• Hey, guest user. Hope you're enjoying NeoGAF! Have you considered registering for an account? Come join us and add your take to the daily discourse.

Next-Gen PS5 & XSX |OT| Console tEch threaD

Status
Not open for further replies.

JLMC469

Banned
My gut feeling is that both will be extremely similar in specs and features (basically the same console made by different companies :messenger_grinning_sweat:).
 

jonnyp

Member
typical fanboys projecting here...

I really don't give a shit, and neither should fanboys like you.
historically speaking the most successful systems and systems people are most fond of each generation have, more often than not, been the least powerful systems of their generation...
NES, PS2, GameBoy, DS, 3DS, Wii, the Switch is on its way too... and even going back to the atari 2600 that was true



because a 5TF RDNA console with 12GB of GDDR6 and a Zen2 cpu would be vastly more capable than an Xbox One X. and if that system is targeting 1080p or 1440p it would be more than enough power to play next gen games with reduced resolution alone.

I really disagree on your use of the word 'vastly' in this regard. It will not be worth upgrading from a One X to that IMHO. But by all means, we will agree to disagree.
 

DeepEnigma

Gold Member
You think 2.0 is feasible for both of these consoles?

My gut tells me that 2.0 is testing what the GPUs can handle with overclocking (as they refine the nodes), etc., not just for consoles, but AMDs RT solution across their next products.

I can see it settling around 1800 for consoles at most. But would be pleasantly surprised if more.
 
Last edited:

R600

Banned
Oberon and Arden are within ~20%.

While Arden has more TF, depending on chip speed they choose (14/16 or 18Gbps) it can be anywhere from 30% advantage Arden to 3% advantage Oberon (if Arden is 14Gbps and Oberon is 18Gbps - because Arden is 320bit bus and Oberon is 256bit bus).

On the other hand, if they have same amount of ROPs Oberon will have advantage in pixel fillrate and geometry processing. All in all ~10-20%
 

xool

Member
Oh cmon, Oberon has BC modes for PS4 (patented by Sony).
jsqjpEp.jpg

You right.

Oberon is PS5 99.9% and that means PS5 has :
* 16GB GDDR6 14Gbps RAM ("for games") working in 16bit dual channel mode - total 448GB/s bandwidth
* 36 CU GPU part, 2GHz clock = 9.2 TF theoretical peak
* 8 core/16 thread Zen2 CPU, unknown clocks or cache
* Some sort of SSD solution

It's basically all there .. we know the specs.

imo the old 12.8 TF leaks were pre-Navi dev kits targeting same performance on GCN as Navi at ~10TF

..

I dunno ... seems like everything is explained, a few specs missing , Xbox still uncertain , if you just accept a avalanche of info from multiple sources


OR .. this is all a huge trol/dis-info campaign .. with multiple fake leakers set up to look legit .. that's tin foil hat territory.

Still waiting on Xbox, but we've been told performance is neck and neck by multiple sources (Klee, Reiner, etc)

JUST ACCEPT
 
Last edited:

Mass Shift

Member
Klee in his last message said he will be back home on 29th . I don’t know why people expect him to post before that .we have around 28 hours to 48 hours before he posts haha.

it would be funny if Klee turns out to be true. Whole internet vs one single man .
Things he confirmed before anyone else :

vrs for both
Hardware raytracing for both
12 Navi TF for xsx
Reveal at TGA for XSX

things left to be proven :
Both boxes are double digit Navi TF
ps5 better than xsx by 10%
Ps meeting is in February

Things he was not aware about :

lockhart

He might be unaware of more than that if the engineering test samples that were revealed turn out to be kosher. And there's very little standing in front of their validity.

I think he's honest, but he might have outdated specs and that would put him and his source further out of the loop. Because it's just not as simple as "I stand by what I said" when you have actual test results being posted.

He would have to counter that with something hard or agree that his info could certainly be challenged by others in that forum. And he wouldn't be able to whine about that either.
 
Some strange stuff happened at Sony PlayStation this last year or so.

