Which team are you on?

  • Double Team (1997)

  • Team Walnut

  • The A-Team

  • Team "No One Can Stop Mr. Domino"

  • Sports Team

  • "I'm a loner, Dottie. A rebel."

  • Team Margarita


Results are only viewable after voting.
Status
Not open for further replies.

Dust

C H A O S
Member
Oct 25, 2017
36,838
I am starting to fear the delay due to the virus, a lot of business are already having problems due to blocked flow of goods.
 

Deleted member 8784

User requested account closure
Banned
Oct 26, 2017
1,502
www.resetera.com

Next-gen PS5 and next Xbox speculation launch thread |OT8| - The Dark Tower (See Staff Post) OT

I’m glad to see we’ve moved pass thinking the Github leak was confirmation of anything.

It isn't confirmation of anything, and (most) of those who put some stock in the Github leak aren't saying it is. Even the poster you're quoting described it as a probability. There is no "confirmation" of anything until it's direct from the horses mouth - everything else is heresay, whether that's from Github, Digital Foundry, or "insiders".

That doesn't mean people can't base their expectations around these rumours though. I do suspect that you'd be less likely to try and shut down other people's expectations if they closer matched your own.
 

KeRaSh

I left my heart on Atropos
Member
Oct 26, 2017
10,884
IIRC DF verified if the data was legit with some of their sources. Of course they also investigated the data itseof and in result they run with 36 CU @ 2 GHz GPU for PS5 since then. Like we all do they also stated that it isn't a confirmation but a likely outcome based on that data leak.
I thought they verified that the info did leak from AMD from actual tests (instead of someone making up numbers). What they did not confirm was that the hardware that was tested is going to be the final hardware that will end up in the PS5.
 

Deleted member 1589

User requested account closure
Banned
Oct 25, 2017
8,576
Except there is, it's in the GitHub leak, and Digital Foundry has verified the data is legit. Stop selectively ignoring.
Well I'm glad you bought that up, since I've been wondering about people who selectively ignored what was written in that Digital Foundry article.

However, having looked into the situation and independently verified the source, the overwhelming evidence is that the data does indeed originate from AMD - and hasn't been doctored.

This is the only thing they could verify.
 

chris 1515

Member
Oct 27, 2017
7,084
Barcelona Spain
It isn't confirmation of anything, and (most) of those who put some stock in the Github leak aren't saying it is. Even the poster you're quoting described it as a probability. There is no "confirmation" of anything until it's direct from the horses mouth - everything else is heresay, whether that's from Github, Digital Foundry, or "insiders".

That doesn't mean people can't base their expectations around these rumours though. I do suspect that you'd be less likely to try and shut down other people's expectations if they closer matched your own.

By Komachi we know the Oberon B0 from June 2019 github* is far from been the final chip, we had Oberon C0, Oberon D0 and Oberon E0 at least and it seems the chip was in testing mode later than Xbox Series X APU.

It was my Google drive content which was deleted by AMD.

What Matt told is that it is not the final chip and it can goes in two way maybe the chip was downclocked but three majors revision for a simple frequency change seems a bit strange or it was a much bigger change.

*May 2019 Oberon test
 
Last edited:

Adum

Member
May 30, 2019
965
No, they are both the same. The issue that I am trying to get at is the CORES part. OMG..... You guys are killing me here.
What I'm getting from your last few posts is this -

"Yes, of course both next gen consoles will have hardware based raytracing. Except the XSeX will have proper hardware in the gpu dedicated to raytracing while the PS5 won't have any such hardware features. No no, but they're both the same. That's what I'm saying see? Also, who's Mark Cerny and why does he keep lying?"
 

Brohan

The Fallen
Oct 26, 2017
2,547
Netherlands
What I'm getting from your last few posts is this -

"Yes, of course both next gen consoles will have hardware based raytracing. Except the XSeX will have proper hardware in the gpu dedicated to raytracing while the PS5 won't have any such hardware features. No no, but they're both the same. That's what I'm saying see? Also, who's Mark Cerny and why does he keep lying?"

