Support NeoGAF

[policestationx]

AMD test leak is really the best leak for 2019

 

[TheStruggler]

And after all these years you're still the same, Dog! You didn't change at all. If you're trying to make us to look like fools....DON'T!

And this tweet is prime example isn't different than bunch of similar tweets of yours week ago, month ago, year ago, years ago.

Lmao does everything half assed! Still, yet they do the ONE thing MS cant do, Sony does it phenomenally.

 

[MadAnon]

A more articulated post that captured my exact thought about this leak. Kudos to Jeffram.

Here are the text of it: Credit to Jeffram from Reeeee:

"Since you snuck this in at the end of the last OT, figured i'd respond here. There are some reasons NOT to put weight into the Github leak

There are 3 reasons to question the implications of the leaked documents:

1. They are full of mistakes and inaccuracies

• There are files in wrong folders, there are sub folders and wrong folders. Again, when people say these are legitimate, they mean that this is what is legitimately in the files, not that the files are correct or accurate all the time.
• The biggest issue is that they claim Spakrman (Lockhart) is 56CUs. Every reason you have to believe Oberon is PS5, you have to believe that Sparkman is Lockhart. Sparkman BC mode indicates it can replicate a One S.

However confident you are that PS5 is 36 CUs, must must be equally confident that Lockhart is 56CUs. Or the other way around, how ever skeptical you are that these leaks confirm Lockhart is 56CUs, you must be equally skeptical these leaks confirm PS5 is 36CUs.

2. The leaks if taken at face value, fly in the face of every credible insider we have, including develeper sources and industry sources, that say the Series X and PS5 are very close in power.

• We've had Jason Schreier News Editor of Kotaku, Andrew Reiner Executive Editor at Game Informer, Our own Mods, and Vetted industry insider and veteran Klee, and other insiders all say the consoles are close in power.
• They specifically claim PS5 and Scarlett were close in power according to those sources. And All of that is in the same time window these leaks are from, which according to DF is the point where it's too late to make changes. None of them have come out and said the the tables have turned. The Github leaks are new to us, but they are actually older than the leaks we've gotten from insiders.

So, Actual insiders who actually know about the target specs of the PS5 and Xbox Series X (rather than drawing conclusions from indirect and uncontextualized data points) with a network of sources (instead of an interns notepad) are saying something to the contrary of what the leak indicates at face value.

3. The documents, if taken at face value, don't pass the sniff test, and an alternative makes more sense

• The test were done without VRR and RT, when we know that at the very least PS5 has RT. This means that the tests we have access to are not the full PS5 in at least one way.
• 36CUs at 2GHz, is not a reasonable way to hit 9.2TFs, if that's the target. Every bit of historical and contextual data we have for consoles points to a wide and slow approach to console APUs. Efficiencies, Thermals, Cooling Costs, and reliability have all proven to be reasons that trump a lower sized and faster APU. Why would Sony feel any different? What technology would change the equation? I'm not aware of any, so I don't expect that approach to change.
• 36CUs would also make this the smallest APU Sony has made in a decade. In a world where Sony is dominating in terms of sales and profitability, where they've come out and aid that PS5 is going to be Niche product aimed at the hardcore who what the best and latest, where that is corroborated by their rumoured bleeding edge SSD tech, a small APU doesn't make sense.
• Digital Foundry Stated in the article that many BC test are being done. If that's true, this could be a BC test right? Well, what would that look like. Presumably, they would want to test a Native PS4 mode, a PS4 Pro mode, and like the PS4 Pro had.... A PS4 Pro Boost mode. What would the Boost mode test look like?
• A PS4 Pro Boost mode compatibility test would look like this: Exactly the same CUs as the PS4 Pro, An unlocked "full" clock, no RT or VRS hardware activated. What do we have in the leak? 36CUs (exactly the same as the PS4 Pro), 2Ghz, the rumoured full clock of the PS5, No RT and no VRS indicated.

The leaks makes little sense as the Full PS5 hardware, indicating you shouldn't take it as a certainty of the PS5s total performance, and a lot of sense that it's a boost mode PS4 Pro compatibility test so that possibility can't be dismissed"

Those boost mode points don't make any sense. Why would you need PS4 pro mode at all then? You would only need PS4 mode and then just run all 36CUs at full clocks to boost just like PS4 pro did. Why there's native mode? To boost the boosted? Complete nonsense.

