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

Status
Not open for further replies.
Scott_Arm said:
Why the assumption that XseriesX has the same number of rops? Looks like RDNA has 4 RBs per shader array. 40 CUs aligns to 4 shader arrays and 60 CUs aligns to 6. It could be a different configuration, but I’m just wondering if there are any numbers to suggest that it is.

edit: confusing codenames
Click to expand...

Would be moot if they don't have sufficient external bandwidth. The simplest would be to just add more DCUs/WGPs per array (also from a die size standpoint in terms of overhead blocks needed for extra arrays).

Edit:
I'd be interested in Navi 10 bandwidth scaling since the 5700 series can be anywhere from 6.7 - 10.1TF on 448GB/s.
 
Last edited:
Looking at the RDNA whitepaper, True Audio Next sounds a lot like what Sony talked about for their 3D audio.

First, the GPU can partition the shader arrays and reserve select compute units solely for audio tasks, ensuring low latency and predictable execution. Second, True Audio relies on two ACE-managed real-time task queues. One queue is used for computing audio ray-tracing using the open-source Radeon Rays library, the other is used to calculate convolutional effects for positional audio and full surround sound rendering.
Click to expand...

I wonder if RDNA2, or whatever it will be called, has further investment in this, or if Sony has customized it further.
 
DSoup said:
It was on Github until this morning when, presumably AMD, had it pulled.
Click to expand...

Someone has broken some rules somewhere :)

AbsoluteBeginner said:
I'm struggling listening to "Klee said"
Click to expand...

But Klee said what they wanted to hear, you must understand that. Everything else is just wrong. Real B3D stuff, this Klee.

AbsoluteBeginner said:
He hasnt said anything of note, just please can we please referrence Klee here? There was nothing technical from here posted...EVER.
Click to expand...

He has also been wrong.

@Scott_Arm
So it is a hardware block on the GPU afterall, RDNA2 GPU's release before PS5 so we probably going to see some tests on it.
 
PSman1700 said:
...
@Scott_Arm
So it is a hardware block on the GPU afterall, RDNA2 GPU's release before PS5 so we probably going to see some tests on it.
Click to expand...

I wouldn't call it a hardware block. It's just using the standard GPU parts to do the work via some driver, software SDK features. That's assuming Sony is just using the standard RDNA features provided by AMD.
 
Scott_Arm said:
Looking at the RDNA whitepaper, True Audio Next sounds a lot like what Sony talked about for their 3D audio.



I wonder if RDNA2, or whatever it will be called, has further investment in this, or if Sony has customized it further.
Click to expand...

What’s different from the GCN version?
 
Silent_Buddha said:
Additionally, if Phil Spencer is to be believed he at least has some type of prototype Xbox Series X device (or devkit) at his home that he is playing games on. Whether it's able to play games other than XBO/X360/Xbox games could be questionable, but I'd bet it's able to run anything that's being developed for Xbox Series X currently.
Click to expand...

He stated it's his primary console now. That tell me it's backward compatible with anything the Xbox One can play.

https://twitter.com/x/status/1202354057047728128

AbsoluteBeginner said:
Kinda funny how Arden and Sparkman are both towns in state of Arkansas. Heh...the coincidence.
Click to expand...

That is funny. I've only been living in Northwest Arkansas for the past 11 years. So it's no wonder I never heard of those towns in the Southern Arkansas.

BTW, maybe the zip codes have something to do with the projects?

Arden, Arkansas 71822
Sparkman, Arkansas 71763

BRiT said:
Just dont bring "Arkancide" into the thread.
Click to expand...

I had to Google that as I don't keep up with the news. LOL

Tommy McClain
 
AbsoluteBeginner said:
Most tabs are "placeholder". As I said, I think Sparkman is copy from Arden, but in additional tab Sparkman didn't have GEN1, and this matches Komachi saying "Sparkman" is "less then Oberon/Arden".

Problem is people looking at 9.2 v 12 and saying OMG XSX is 30% faster. Its not. Period. Not as per leaked specs.

1. More CUs (56 vs 36) at lower clocks (1.7 vs 2.0) with same amount of ROPSs will result in worse pixel fillrate (25% advantage for PS5)
2. More CUs at 560 (or even 640GB/s) bandwidth will result in smaller BW per TF then 36 CUs at 512/576GB/s

From Sony's perspective, and knowing their BC method, they were pretty much "pushed" towards 36CUs at very high clocks. This will result in straightforward BC, smaller chip and less TF (but more pixel fillrate).

