Support NeoGAF

[HawarMiran]

Except that’s not what devs are telling me. The system is not 9.2TF. I would say it’s between 11.5-13TF...

In Fact had a Dev tell me in a message off site that a recent leak was 100% right with PS5. Just waiting on him to tell me which one.

For the record I can back what I’m saying up if Mod of War wants to verify it. That I’m waiting on a response to which leak by a Dev.

It is me Sir. Your mod

 

[HeisenbergFX4]

Except that’s not what devs are telling me. The system is not 9.2TF. I would say it’s between 11.5-13TF...

In Fact had a Dev tell me in a message off site that a recent leak was 100% right with PS5. Just waiting on him to tell me which one.

For the record I can back what I’m saying up if Mod of War wants to verify it. That I’m waiting on a response to which leak by a Dev.

I hope things go well for you.

Several things you have said has been spot on.

 

[WhispersintheWind]

Mod of War Check you PM Conversations. I have sent you evidence.

 

[thicc_girls_are_teh_best]

Except that’s not what devs are telling me. The system is not 9.2TF. I would say it’s between 11.5-13TF...

In Fact had a Dev tell me in a message off site that a recent leak was 100% right with PS5. Just waiting on him to tell me which one.

For the record I can back what I’m saying up if Mod of War wants to verify it. That I’m waiting on a response to which leak by a Dev.

Well why can't we find any actual chips matching that specific performance range, when it's already middle of March 2020 (same year PS5 is supposedly launching)?

That's been my biggest skepticism behind some of these newer spec rumors. They all feel almost reactionary, and continuously creeping upward, while varying wildly almost as consistently.

But to give you an example, you can technically reach 11.5TF with a 40CU chip if you have all 40CUs on clocked @2.25GHz. I mean, if Oberon has already been tested at 2GHz what's to say they have not pushed it further than that...

...is how I try relating the testing data with, say, your claim here. Except the problem is there is no clock setting in any of that data actually reflecting 40CUs @ 2.25GHz. So does that leave another yet-undocumented Oberon revision? And if so what's the setup for that revision differing it from the most recently-known one?

Because your word alone isn't going to get me to discard actual data completely. At the very least, it's speculation I can combine with that data and try finding something in the middle. But the middle doesn't necessarily have to be the obvious middle I suppose.

 

[WhispersintheWind]

I hope things go well for you.

Several things you have said has been spot on.

Thank You for your Kind words. I’m glad you have been able to verify some of the info I have passed on. That I have received.

 

[SlimySnake]

Except that’s not what devs are telling me. The system is not 9.2TF. I would say it’s between 11.5-13TF...

In Fact had a Dev tell me in a message off site that a recent leak was 100% right with PS5. Just waiting on him to tell me which one.

For the record I can back what I’m saying up if Mod of War wants to verify it. That I’m waiting on a response to which leak by a Dev.

did you ever find out if the new ps5 devkit caesar was over 11 tflops?

ram type, ssd speed extra? you said you'd ask.

 

[WhispersintheWind]

did you ever find out if the new ps5 devkit caesar was over 11 tflops?

ram type, ssd speed extra? you said you'd ask.

I have asked. Haven’t been told anything in detail. Trust me I Try weekly to contact Few people to get those fine details. Everybody talks in very Vague without giving away the fine details.

I guess because people are scared to lose their Job or get in Trouble.

 

[SlimySnake]

I have asked. Haven’t been told anything in detail. Trust me I Try weekly to contact Few people to get those fine details. Everybody talks in very Vague without giving away the fine details.

I guess because people are scared to lose their Job or get in Trouble.

cool.

and just to be sure, your sources are devs in the industry?

 

[WhispersintheWind]

cool.

and just to be sure, your sources are devs in the industry?

Yes they work in the gaming industry as devs. Mod of War can Verify when he looks at the evidence. That they do In fact work in the Industry. This is only one of the people I have talk to and still talk to.

 

[M.Night Shyamalan]

I hope things go well for you.

Several things you have said has been spot on.

Link to what was correct, please?