 

[Nikana]

Those boost mode points don't make any sense. Why would you need PS4 pro mode at all then? You would only need PS4 mode and then just run all 36CUs at full clocks to boost just like PS4 pro did. Why there's native mode? To boost the boosted? Complete nonsense.

Who is jeffram

 

[mckmas8808]

Those boost mode points don't make any sense. Why would you need PS4 pro mode at all then? You would only need PS4 mode and then just run all 36CUs at full clocks to boost just like PS4 pro did. Why there's native mode? To boost the boosted? Complete nonsense.

Do you love console warring this much? Like do you honestly believe Sony will only have 36 active CUs in the PS5?

 

[MadAnon]

Do you love console warring this much? Like do you honestly believe Sony will only have 36 active CUs in the PS5?

How's this console warring? Couldn't care less which one ends up better. I'm just adressing a point which makes no sense.
The whole Sparkman, Lockhart is also nonsense because there's no data there about Sparkman's real CU count. There was info about Arden.

He definitely makes some logical statements though.

 

[BattleScar]

A user already called him for bs in that thread.

DF > Jeffram's headcanon

By DF, you mean Digital Foundry?
Ah yes. They have such a wonderful track record don't they. Those early PS4 sales certainly didn't mean much in the long run eh?

If you're referring to the GitHub repo, then sure. But Digital Foundry are really only relevant when examining game performance on existing hardware. They are absolutely no authority on stuff that is yet to be released.

 

[it_wasn't_me]

Who is jeffram

Ordinary member with maybe some common sense and logical conclusions.

 

[Deleted member 471617]

I'll say PS5 is 9-10TF and $400 at launch. PS5 Pro will be 12TF and $400 in 2023.

 

[Aceofspades]

Those boost mode points don't make any sense. Why would you need PS4 pro mode at all then? You would only need PS4 mode and then just run all 36CUs at full clocks to boost just like PS4 pro did. Why there's native mode? To boost the boosted? Complete nonsense.

Because they are testing native compatibility and want to run PS4 games higher than 4k?
Sony basically co developed Navi, they want to ensure best PS4 games transition.

 

[BattleScar]

Those boost mode points don't make any sense. Why would you need PS4 pro mode at all then? You would only need PS4 mode and then just run all 36CUs at full clocks to boost just like PS4 pro did. Why there's native mode? To boost the boosted? Complete nonsense.

For compatibility/stability? Sony's intention is to make it easy for the developer so they don't have to do any extra work in to the "Pro" or indeed the "PS5" versions. If the code doesn't run properly at the full GPU clocks, without a patch, then its just easier to run the game at Pro clocks, and if that fails, at PS4 clocks. Its always nice to have a failsafe solution. It makes for a better product for the end user.

 

[mckmas8808]

For compatibility/stability? Sony's intention is to make it easy for the developer so they don't have to do any extra work in to the "Pro" or indeed the "PS5" versions. If the code doesn't run properly at the full GPU clocks, without a patch, then its just easier to run the game at Pro clocks, and if that fails, at PS4 clocks. Its always nice to have a failsafe solution. It makes for a better product for the end user.

Plus........even if Sony doesn't end up doing it that way, they have to AT LEAST test it.

 

[MadAnon]

Because they are testing native compatibility and want to run PS4 games higher than 4k?
Sony basically co developed Navi, they want to ensure best PS4 games transition.

How exactly you will run ps4/pro games at higher resolutions without patching them to take advantage of extra power? As I said, such mode makes no sense in the first place.

 

[Proelite]

Seems I missed quite a bit, If you think there is that much of a difference in these machines power then you're going to be disappointed. They are both very close and Neither is under 10. I will stand on what I said they are both very competitive and the difference is minimal. You have a month a week and a few days until its official.

Do you stand on Ps5 > Scarlett? You said that too.