Rumours of next gen supposed to be this year.
Oddly skipped E3
Shaun Layden mysteriously leaves the company

🤔

Almost forgot about these bits. The Layden one is probably the most puzzling still; just up and left, not even much of an official parting word IIRC. Either he had some emergency family stuff to tend to, emergency health situation, or there was some type of disagreement between them and superiors over...something.

Which could be anything, but given his position I would assume it'd of been about PS. I try being an optimist in those kind of situations but I can't picture him leaving for a reason beyond these three (if he got offered a better job at another company news would've came out about it), and out of those three it'd most likely had been PS-related.

Unless he left because of some crazy epiphany and he decided to travel to a Buddhist temple in pursuit of monkhood xD
 
because a 5TF RDNA console with 12GB of GDDR6 and a Zen2 cpu would be vastly more capable than an Xbox One X. and if that system is targeting 1080p or 1440p it would be more than enough power to play next gen games with reduced resolution alone.

You're already banned from the thread (for obvious reasons) but I've seen this thrown around way too many times. Possible lockhart is much more capable & a generational difference than One X, but at the same time will retail $200 cheaper, despite the One X still retailing for $499 in 2019.

There's no way based on the alleged leak specs of Lockhart, will it be less than $399. It's not as if Phil Spencer came along right after the One X. If MS wanted to be aggressive with their pricing, it would have happened two years ago.
 

bitbydeath

Gold Member
You right.

Oberon is PS5 99.9% and that means PS5 has :
* 16GB GDDR6 14Gbps RAM ("for games") working in 16bit dual channel mode - total 448GB/s bandwidth
* 36 CU GPU part, 2GHz clock = 9.2 TF theoretical peak
* 8 core/16 thread Zen2 CPU, unknown clocks or cache
* Some sort of SSD solution

It's basically all there .. we know the specs.

imo the old 12.8 TF leaks were pre-Navi dev kits targeting same performance on GCN as Navi at ~10TF

..

I dunno ... seems like everything is explained, a few specs missing , Xbox still uncertain , if you just accept a avalanche of info from multiple sources


OR .. this is all a huge trol/dis-info campaign .. with multiple fake leakers set up to look legit .. that's tin foil hat territory.

Still waiting on Xbox, but we've been told performance is neck and neck by multiple sources (Klee, Reiner, etc)

JUST ACCEPT

The leaks point to 72CU, Navi10 and no RT.
If you were to believe them.
 

stopitgetsomehelp

Neo Member
Oberon is PS5 99.9%

Maybe yes... maybe no. But it's highly possible.

the One X still retailing for $499 in 2019

You can buy One X bundle for $350 on MS Store.

The leaks point to 72CU

I missed smth? 72CU?

Oberon and Arden are within ~20%.

While Arden has more TF, depending on chip speed they choose (14/16 or 18Gbps) it can be anywhere from 30% advantage Arden to 3% advantage Oberon (if Arden is 14Gbps and Oberon is 18Gbps - because Arden is 320bit bus and Oberon is 256bit bus).

On the other hand, if they have same amount of ROPs Oberon will have advantage in pixel fillrate and geometry processing. All in all ~10-20%

Interesting +1
 

bitbydeath

Gold Member
Maybe yes... maybe no. But it's highly possible.



You can buy One X bundle for $350 on MS Store.



I missed smth? 72CU?



Interesting +1

The leaks claim there are two 36CU GPU’s which have been around for over a year. It’s fudd.

Both being around for that long rules out any chance of them being incremental.
 
Last edited:
I wouldn't be surprised if the retail PS5 is a sub 9TF machine. It seems that Sony has been trying to increase the graphics clock frequency from 1.8GHz to 2GHz as a some kind of last minute emergency measure to better compete with the new MS monster console, and I'm not sure they can pull it off.

I mean, can they get good enough yields for chips reaching that high in a retail console box, with retail cooling solution? They can't do binning like AMD is doing with dGPU parts. If they can't reach that high, PS5 TF number might start with an 8.

