in that leak?

Seems very unlikely they would use the top bin, it would be wreckless while everything else in the ps5 rumor is precisely balanced to reach a good price/performance.

512 is quite possible though.

There were old theoretical values (448 GB/s) in repo, but as DF noted, newer Oberon had 530GB/s achieved. I think 448GB/s was left over template from early Ariel (14Gbps).

So, either overclocked 16Gbps or down clocked 18Gbps.

I am thing they will go for 18Gbps chips, but under clocked.

 

[AlphaWolf]

Have we confirmation that Sony is using rdna >1 btw ? What not a normal rnda arch+customs modifications but not linked to rdna2 ?

You can claim that any customizations to the base of 1 automatically moves you to >1. However it's possible that doesn't necessarily make it closer to 2.

 

[turkey]

Have we confirmation that Sony is using rdna >1 btw ? What not a normal rnda arch+customs modifications but not linked to rdna2 ?

If we take the leak at face value it has hardware back comparability modes which pretty much guarantee its custom to some degree. If that is also with IP from after RDNA1 or just old GCN stuff we do not know, but it's almost certainly not a vanilla pc part.

Cerny was quite forward looking with compute on base PS4, then checkerboard with pro. I think it's likely they will tout a new efficiency tweak or two.
I would think psvr will have its own needs met, a new headset must be in incubation.

Edit: ninja'd

 

[Deleted member 11852]

Well, i can answer this from my perspective, being someone who depends on the TF number.

1 Navi TF > 1 GCN TF? Bullshit! 1==1, period. AFAIK Navi has improved on rasterization, geometry throughput... things like that. But compute does not benefit from this directly.

OMG. You need to email AMD right now. Their RNDA Whitepaper is completely wrong.

One clock cycle by crystal frequency = one clock cycle. The amount of work any particular chip can achieve in one clock cycle = variable. I'm old so I remember when no processor could perform a single operation in a one clock cycle unless it was NOP (No OPeration).

 

[RDGoodla]

Arden has 56CUs active. If there are also some disabled CUs for better yield (since perfect die is very expensive) then total CUs will reach 60~64.
Chip size is 350mm2 then it is very likely in 7nm+ for 60~64 CUs.

How could 7nm/N7P be possible?

 

[Deleted member 11852]

You can claim that any customizations to the base of 1 automatically moves you to >1. However it's possible that doesn't necessarily make it closer to 2.

Are you Google's accountants?