 

[Fortuneelsewhere]

It baffles me how for so long we already know that both SEX and PS5 will be like twins regarding GPU power but some recuse to believe it.. the information is there for long enough with respectable people in the business telling/hinting us so... even in this forum people in the know that rightfully so keep low profile hinted at that in some posts.. (you just need to pay attention)

But Some prefer to believe and put more weight in some incomplete and out of context data overcomplicating the obvious just to feed their ego and system of choice...

For some the awakening will be rude..

 

[xool]

RedGamingTech reports messages from "sources'" that Sony are making changes that *may* delay the PS5.

RGT are sceptical.

Imagine the freedom to be sceptical about your own made up BS

youtube was a mistake

 

[Reindeer]

Common, we all know how those RDX guys roll. They can act like they're not fanboys, but time and time again they expose their bias.

 

[Aceofspades]

How exactly you will run ps4/pro games at higher resolutions without patching them to take advantage of extra power? As I said, such mode makes no sense in the first place.

I believe Sony would know how to do it, that's why they are experimenting with Oberion chip. I can't answer you since your question is beyond my knowledge in this field.

 

[JordanN]

I'll say PS5 is 9-10TF and $400 at launch. PS5 Pro will be 12TF and $400 in 2023.

Serious question: how would the games even benefit from this?

GPU alone is not enough to make a big difference, that PS5 Pro would still be stuck with the same CPU that's already in the base PS5.

Just my opinion, but I think releasing a mid-gen refresh that's only 30% faster but with all the other same guts in it seems like a dumb idea.

Even the Gamecube to Wii jump was a lot bigger than this...

 

[CyberPanda]

 

[Celcius]

With next gen systems using nvme ssd’s but Sony being price conscious, I could see them releasing a 1tb console for one price and then a 2tb console for a higher price, like they did with ps3 back in the day.

 

[R600]

Because they are testing native compatibility and want to run PS4 games higher than 4k?
Sony basically co developed Navi, they want to ensure best PS4 games transition.

But why is Arden Native 56CUs? If Oberon has them missing, why are Ardens 56CUs there?

 

[Reindeer]

I'll say PS5 is 9-10TF and $400 at launch. PS5 Pro will be 12TF and $400 in 2023.

That's very conservative. I think PS5 will be around 2X as powerful as base console or there isn't really much of reason to buy it. Also I doubt they will launch just 3 years after again as they would want to avoid launching another weak Pro console.

 

[DJ12]

Because they are testing native compatibility and want to run PS4 games higher than 4k?
Sony basically co developed Navi, they want to ensure best PS4 games transition.

I personally think this is for ms's xcloud servers which will of course need playstation compatibility going forwards as well as Xbox. This is the real reason there is Oberon and Arden in the same set of data in my opinion.

I know the 'I have spoken' posse will say this isn't a possibility but it won't be long before they've all gone when it's confirmed ps5 isn't 9 tflops and/or Xbox isn't really 12 tflops rDNA.

 

[xool]

“See lots of robots”

PS5 reveal teaser with HZD2 footage?

He posted an “end transmission” satellite in a tweet a couple of weeks back, now this...

Lost In Space game confirmed!

("end transmission" is possible HZD reference right ?)

Herman Hulst is gone from Guerilla. What if Balrog is directing HZD2 ??

 

[Tiago07]

What do you guys think about part of 56 CU's of Series X be dedicated to Ray Tracing performance ? We have no mention of a dedicate chipset to process Ray Tracing in Xbox, but we know that Sony is making a solution to them

 

[Marlenus]

What do you guys think about part of 56 CU's of Series X be dedicated to Ray Tracing performance ? We have no mention of a dedicate chipset to process Ray Tracing in Xbox, but we know that Sony is making a solution to them

pointless. they will be gpgpu shaders and devs can choose how to use them.

 

[demigod]

9.2TF folks are going to look really dumb when Sony reveals the specs.

 

[DJ12]

But why is Arden Native 56CUs? If Oberon has them missing, why are Ardens 56CUs there?

Maybe they are there, they just weren't part of the test. Just like you claim about RT.

Both are unlikely but hey if you think that's a good enough excuse I'll use it too. Guess this will be complete rubbish unlike your hypothasis though... Right? lol

 

[Kazekage1981]

Generally speaking why are PS5 and Xsex limited to 36-56 Compute Unit Range? Why cant it be 64-128 Compute Unit Range? Is it better to have higher GPU speed frequency and less compute units, or more compute units and less GPU speed frequency? I am a noob.

