Gamernyc78
Banned



Last edited:
Good boi.K mom.
Don't be surprised if someone quotes your post in era with the comic and make a big deal about and say "omg gaf is so so baddddd, those assholes"
Xbox plan was 12 tflops from the start. Throw it outAbsolute BS.
Sure more info has leaked, but the 9.2tflop number was apart of the leak back in August, which they wouldn't have known unless they had the CU count.Its because only details regarding Oberon that leaked by Komachi were clocks (native, BC1 and BC0).
Now CUs, bandwidth, cache bandwidth and million other things leaked.
Lol not surprised in the slightest. Resetera will get rid of you at the drop of a hat if you don't align with their political views.But leave it to ERA to make me look like a jackass.
![]()
Even ree/reddit is doubing this speculation, unless they want to invalid all devs saying the gap is close.I'm expecting both consoles to be pretty close. I don't buy the PS5 ~9TF if the Xbox is targeting 12TF. Just pure opinion of course,
I'm expecting both consoles to be pretty close. I don't buy the PS5 ~9TF if the Xbox is targeting 12TF. Just pure opinion of course,
Actually, data in repo says they are close. Closer then Xbone vs PS4, and even closer then Pro vs X. So, yea...close.Even ree/reddit is doubing this speculation, unless they want to invalid all devs saying the gap is close.
Klee and other insiders have said that nothing has changed since Xbox Series X reveal regarding the power differences. PS5 still has the edge.I'm expecting both consoles to be pretty close. I don't buy the PS5 ~9TF if the Xbox is targeting 12TF. Just pure opinion of course,
I THINK WE CRACKED IT!
Howdy folks. A user on Era pointed out that the GPU being used in Github leaks for Oberon are pretty much identical to a 5700. Upon realizing this it all clicked into place for me. The unit being tested here must be the early low speed devkit Cerny showed off to Wired in April (the big silver tower) that was running a 5700 at high clocks to get as much performance out of it as possible. This explains why the unit in the Github leaks had no RT hardware and no VRS. It also explains why there were rumors going around GDC that the PS5 was 9.2 TF because that was the speed of the early low speed devkits up until that point
After GDC Sony began shipping out the new V shaped devkits and since then we've been hearing about a much more powerful PS5 from insiders, this seems to be the unit where all the current rumors stem from.
So basically, it looks like the Github leak was just using an early, low speed devkit that had a re-purposed 5700 as the GPU.
Comments?
you cant just upclock a gpu and get more power it doesn't scale in a linear way. im sure somebody can post a Hz power wattage chart.
Read it. Couldnt be more wrong.I THINK WE CRACKED IT!
Howdy folks. A user on Era pointed out that the GPU being used in Github leaks for Oberon are pretty much identical to a 5700. Upon realizing this it all clicked into place for me. The unit being tested here must be the early low speed devkit Cerny showed off to Wired in April (the big silver tower) that was running a 5700 at high clocks to get as much performance out of it as possible. This explains why the unit in the Github leaks had no RT hardware and no VRS. It also explains why there were rumors going around GDC that the PS5 was 9.2 TF because that was the speed of the early low speed devkits up until that point
After GDC Sony began shipping out the new V shaped devkits and since then we've been hearing about a much more powerful PS5 from insiders, this seems to be the unit where all the current rumors stem from.
So basically, it looks like the Github leak was just using an early, low speed devkit that had a re-purposed 5700 as the GPU.
Comments?
AMD can't supply what not exist. Those first devkits probably used VEGA.Why would they go to the trouble of making an early APU just for temporary dev kits.
Why would they go to the trouble of making an early APU just for temporary dev kits. At that stage of the game an alpha kit PC with standard parts would suffice.
I have to say mods here on neogaf are awesome. I havent seen more polite and cool mods anywhere else. These guys still have borders and will punish people for their actions but they act more like friends, and not like policemen.
AMD can't supply what not exist. Those first devkits probably used VEGA.
Read it. Couldnt be more wrong.
Its like saying PS4 or PS4Pro dont exist because, in AMD test repo their numbers coorespond to 7870 and RX480. Just bizzare...
Early kits, released in 2018, were almost 100% Vega based. So yea, after GDC, devs got their Oberons.Yet we know an early, low speed devkit exists and was in use when Cerny talked to Wired in April and these test are from around June. And we know a new V shaped devkit went out to devs just after GDC.
Can you come up with a better explanation that fits all the data including the multiple reports we've heard from insiders and journalists that indicate a PS5 with more power? Why does this devkit lack RT when we know RT is in the PS5's GPU?
It all fits with this theory.
They will start to eat their own. They always do. These self imposed utopias never last. They are all putting up purity tests they themselves cant pass. (To quote Pete)G
Lol not surprised in the slightest. Resetera will get rid of you at the drop of a hat if you don't align with their political views.
Early kits, released in 2018, were almost 100% Vega based. So yea, after GDC, devs got their Oberons.
People dont realize...Its not that Oberon is 64CU chip clocked at 1.0GHz, its 36CU chip clocked at 2.0GHz.
Basically, any changes you wanted to do on logic side, that means materials and actual physical characteristics had to be done in summer at latest..
Summer of 2018 that is.
You are trying too hard. The evidence it pretty obvious. If Sony has done something different that is outside these GPUs then ok, but we dont know that. All evidence points towards a 9.2tflop 36 CU GPU. However if new evidence comes out, maybe from Sony themselves, then people will re address their views based on new evidence.I THINK WE CRACKED IT!