It's so damned transparent that i'm amazed that he thinks that people wont see what he's actually saying.
 

P40L0

Member
Jun 12, 2018
8,550
Italy
I've only said the 2GHz configuration is feasible, I did not imply it will be the final form.
340


I'm out of the speculation for a very long time.

Can someone really TL;DR what we're currently at?

XSX slightly more powerful thant PS5 (something like 12TF vs 10TF) but mostly identical and both at 500$ at launch. Right?
 

beta

Member
Dec 31, 2019
176
No, that's all incorrect. I don't even know who that is. The lead coder for The coalition Studio said "dedicated hardware CORES", which is NOT the same as hardware added to the GPU. It's not supposed to be revealed yet.

Microsoft said hardware raytracing in their video, but they didn't go into more details, the coder from "The Coalition" did because he thought Microsoft already had mentioned it.

I never mentioned Mark Cerny's name on this. It has nothing to do with him. He said it's built into the GPU, that's all he has said.
I also never said anything about software as far as the PS5.

We know that the PS5 has hardware for raytracing put in the GPU.

Software means it's purely a program using a regular CPU. Raytracing code has been added to the GPU hardware, thus it's using hardware with driver support.
Got it?

The fact that I have to point this out to not offend anyone is just absurd.

Please, before you accuse me of something, at least ask me where I got the information first.

GPUs run software, its what shaders are. Unless there is dedicated hardware to run instructions on the GPU then its software and not in anyway hardware. It is clear that the raytracing is hardware accelerated using fixed function dedicated hardware in both the PS5 and XSX and there's no evidence otherwise, other than people fundamentally misunderstanding what the words 'software' and 'hardware' mean.
 

chris 1515

Member
Oct 27, 2017
7,084
Barcelona Spain

One revision would have been more than enough for doing it. ;) It is how it was gone from 1 GHz to 1.8 Ghz and finally 2 Ghz. No need to do three steppings for a simple downclock or upclock. I just say than github leak is not the final chip and the end of the story. The last people saying than the two consoles are very close is Jason Schreier on era the 4th and 9th december 2019 and I trust him.

EDIT: Probably XSX comes a little bit on top.
 
Last edited:

III-V

Member
Oct 25, 2017
18,845
The information was from GDC, there was nothing specified either way. I don't see a problem with that.

If the information matches the same exact 9,216 TF number given from GitHub then I think it has to be looked at with some seriousness.

Look, you can have your 12 TF PS5 retail dreams. I am not out to destroy anyone's dreams.

I am just saying that I personally don't believe that. That's it. I believe in the GitHub rumors because they match the exact number that was found at GDC 2019. If you match an exact number such as 9,216 TF from GDC 2019, I am going to take note of it and that same number matches the GitHub leak.

That data point was not from a dev at GDC. Remember that Gonzalo was found way before GitHub. Gonzalo was 20dCU at 1.8 GHz equals 9.216 TF. Which also matches 18dCU at 2.0 GHz equals 9.216 TF. It's crazy, I know, but they had been testing this Oberon/Gonzalo thing from as early as late 2018. So Josh cobbled together a rumor around GDC time based on this info. If you really TRUST the FACTS from twitter Josh, here is what he recently said about the Series X:






So is that where we are at now? Seems like a reputable source, amirite!


The actual quote to bloomberg was "Usually cooling is less than a buck for us but since we designed this console for 2018 we had to jack up the clocks so high that cooling actually costs $40 and the BOM is $450. This is totally normal. We cool it with baby penguins."
LMAO
Just accept ps5 is 36cu, then be pleasantly surprised if it's higher. Thou highly likely it ain't..
Thanks, user hwraytracer! Anything else you know?
Yes! Back to 8 TF!
 
Last edited:

Brohan

The Fallen
Oct 26, 2017
2,547
Netherlands
340


I'm out of the speculation for a very long time.

Can someone really TL;DR what we're currently at?