 

[M.Night Shyamalan]

Well why can't we find any actual chips matching that specific performance range, when it's already middle of March 2020 (same year PS5 is supposedly launching)?

That's been my biggest skepticism behind some of these newer spec rumors. They all feel almost reactionary, and continuously creeping upward, while varying wildly almost as consistently.

But to give you an example, you can technically reach 11.5TF with a 40CU chip if you have all 40CUs on clocked @2.25GHz. I mean, if Oberon has already been tested at 2GHz what's to say they have not pushed it further than that...

...is how I try relating the testing data with, say, your claim here. Except the problem is there is no clock setting in any of that data actually reflecting 40CUs @ 2.25GHz. So does that leave another yet-undocumented Oberon revision? And if so what's the setup for that revision differing it from the most recently-known one?

Because your word alone isn't going to get me to discard actual data completely. At the very least, it's speculation I can combine with that data and try finding something in the middle. But the middle doesn't necessarily have to be the obvious middle I suppose.

You find what they want you to find.

 

[tflopsmatter]

GitHub was not a hoax.
GitHub was not wrong.
GitHub has real information.

However, the GPU in GitHub is not the GPU in the PS5.

 

[M.Night Shyamalan]

GitHub was not a hoax.
GitHub was not wrong.
GitHub has real information.

However, the GPU in GitHub is not the GPU in the PS5.

This.

 

[thicc_girls_are_teh_best]

GitHub was not a hoax.
GitHub was not wrong.
GitHub has real information.

However, the GPU in GitHub is not the GPU in the PS5.

Based on feelings?

I thought the general idea was that it wasn't a full picture on the chip itself, not that it wasn't the chip whatsoever.

Also what do you even mean by "chip"? A whole new chip not Oberon? So you mean Flute? Or something no existence of which has been retrieved for yet? Do you mean a revision? Or a stepping?

Sounds like you're the one trying to push a conspiracy theory here. I'm just keeping possibilities in the running given one type of source isn't absolutely overruling another to the point of disregarding said other.

 

[HawarMiran]

Well why can't we find any actual chips matching that specific performance range, when it's already middle of March 2020 (same year PS5 is supposedly launching)?

That's been my biggest skepticism behind some of these newer spec rumors. They all feel almost reactionary, and continuously creeping upward, while varying wildly almost as consistently.

But to give you an example, you can technically reach 11.5TF with a 40CU chip if you have all 40CUs on clocked @2.25GHz. I mean, if Oberon has already been tested at 2GHz what's to say they have not pushed it further than that...

...is how I try relating the testing data with, say, your claim here. Except the problem is there is no clock setting in any of that data actually reflecting 40CUs @ 2.25GHz. So does that leave another yet-undocumented Oberon revision? And if so what's the setup for that revision differing it from the most recently-known one?

Because your word alone isn't going to get me to discard actual data completely. At the very least, it's speculation I can combine with that data and try finding something in the middle. But the middle doesn't necessarily have to be the obvious middle I suppose.

 

[01011001]

GitHub was not a hoax.
GitHub was not wrong.
GitHub has real information.

However, the GPU in GitHub is not the GPU in the PS5.

what is it for then?

 

[Rusco Da Vino]

In Fact had a Dev tell me in a message off site that a recent leak was 100% right with PS5. Just waiting on him to tell me which one.

For the record I can back what I’m saying up if Mod of War wants to verify it. That I’m waiting on a response to which leak by a Dev.

Are you expecting to get confirmation of that legit leak by today?

That would be huge!

 

[Batiman]

At this point I feel I’ve heard every combination of teraflop numbers from each insider here. It’s like almost impossible for any of them to be wrong lol. It’s way too easy to be an insider nowadays. Should of called myself Johnny smith and post range of numbers with always leaving an out for myself.

-Different dev kits
-Plans change
-What I was told
-One of my sources said this but another said that. Etc.

people will believe whatever they want to hear.

 

[WhispersintheWind]

Are you expecting to get confirmation of that legit leak by today?

That would be huge!