36CU@1.85GHz = 8.52 TF
36CU@1.90GHz = 8.76 TF
36CU@1.95GHz = 8.99 TF
 
Last edited:

MadAnon

Member
I'm still thinking about those codenames. I believe Arden has nothing to do with Xbox.

On reddit I found this...



Not his assertions but the data he found in those same documents is interesting. There's certainly a connection between Arden and Oberon.
 
Last edited:
  • Thoughtful
Reactions: TLZ

R600

Banned
I'm still thinking about those codenames. I believe Arden has nothing to do with Xbox.

On reddit I found this...



Not his assertions but the data he found in those same documents is interesting. There's certainly connection between Arden and Oberon.

Well, considering Arden and Sparkman both have Xbox BC1 data I'd say this is wrong take...

Proelite Proelite can chim in, but this guy pulled Arden test when making regression test for Oberon. Literally mix up.
 
Last edited:

PocoJoe

Banned
here is my guess, loosely based on all the leaks and speculations:

Xbox Series S: 5TF, 12GB GDDR6, no disc drive, 300€
PS5: 9.2TF, 16GB GDDR6, 400€
Xbox Series X: 12TF, 20GB GDDR6, 450€ (selling at a considerable loss, trying to get the money back through gamepass subscriptions and Gold)

let's see if I'm Nostradamus 🤣


no matter the specs and prices tho, I'll get the Series X first, because Halo and it will most likely be the more powerful system, and then later the PS5... and maybe a Series S as a secondary system later on for local multiplayer in games with no splitscreen. that's my current setup too, I got a base PS4 and an Xbox One S on my PC monitor, and a Pro and One X on the 4k TV, easy local coop and stuff.

You really think that MS makes money by selling huge amount of games by few coins/month? (Game pass)

I think they are making huge losses with game pass, no way it makes money.
 
Last edited:

xool

Member
jsqjpEp.jpg


So matching the clocks of PS4, and PS4 Pro, but different CUs (less than pro) and on a +gen architecture ..

..why, what's the point ?? that isn't going to help backwards compat.. like AT ALL.

I had this feeling before about Komachi, but something makes me not believe this..

.. it's info designed to lead the gullible into thinking they're onto something, giving just enough info to make it look believable, whilst massaging the reader's ego into thinking they "understood" ..

..but it doesn't add up - there's no reason to run a RDNA GPU with less CUs at the same frequency as a GCN one, because .. it makes no sense..

wtf is going on with this. ELI5 please
 
Last edited:

R600

Banned
jsqjpEp.jpg


So matching the clocks of PS4, and PS4 Pro, but different CUs (less than pro) and on a +gen architecture ..

..why, what's the point ?? that isn't going to help backwards compat.. like AT ALL.

I had this feeling before about Komachi, but something makes me not believe this..

.. it's info designed to lead the gullible into thinking they're onto something, giving just enough info to make it look believable, whilst massaging the reader's ego into thinking they "understood" ..

..but it doesn't add up - there's no reason to run a RDNA GPU with less CUs at the same frequency as a GCN one, because .. it makes no sense..

wtf is going on with this. ELI5 please
Its internal AMD GPU test.

It is legit.

It contains other chips (Renoire - next years Notebook APU and Arden/Sparkman (Xbox)

Komachi didnt want to share more. He literally said its gonna get someone into trouble.

Xbox chips match with Scarlett video (meaning, 320bit bus and 102983 lines of GPU data)

Again, its 100% legit. Navi CUs have legacy method as well, and for Sony its important that CU and clock match up perfectly (just watch Cernys interview about it. Even few MHZ can throw emulation of).
 

xool

Member
You really think that MS makes money by selling huge amount of games by few coins/month? (Game pass)

I think they are making huge losses with game pass, no way it makes money.

I did some sums ~ at post promotion prices it's $120 per year ($10 a month) .. for a bunch of games that I (average gamer) didn't buy, and probably wouldn't buy even at discount. .. yes they have AAA like gears5 and halo - but one of those comes out every 3+ years of something ..