XSX slightly more powerful thant PS5 (something like 12TF vs 10TF) but mostly identical and both at 500$ at launch. Right?

We are at XsX @12tf and PS5 @9-14tf. Both with hardware raytracing but the XsX is hardware based while the PS5 is hardware accelerated. Something along those lines.
 

Pheonix

Banned
Dec 14, 2018
5,990
St Kitts
Then please make a case why gen2 is called native mode? To me native sounds like using all available resources.
I think I did.... native mode would be either... the PS5 running PS4 games in their naive mode; i.e Native GPU CU number and Native clocks for the games that for whatever reason doesn't boost well or doesn't have a PS4pro patch or version. And running games with a dedicated PS4pro mode in their "native mode" without giving it a PS5 boost.

The point here is, we simply cannot know (regardless of the literature) if that is all the available resources the system has when what that leak showed was clearly BC testing.

Another point is this... if for any reason the PS5 is 9ish TF compared to the XSX 12ish TF.... I doubt a single person out there would be saying they are close in power much less some even saying he PS5 has the edge. There is just too much conflict between the two threads of information we have.
 

MrKlaw

Member
Oct 25, 2017
34,907
That data point was not from a dev at GDC. Remember that Gonzalo was found way before GitHub. Gonzalo was 20dCU at 1.8 GHz equals 9.216. Which also matches 18dCU at 2.0 equals 9.216. It's crazy, I know, but they had been testing this Oberon/Gonzalo thing from as early as late 2018. So Josh cobbler together a rumor around GDC time based on this info. If you really TRUST the FACTS from twitter Josh, here is what he recently said about the Series X:






So is that where we are at now? Seems like a reputable source, amirite!


LMAO
Thanks, user hwraytracer! Anything else you know?

Yes! Back to 8 TF!


this could be fun. Xsx 12TF Vega vs 9TF Navi?
 

anexanhume

Member
Oct 25, 2017
12,957
Maryland
I think ray tracing in the PS5 is hardware encouraged.

"Go get those rays, son!"

"Aw man, you're tracing the hell out of those paths dog!"

Cerny seems all about positive reinforcement.
 

Deleted member 1589

User requested account closure
Banned
Oct 25, 2017
8,576
I swear its like people don't posts that aren't their own.
Yep.

The discourse is mostly 80% pointing out the problems with the github leak and what spawned out of it.

Says a lot that they would rather talk about this than speculate about the other console, and then complain that the quality of discourse has dropped.
 

msia2k75

Member
Nov 1, 2017
601
That data point was not from a dev at GDC. Remember that Gonzalo was found way before GitHub. Gonzalo was 20dCU at 1.8 GHz equals 9.216. Which also matches 18dCU at 2.0 equals 9.216. It's crazy, I know, but they had been testing this Oberon/Gonzalo thing from as early as late 2018. So Josh cobbler together a rumor around GDC time based on this info. If you really TRUST the FACTS from twitter Josh, here is what he recently said about the Series X:






So is that where we are at now? Seems like a reputable source, amirite!


LMAO
Thanks, user hwraytracer! Anything else you know?

Yes! Back to 8 TF!


Looks like Tidux level of "sourcing"....
 

eso76

Prophet of Truth
Member
Dec 8, 2017
8,679
We're 14 teraflopping now ?
I reckon if Sony doesn't make an official statement by April we might approach 16tf realm.
 

Brohan

The Fallen
Oct 26, 2017
2,547
Netherlands
I think ray tracing in the PS5 is hardware encouraged.

"Go get those rays, son!"

"Aw man, you're tracing the hell out of those paths dog!"

Cerny seems all about positive reinforcement.
Hardware Suggested Ray Tracing:

"Now you're feeling like tracing some rays"

"Slowly you drift into a contact hardening penumbra"

No no, it's actually "Hardware Passed Raytracing"

"Why are you passing me by Ray? I need you to go through me, to use me. I traced you all the way here! You Can't just ignore me again!!"
 