 

[Raploz]

I don't see a 9.2TF PS5 as a bad thing as long as it's priced well. The success of a new console depends on the perceived value.

A powerful Xbox Sx at 500$ would probably sell more than a PS5 at 400$, but a 600$ Xbox vs a 400$ PS would fail, even if it's more powerful, because the additional power wouldn't justify a 200$ price difference.

Anyway, I'm happy with the new generation, even if there's a weaker console, it's still a significant upgrade over base consoles. Games will look amazing.

 

[shenmue mugen]

People, consider this...

What doesn't make sense to me is the way Phil Spencer talked about XSEX power. Instead of shouting the TF value like they did every single oportunity with the Xbox One X he said: "more than 8 times base Xbox One GPU and 2 times Xbox One X GPU" (he omitted the word "more" when comparing to the One X GPU)

Well. If we use this same rethoric of using performance INSTEAD of raw TF value the rumoured 9.2 TF PS5 RDNA GPU (which would translate to 11.5 GCN teraflops) fits perfectly in the description of the X Series GPU.

PS5 9.2 TF RDNA GPU = (roughly) 11.5 TF GCN GPU.

11.5 GCN teraflops is exactly 8.8 times base Xbox One GPU (which falls perfectly in the "MORE than 8 times base Xbox One GPU"...which means higher than 8 but LOWER than 9)

11.5 GCN teraflops is exactly 1.9 times Xbox One X (which falls perfectly in the "2 times Xbox One X GPU" ....because, as I said before he didn't use the word "more" like when he was comparing with base Xbox One becuse it is actually 1.9 times)

If XSEX GPU is 12 RDNA TF that means it's GPU is roughly 14.3 GCN TF.

If this is the case Phil Spencer's speech simply does not make any sense because that would make XSEX' s GPU exactly 11 times more powerful than base Xbox (14.3÷1.3) and 2.38 times more powerful than the One X GPU (14.3÷6...and he never said the word more like he did when comparing to base Xbox One)

I mean...performance wise, he just described the rumoured PS 5 9.2 RDNA TF GPU.

Can anyone make sense of all this? Because I sure as hell can't.

 

[SportsFan581]

Kazekage1981 , in the scheme of things, these are cheap boxes and always will be. There is a limit to what size die is reasonable in both cost and yield.

 

[onQ123]

I prefer 8TF GPU if it means that more room was used for accelerators & stuff.

 

[Marlenus]

People, consider this...

What doesn't make sense to me is the way Phil Spencer talked about XSEX power. Instead of shouting the TF value like they did every single oportunity with the Xbox One X he said: "more than 8 times base Xbox One GPU and 2 times Xbox One X GPU" (he omitted the word "more" when comparing to the One X GPU)

Well. If we use this same rethoric of using performance INSTEAD of raw TF value the rumoured 9.2 TF PS5 RDNA GPU (which would translate to 11.5 GCN teraflops) fits perfectly in the description of the X Series GPU.

PS5 9.2 TF RDNA GPU = (roughly) 11.5 TF GCN GPU.

11.5 GCN teraflops is exactly 8.8 times base Xbox One GPU (which falls perfectly in the "MORE than 8 times base Xbox One GPU"...which means higher than 8 but LOWER than 9)

11.5 GCN teraflops is exactly 1.9 times Xbox One X (which falls perfectly in the "2 times Xbox One X GPU" ....because, as I said before he didn't use the word "more" like when he was comparing with base Xbox One becuse it is actually 1.9 times)

If XSEX GPU is 12 RDNA TF that means it's GPU is roughly 14.3 GCN TF.

If this is the case Phil Spencer's speech simply does not make any sense because that would make XSEX' s GPU exactly 11 times more powerful than base Xbox (14.3÷1.3) and 2.38 times more powerful than the One X GPU (14.3÷6...and he never said the word more like te did when comparing to base Xbox One)

I mean...performance wise, he just described the rumoured PS 5 9.2 RDNA TF GPU.

Can anyone make sense of all this? Because I sure as hell can't.

RDNA Tflops are equal to GCN Tflops. What is not the same is the fps per TFlop that RDNA can extract Vs GCN.