.. and even if Witcher 3 (or whatever) is on the pass now, there's no way me (average gamer) is going to wait 18 months or whatever to play it - I buy stuff like that at launch, or a month later .. but not a year.

It reminds me of the free films you get on Amazon Prime - the good ones I already saw, the rest aren't worth watching.

Pretty sure it makes money. Could work for someone who likes all you can eat offeres, and very core Xbox gamers. Not a loss maker at $10/month though -
 
Last edited:

CyberPanda

Banned
Its internal AMD GPU test.

It is legit.

It contains other chips (Renoire - next years Notebook APU and Arden/Sparkman (Xbox)

Komachi didnt want to share more. He literally said its gonna get someone into trouble.

Xbox chips match with Scarlett video (meaning, 320bit bus and 102983 lines of GPU data)

Again, its 100% legit. Navi CUs have legacy method as well, and for Sony its important that CU and clock match up perfectly (just watch Cernys interview about it. Even few MHZ can throw emulation of).
Komachi is a female.😆
 
Silly question, but cheap the console will not be thats for sure.
Silly? Not really. In the gen before this one thats exactly what happened. IBM chips, with Nvidia GPUs on PS3, and IBM and AMD on the 360.
Intel chips are still better for gaming than AMD, and Nvidia make better GPUs. I think both companies would do the best deals they could to get the console business away from AMD.
 

xool

Member
Its internal AMD GPU test.

It is legit.

It contains other chips (Renoire - next years Notebook APU and Arden/Sparkman (Xbox)

Komachi didnt want to share more. He literally said its gonna get someone into trouble.

Xbox chips match with Scarlett video (meaning, 320bit bus and 102983 lines of GPU data)

Again, its 100% legit. Navi CUs have legacy method as well, and for Sony its important that CU and clock match up perfectly (just watch Cernys interview about it. Even few MHZ can throw emulation of).

But Instructions Per Clock is completely different from GCN to RDNA - this is my point - matching clocks in this case achieves nothing - it's Computer Science Nonsense

And how do you/we know it's legit?, and not say, something he/she pulled out his ass to get clicks on twitter (worst case)
 
Last edited:

TLZ

Banned
The leaks point to 72CU, Navi10 and no RT.
If you were to believe them.
Very wild speculation here, but what if those two are in the PS5 running at different speeds? Didn't we have two in the Pro iirc? So one is 36CU @2ghz (9.2tf), while the other 36CU @0.9ghz (4.1tf) for BC. Can this a possibility? That would bring the total to 13.3tf. Maybe this is why we have a V shape?

Just throwing ideas out there.
 
Last edited:
So have any of the ERA insiders such as Matt or Klee come out with anything after the 9.2flop reveal?
If they had official spec sheets, as they claim, they would be pretty confident to comment on it.
 

stopitgetsomehelp

Neo Member
Sparkman has BC mode for XO:



Arden? idk, linked to Oberon (GitHub leak, Komachi source??)




"It was a typo mistake in config file that caused Arden tests results to show up in one of the files."
 

bitbydeath

Gold Member
Sparkman has BC mode for XO:



Arden? idk, linked to Oberon (GitHub leak, Komachi source??)




"It was a typo mistake in config file that caused Arden tests results to show up in one of the files."


They’re linked because it’s just AMD hardware and nothing to do with next-gen. Every time I point out the holes in everyone’s theory they fall silent.
 
They’re linked because it’s just AMD hardware and nothing to do with next-gen. Every time I point out the holes in everyone’s theory they fall silent.

Yeah i struck me also very late i must admit - in the aftermath of this Attack against Sony it came clear that Arden actually is also connected to that whole bunch of Shakespear Codenames. obviosly not fitting with the Theory that those where Sony only related, when i real it was all about AMDs own codenames for their Products (wich may still include some of the APIs from Console manufacturers)
 
Last edited:
Status
Not open for further replies.
Top Bottom