If they went with 56CUs, they would have to lower clocks (maybe even resulting in ~11TF), resulting in 35% lower pixel fillrate (!!), bigger chip and lower yields for ~2TF more. Pushing for 12-13TF would result in VERY expensive premium console going against competitor pushing one at 299$.

IMO, in worst possible case (12TF vs 9.2TF, 640GB/s vs 576GB/s) difference will not be bigger then 20%. And that is CLOSE.
Click to expand...

Yo! Nice calculation. One more thing. I'm pretty sure bunch of you guys heard for Hmqgg ( member on ERA ), one of Xbox insiders. He claimed that regarding performance Scarlet chip is equal to Dante ( devkit ) chip. And also number was this one : Custom AMD Navi @ 1475MHz 64cu - which is also 12 TFLOPS ( maybe RDNA, maybe not, i dunno. Can you calculate pixel fillrate again with 1475 Mhz versus 2.0 in PS5? Thanks!
 
dobwal said:
What’s different from the GCN version?
Click to expand...

gcn version was a dedicated dsp and it didn’t have the 3D audio capability. True Audio Next is strictly gpgpu and has no backwards compatibility with gcn True Audio. That’s interesting with respect to BC on PS5. I wonder if they’ll include the dsp just to make it easier. I think it’s pretty small.
 
scarythings said:
Yo! Nice calculation. One more thing. I'm pretty sure bunch of you guys heard for Hmqgg ( member on ERA ), one of Xbox insiders. He claimed that regarding performance Scarlet chip is equal to Dante ( devkit ) chip. And also number was this one : Custom AMD Navi @ 1475MHz 64cu - which is also 12 TFLOPS ( maybe RDNA, maybe not, i dunno. Can you calculate pixel fillrate again with 1475 Mhz versus 2.0 in PS5? Thanks!
Click to expand...
Its 94,40 GP/s

But its not 64CU chip clocked at 1475MHZ.

If that were the case I would say, even though its missing 2.8TF, PS5 hypotetical chip would be a better way to go. I would say, for 64CU part clocked at 1475MHZ,t it is waste of silicon (considering Navi clock sweet spot)
 
AbsoluteBeginner said:
Its 94,40 GP/s

But its not 64CU chip clocked at 1475MHZ.

If that were the case I would say, even though its missing 2.8TF, PS5 hypotetical chip would be a better way to go. I would say, for 64CU part clocked at 1475MHZ,t it is waste of silicon (considering Navi clock sweet spot)
Click to expand...

That's for XSX, right? How much pixel fill rate would be then in PS5 if TF are 9.2?
 
scarythings said:
Yo! Nice calculation. One more thing. I'm pretty sure bunch of you guys heard for Hmqgg ( member on ERA ), one of Xbox insiders. He claimed that regarding performance Scarlet chip is equal to Dante ( devkit ) chip. And also number was this one : Custom AMD Navi @ 1475MHz 64cu - which is also 12 TFLOPS ( maybe RDNA, maybe not, i dunno. Can you calculate pixel fillrate again with 1475 Mhz versus 2.0 in PS5? Thanks!
Click to expand...

That's made up.

Arden is 56CUs at ~1700mhz
 
scarythings said:
That's for XSX, right? How much pixel fill rate would be then in PS5 if TF are 9.2?
Click to expand...
2000 x 64 ROPs - 128 GP/s

This is why you would rather push the clocks and get most out of your chip, then go wide and slow. This is why Navi clocks higher then any AMD chip, because console makers want as much freq as possible. Makes for smaller chip, and more bang for a buck.

PS5 at 36CU 2.0GHZ and Zen2 with 1/4 of the cache would be, IMO, beautiful and elegant chip.
 
Proelite said:
Made up by some else. Hmqgg never confirmed it.
Click to expand...

Somehow, i've google it and found this one. LOL

JDAyNqD.jpg
 
Status
Not open for further replies.

Similar threads

A
  • Locked
Baseless Next Generation Rumors with no Technical Merits [pre E3 2019] *spawn*
55 56 57
Replies
1K
Views
124K
Globalisateur
Globalisateur
BRiT
The Annual E3 Microsoft press conference thread, 2019 edition
8 9 10
Replies
191
Views
25K
eastmen
E
D
  • Locked
Next Generation Hardware Speculation with a Technical Spin [post E3 2019, pre GDC 2020] [XBSX, PS5]
178 179 180
Replies
4K
Views
535K
BRiT
BRiT
Back
Top