Deleted member 31104

User requested account closure
Banned
Nov 5, 2017
2,572
I'm not writing off the github leak at all, I just think fast and narrow is such a weird choice given the little we know about AMD's new line up than if it's correct I'd really be interested in finding out the rationale.

Perhaps it's a simple as: better yield even after allowing for clocking tolerance = Acceptable performance for lower total cost even after cooling considerations.

But it seems a bit of a limiting route.
 

Pheonix

Banned
Dec 14, 2018
5,990
St Kitts
One revision would have been more than enough for doing it. ;) It is how it was gone from 1 GHz to 1.8 Ghz and finally 2 Ghz. No need to do three steppings for a simple downclock or upclock. I just say than github leak is not the final chip and the end of the story. The last people saying than the two consoles are very close is Jason Schreier on era the 4th and 9th december 2019 and I trust him.

EDIT: Probably XSX comes a little bit on top.
Exactly. It wouldn't surprise me if the XSX comes out on to, but no way in hell is that going to be to the tune of 8-9TF vs 12TF. Its even worse when you consider that to get to that 8-9TF we are also to believe that sony is taking a small chip and pushing it as hard as they possibly can with expensive cooling and all that.

I'm not writing off the github leak at all, I just think fast and narrow is such a weird choice given the little we know about AMD's new line up than if it's correct I'd really be interested in finding out the rationale.

Perhaps it's a simple as: better yield even after allowing for clocking tolerance = Acceptable performance for lower total cost even after cooling considerations.

But it seems a bit of a limiting route.
I can see fast and narrow being a thing. The issue here is "how narrow" Narrow to the tune of a 36CU GPU makes absolutely zero sense to me. Especially if MS is getting 56-60CU!!!! Like come on, somehow the best sony could do was go for 36CU after all the years of designing the PS5 closely with AMD?

40CU, 44CU...48CU, those all seem more likely than just going with 36CU for reasons.
 

anexanhume

Member
Oct 25, 2017
12,957
Maryland
┏┓
┃┃╱╲ in
┃╱╱╲╲ this
╱╱╭╮╲╲house
▔▏┗┛▕▔ we
╱▔▔▔▔▔▔▔▔▔▔╲
ONLY RASTERIZE
╱╱┏┳┓╭╮┏┳┓ ╲╲
▔▏┗┻┛┃┃┗┻┛▕▔
 

KeRaSh

I left my heart on Atropos
Member
Oct 26, 2017
10,884
|OT12| It's Groundhog Day (again)!

Seriously, we're talking about the GitHub leak (again) and have posters doubting the RT implementation of the PS5 (again)...
 

Deleted member 8784

User requested account closure
Banned
Oct 26, 2017
1,502
Like come on, somehow the best sony could do was go for 36CU after all the years of designing the PS5 closely with AMD?

But it's not just about "the best they can do". If the best they can do was about raw specs and nothing else, they'd all be releasing multi thousand dollar god machines.

It's about the best performance you can get, for the price point your aiming for, within the form factor you want, and any other design goals you might have.

and have posters doubting the RT implementation of the PS5 (again)...

Are people genuinely doing this?
 

KeRaSh

I left my heart on Atropos
Member
Oct 26, 2017
10,884
But it's not just about "the best they can do". If the best they can do was about raw specs and nothing else, they'd all be releasing multi thousand dollar god machines.

It's about the best performance you can get, for the price point your aiming for, within the form factor you want, and any other design goals you might have.



Are people genuinely doing this?
Not sure if they're doing it "genuinely" but they're doing it.
Just look at recent posts from TimStone .
 

Brohan

The Fallen
Oct 26, 2017
2,547
Netherlands
|OT12| It's Groundhog Day (again)!

Seriously, we're talking about the GitHub leak (again) and have posters doubting the RT implementation of the PS5 (again)...

Even though I hate the fact that we keep going in circles with the Github stuff at least I understand it. But there is 0 excuse for people to still question the Raytracing implementation of the PS5.
 
Status
Not open for further replies.