They are going to talk about an objective value like the TFlop figure rather than a more variable and subjective figure like performance as that will vary by game.

 

[Raploz]

Generally speaking why are PS5 and Xsex limited to 36-56 Compute Unit Range? Why cant it be 64-128 Compute Unit Range? Is it better to have higher GPU speed frequency and less compute units, or more compute units and less GPU speed frequency? I am a noob.

Unless you're willing to pay 2000$ for a new console a 128CU chip won't happen. There's also scalability limitations of the architecture, heat output, power consumption and a bigger chip would also be harder and more expensive to manufacture.

 

[Deleted member 471617]

Serious question: how would the games even benefit from this?

GPU alone is not enough to make a big difference, that PS5 Pro would still be stuck with the same CPU that's already in the base PS5.

Just my opinion, but I think releasing a mid-gen refresh that's only 30% faster but with all the other same guts in it seems like a dumb idea.

Even the Gamecube to Wii jump was a lot bigger than this...

The games really wouldn't benefit from it. I think the 9-10TF is to keep the price at $400. And the 12TF is just to bump the resolution to 8K CBR. But the extra GPU boost would make games look overall better. That would probably be the benefit along with the resolution bump.

I'm personally hoping for the best Sony can give me for $500, no mid-gen and instead, November 2025, Sony launches PS6 with Native 8K and whatever else at that time for $500. One thing I firmly believe is that if there's a PS5 Pro, it's just to get to 8K CBR resolution. Anything more would be a bonus.

That's very conservative. I think PS5 will be around 2X as powerful as base console or there isn't really much of reason to buy it. Also I doubt they will launch just 3 years after again as they would want to avoid launching another weak Pro console.

My prediction a year ago or whenever I voted was 8TF just so it can hit Native 4K but that was before knowing about Ray Tracing being on PS5. Even if PS5 is only 8TF, the sequels to all the exclusives from PS4 will be more than enough for me to buy PS5 and tens of millions of PS4 gamers will be the same.

 

[Gamernyc78]

9.2TF folks are going to look really dumb when Sony reveals the specs.

I am waiting anxiously for any info.

 

[Kazekage1981]

The holidays are over, and the new year has begun! Get ready for the hype train! Choo CHooooo!!!!

-MEGA HYPEREALISTIC PHOTOSYNTHESIS ULTRA MEGATON! SO HYPERPHOTO REAL, YOU WILL GO BLIND IN 10 MINUTES!
-SUPER DOOPER 3D AUDIO SOUND THAT IS MORE REAL SOUNDING THAN REAL LIFE! HEAR THE ZOMBIE CREAPING UP FROM YOUR BEHIND AND THEN ZOOM PAST YOU AS IT BEGINS TO FIGHT THE MANBEARPIG!
-TRILLIONS OF TEXTURE MAPPED POLYGONS WITH THE POWER OF ZEN CPU AND RDNA GPU TO MAKE A SPHERE THAT REFLECTS LIGHT LIKE REAL LIFE, EXCEPT ITS BETTER CAUSE ITS 4K HDR AND USING GPUGPU COMPUTE WITH THE NEW HYPER ADVANCED SSD TO MAKE MORE SPHERES THAT REFLECTS LIGHT BASED ON REAL LIFE!
-XSEX USING THE POWER OF THE CLOUD TO TALK TO OTHER XSEX'S USING NEURAL NETWORK SKYNET BRAIN! XSEX NEURAL NETWORK SKYNET BRAIN IS SUPER ADVANCED AI THAT IS SUPER SMART THAT IS BEYOND ULTRA HARD SETTINGS, SO THAT WHEN YOU AS THE PLAYER FIGHT THE MANBEARPIG WITH YOUR XBOX ELITE CONTROLLER...it will be smart.

Oh and dont forget the switch, thats where you can play mario, zelda, metroid

 

[Snake29]

Funny thing is that we already expected at least 10TF for the next-gen consoles. So why would 9.2/10TF now something bad or "underperforming", as if all other components in the console won't mean much.

I can't wait for the official reveal and and the PS5 is a no brainer for me.

 

[mckmas8808]

RDNA Tflops are equal to GCN Tflops. What is not the same is the fps per TFlop that RDNA can extract Vs GCN.