He Messaged me On Thursday. I asked a lot questions to do with all recent leaks to pinpoint which one. I also was asking questions about projects they are working as a studio. He hasn’t respond yet. But sometimes Devs take days to respond. I have even had some take weeks. As soon as I know anything I will be glad to share.

 

[Matreios Raven]

A few possible scenarios, in no specific order-

Sony is going for an affordable console, so 8-9.2, with a possible PRO model later.

RDNA2 allows Sony to overclock chip even with less CUs to just barely reach 11.5-12 TF.

The chip in question are not the final chip, and the PS5 GPU is 11 to 13 (max) TF with 52-56 CUs.

 

[Matreios Raven]

Well why can't we find any actual chips matching that specific performance range, when it's already middle of March 2020 (same year PS5 is supposedly launching)?

That's been my biggest skepticism behind some of these newer spec rumors. They all feel almost reactionary, and continuously creeping upward, while varying wildly almost as consistently.

But to give you an example, you can technically reach 11.5TF with a 40CU chip if you have all 40CUs on clocked @2.25GHz. I mean, if Oberon has already been tested at 2GHz what's to say they have not pushed it further than that...

...is how I try relating the testing data with, say, your claim here. Except the problem is there is no clock setting in any of that data actually reflecting 40CUs @ 2.25GHz. So does that leave another yet-undocumented Oberon revision? And if so what's the setup for that revision differing it from the most recently-known one?

Because your word alone isn't going to get me to discard actual data completely. At the very least, it's speculation I can combine with that data and try finding something in the middle. But the middle doesn't necessarily have to be the obvious middle I suppose.

With the way RDN2A2 is more efficient, and that double side chip cooling patent, perhaps Sony could squeeze out a higher TF number, comparable to the XSX by increasing the clock speed?

 

[semicool]

Yes they work in the gaming industry as devs. Mod of War can Verify when he looks at the evidence. That they do In fact work in the Industry. This is only one of the people I have talk to and still talk to.

He could be telling you GCN TFs and not RDNA2 from the dev kits, just a distinct possibility

 

[WhispersintheWind]

He could be telling you GCN TFs and not RDNA2 from the dev kits, just a distinct possibility

The Studio this Dev works at is a Major Game Dev. With many awards under their belt. I think I’m safe to assume they have a recent Dev Kits and Debug Kits. Debug Kits are the actually PS5 console devs run test could and programs on. The PSV V shapes console is a debug kit.

 

[RaySoft]

My take: PS5 has lower TF number than XSX, but not by much.. around 11 maybe. Nevertheless.. Even though it ships with lower TF count, it makes up for it in smart architectural nuances i.e. the larger compute in PS4 APU compared to the norm at the time. Cerny is a smart man, he (as the opposition ofc) knows fairly well what's needed down the road. TF's isn't be-all here. I would rather have a SOC that's optimized towards the functions that are actually used by games instead of a raw TF benchmark monster. My bet is PS5 will perform better than XSX even tho' it calculates less TF's. Chew on that!

 

[Hatsuma]

At this point I feel I’ve heard every combination of teraflop numbers from each insider here. It’s like almost impossible for any of them to be wrong lol. It’s way too easy to be an insider nowadays

Well one thing the Insiders are consistent about is that it very close, above 9.2, and double digits. Variance lies in how high. Majority claim above 12. Odium and Heisenberg say it is slightly in Xbox favor, so 11 close to 12

 

[saintjules]

At this point I feel I’ve heard every combination of teraflop numbers from each insider here. It’s like almost impossible for any of them to be wrong lol. It’s way too easy to be an insider nowadays. Should of called myself Johnny smith and post range of numbers with always leaving an out for myself.

-Different dev kits
-Plans change
-What I was told
-One of my sources said this but another said that. Etc.

people will believe whatever they want to hear.

Yup. I just want to see the games, the design(s), and know the price at this point. I'm ready to drop the money on a pre-order asap

 

[Rusco Da Vino]

He could be telling you GCN TFs and not RDNA2 from the dev kits, just a distinct possibility

 

[WhispersintheWind]

For the Record. I can give my opinion what I think the PSV is. But it’s only Speculation. I have not factual data to verify my opinion.