Howdy folks. A user on Era pointed out that the GPU being used in Github leaks for Oberon are pretty much identical to a 5700. Upon realizing this it all clicked into place for me. The unit being tested here must be the early low speed devkit Cerny showed off to Wired in April (the big silver tower) that was running a 5700 at high clocks to get as much performance out of it as possible. This explains why the unit in the Github leaks had no RT hardware and no VRS. It also explains why there were rumors going around GDC that the PS5 was 9.2 TF because that was the speed of the early low speed devkits up until that point
After GDC Sony began shipping out the new V shaped devkits and since then we've been hearing about a much more powerful PS5 from insiders, this seems to be the unit where all the current rumors stem from.
So basically, it looks like the Github leak was just using an early, low speed devkit that had a re-purposed 5700 as the GPU.
Comments?
Do you believe that ps5 will be 36cu? Or we can deduct that from these leaks? Also where is everyone getting 56cu for Xbox Sx? Same leak. Thanks, enjoy your posts"Average member of Resetera reading Xbox One SOC test...16Rops? HD7700?? Lol, yea right..."
"Average member of Resetera reading PS4 SOC test...Hmmm....sounds like underclocked HD7870. This cant be it."
"Average member of Resetera reading PS4Pro SOC test....Lol this is underclocked RX480, this is not Pro."
"Average member of Resetera reading Scorpio SOC test...Hmmm....sounds like RX580. This cant be it."
No? There is no RDNA1 or 2 for consoles.Didn't the Github leak indicate RDNA1 was being used? There was a lot of info being tossed around but I do recall people mentioning that.
Navi was announced to the public in the middle of 2019, meaning it had to exist in some way, shape or form well before that. Sony might have been able to secure some very early variants from AMD for use in their early devkits.
Its clear as a day that it is 36CU chip.Do you believe that ps5 will be 36cu? Or we can deduct that from these leaks? Also where is everyone getting 56cu for Xbox Sx? Same leak. Thanks, enjoy your posts
Interesting , thats what i have heard as well. I dont get how anything can be even touching 10 tflop.No? There is no RDNA1 or 2 for consoles.
Console makers took building block from Navi and added AMDs roadmap features on chip. Thats why its called - custom.
RDNA1 and 2 will be 99% the same under the skin, but VRS and RT will come equiped + better thermals duo to node and process maturation. It same arch.
Its clear as a day that it is 36CU chip.
Xbox will most likely be a 12 tflops RDNA GPU. If it's 56 CU, then it's going to be really easy to get 12tflops from that without going mental on the frequency.In reality, anything is possible at this point. MS has actually never stated that Xbox is at 12TF, they said it had a GPU that was 2x more powerful than Xbone X. Maybe they think a RDNA 2 card at 8.5TF is equivalent to 2x the 6TF on the older architecture. In which case, both the 9.2 leak for PS5 and the rumor that it nudges past Xbox could both be true. I'd secretly love to see 12/13TF consoles, but we don't know for sure if that is the plan.
You are trying too hard. The evidence it pretty obvious. If Sony has done something different that is outside these GPUs then ok, but we dont know that. All evidence points towards a 9.2tflop 36 CU GPU. However if new evidence comes out, maybe from Sony themselves, then people will re address their views based on new evidence.
Ball is in Sony's court.
They could, if they would raise the clocks 200MHZ more. It would be crazy good chip, especially for that size.Interesting , thats what i have heard as well. I dont get how anything can be even touching 10 tflop.
Also the 2 SKU Xbox approach makes the most sense if they fall either side of the PlayStation.
We're talking about a setup specifically configured by AMD for a big, loud devkit tower. What does it matter if it doesn't scale linearly? More power is more power.
If its indeed 36cu they can only do so much. Either we believe in this leaks, or we throw them out, thats for all of them. I highly doubt that tho. Arden was told to me to be in xbox codeXbox will most likely be a 12 tflops RDNA GPU. If it's 56 CU, then it's going to be really easy to get 12tflops from that without going mental on the frequency.
To me it looks like Sony for next gen designed a system that was twice as powerful as their PS4 Pro, which is exactly what MS did going twice as powerful as the X.
Sony got wind of the new Xbox power levels and has tried to ramp up the frequency to reduce the gap. Whether this can be done without cooking a heap of consoles out in people houses where the console is sitting in places it cant get enough cool airflow through the unit is the big question.
No? There is no RDNA1 or 2 for consoles.
Console makers took building block from Navi and added AMDs roadmap features on chip. Thats why its called - custom.
RDNA1 and 2 will be 99% the same under the skin, but VRS and RT will come equiped + better thermals duo to node and process maturation. It same arch.
Its clear as a day that it is 36CU chip.
To be brutally honest, as far as the Next-Gen threads go, I see no difference between here and Era, other than the fact that GAF is apparently considered taboo'ed over there and mentioning it as a source of info invites the wrath of the mods.
not always, cause then you will come to a point where you need 2x more power for just 10% more tflops. not worth it.
I would almost say that the two sku approach would work best if both were at equal price on the high-end. PS at $400 would make Lockhart a weak value (for only $100 or so more you can step to the full powered PS). If both the PS and Xbox are at $500 and Lockhart is at $300 or $250, even though it's much lower end, it's almost half the cost.
I think you do not understand the actual chip design process and how much time and money it takes to put such a big and fast chip on the market.It is indeed clear that it was a 36CU chip used in that particular test, yup. Think about it, if Sony needed AMD to kitbash together a powerful devkit roughly 2 years out from launch and AMD's preparing their new Navi chips then why not use those in the devkit until they can get a proper APU up and running for the next revision? So in the early low speed kit we see a 5700 equivalent chip that's been clocked higher for use in the devkit and it gets up to 9.2 TF. For the next devkit they're able to get an APU going that's closer to the final product and we now have the V shaped devkits.
As I said, it seems to fit the timelines and the rumors absolutely perfectly. Anyways, I'm not going to try and convince you. It's a theory that I like and will continue to run with until we hear something that negates it.