They are going to talk about an objective value like the TFlop figure rather than a more variable and subjective figure like performance as that will vary by game.

You don't know this. Depends on who is controlling the message.

 

[SportsFan581]

8k is 4x the pixels, don't expect these consoles to do anything serious at 8k (possible mid gens included).

 

[Lunatic_Gamer]

Dualshock 4 Will Work With PS5

Users won't have to spend extra money to get additional controllers with PS5. Dualshock 4 will work just fine with your PlayStation 5.

respawnfirst.com

 

[XcloudTimdog]

People, consider this...

What doesn't make sense to me is the way Phil Spencer talked about XSEX power. Instead of shouting the TF value like they did every single oportunity with the Xbox One X he said: "more than 8 times base Xbox One GPU and 2 times Xbox One X GPU" (he omitted the word "more" when comparing to the One X GPU)

Well. If we use this same rethoric of using performance INSTEAD of raw TF value the rumoured 9.2 TF PS5 RDNA GPU (which would translate to 11.5 GCN teraflops) fits perfectly in the description of the X Series GPU.

PS5 9.2 TF RDNA GPU = (roughly) 11.5 TF GCN GPU.

11.5 GCN teraflops is exactly 8.8 times base Xbox One GPU (which falls perfectly in the "MORE than 8 times base Xbox One GPU"...which means higher than 8 but LOWER than 9)

11.5 GCN teraflops is exactly 1.9 times Xbox One X (which falls perfectly in the "2 times Xbox One X GPU" ....because, as I said before he didn't use the word "more" like when he was comparing with base Xbox One becuse it is actually 1.9 times)

If XSEX GPU is 12 RDNA TF that means it's GPU is roughly 14.3 GCN TF.

If this is the case Phil Spencer's speech simply does not make any sense because that would make XSEX' s GPU exactly 11 times more powerful than base Xbox (14.3÷1.3) and 2.38 times more powerful than the One X GPU (14.3÷6...and he never said the word more like he did when comparing to base Xbox One)

I mean...performance wise, he just described the rumoured PS 5 9.2 RDNA TF GPU.

Can anyone make sense of all this? Because I sure as hell can't.

i can, its rdna 12 tflop.

 

[Marlenus]

You don't know this. Depends on who is controlling the message.

They will use something incontrovertible after the Xbox one launch fiasco with always online.

 

[thicc_girls_are_teh_best]

I personally think this is for ms's xcloud servers which will of course need playstation compatibility going forwards as well as Xbox. This is the real reason there is Oberon and Arden in the same set of data in my opinion.

This is actually plausible. But I'm also thinking, since Sony will be using Azure for PS cloud/streaming next gen and MS has already been on record saying they'll be using next-gen Xboxes as servers (kind of like what they've done with XBX), doesn't that lend some credibility to the idea that XSEX will be at least slightly more powerful than PS5? How is a weaker console going to serve as a host server emulating a more powerful system for game streaming and cloud purposes?

Unless maybe MS will still be using purpose-speced PCs in server racks for that purpose as well. Which is possible, but sounds a bit cumbersome of a setup.

 

[FranXico]

Dualshock 4 Will Work With PS5

Users won't have to spend extra money to get additional controllers with PS5. Dualshock 4 will work just fine with your PlayStation 5.

respawnfirst.com

Likewise, I expect a PS4 firmware update to give it DS5 support (like they did for the PS3).

 

[FranXico]

Unless maybe MS will still be using purpose-speced PCs in server racks for that purpose as well. Which is possible, but sounds a bit cumbersome of a setup.

Cumbersome? That actually is more economically viable than custom blades.

 

[shenmue mugen]

i can, its rdna 12 tflop.

But that would put XSEX performance compared to base model in a place that makes Phil's statement grossely misguided.

It's like Phil subtracted 3 teraflops from the performance delta between the XSEX and base Xbox One.

 

[XcloudTimdog]

But that would put XSEX performance compared to base model in a place that makes Phil's statement grossely misguided.

It's like Phil subtracted 3 teraflops from the performance delta between the XSEX and base Xbox One.

Its 12 tflop, its in the code, its also been explicitly told to me .