My Opinion is based off what Devs have said with out going into details.

 

[semicool]

The Studio this Dev works at is a Major Game Dev. With many awards under their belt. I think I’m safe to assume they have a recent Dev Kits and Debug Kits. Debug Kits are the actually PS5 console devs run test could and programs on. The PSV V shapes console is a debug kit.

Just a point. If he truly is a Dev and an actual firsthand insider, he's breaking his NDA, which means he implicitly cannot be trusted on some level. Implied integrity issue. That's why it's inherently hard to trust "insiders", not to mention people straight up lying like TommyFisher. Unintended leaks on the other hand....would be classified as mistakes.

 

[WhispersintheWind]

Just a point. If he truly is a Dev and an actual firsthand insider, he's breaking his NDA, which means he implicitly cannot be trusted on some level. Implied integrity issue.

I agree and that is why I don’t think they have spoken in detail. I believe they all have been very loose and causal with that have said with out leaking details. I have no doubt they are being truthful though.

 

[SlimySnake]

For the Record. I can give my opinion what I think the PSV is. But it’s only Speculation. I have not factual data to verify my opinion.

My Opinion is based off what Devs have said with out going into details.

go on. we are listening.

 

[Rusco Da Vino]

I agree and that is why I don’t think they have spoken in detail. I believe they all have been very loose and causal with that have said with out leaking details. I have no doubt they are being truthful though.

Man, the trolls here are going to try to discredit you, your source and all what you are going to share with us from now on.

They are fucking desperate

 

[01011001]

For the Record. I can give my opinion what I think the PSV is. But it’s only Speculation. I have not factual data to verify my opinion.

My Opinion is based off what Devs have said with out going into details.

if what the devs tell you is so vague that you have to then form your own opinion and speculate on what this means... how the hell is this then at all helpful?

and how comes that these insiders happily tell people certain details but are seemingly always unable to give basic info about the hardware they have?
it's not like leaking target specs could be used to identify you any more than any other unreleased info could.

let's say some dev would tell you "it's currently targeting exactly 11.36TF"
how is that in any way more risky than talking to you in the first place?

 

[siekro1983]

Round and round we go. I just want sony to reveal specs, and that's it.

 

[01011001]

Man, the trolls here are going to try to discredit you, your source and all what all what are going to share from now on.

They are fucking desperate

desperate is automatically jumping on someone's ship that is not yet verified and also can not give any details and has to speculate on vague info they get from an unknown source.

☝that is desperate

 

[OrionNebula]

For the Record. I can give my opinion what I think the PSV is. But it’s only Speculation. I have not factual data to verify my opinion.

My Opinion is based off what Devs have said with out going into details.

Just a quick question, most likely unrelated but, I gotta ask, since you keep putting random letters in caps when you write

Are you Speaking in Code?

 

[Rusco Da Vino]

desperate is automatically jumping on someone's ship that is not yet verified and also can not give any details and has to speculate on vague info they get from an unknown source.

☝that is desperate

☝
Troll spotted?

 

[WhispersintheWind]

Just a quick question, most likely unrelated but, I gotta ask, since you keep putting random letters in caps when you write

Are you Speaking in Code?

No Not speaking in code. My New IPhone 11 Pro Max does it on its own. I have big hands also.

 

[01011001]

Troll spotted?

so you don't think it is desperate to automatically side with an unknown user that is unverified and says his info is vague and he has to share it through his own interpretation?

any arguments on how this is not desperate?
for real. we have people that go through pages and pages of data to find data leaks... but these data leaks are obviously not trustworthy... because... reasons.

then comes a guy noone knows, who says he knows someone who doesn't tell specifics... and you and many others instantly jump on that boat and go with it...

no sign of scepticism whatsoever... simply because you like the kind of info they share.

don't tell me that's not the case because it so obviously is.

 

[Playboi Carti]

For the Record. I can give my opinion what I think the PSV is. But it’s only Speculation. I have not factual data to verify my opinion.

My Opinion is based off what Devs have said with out going into details.

Go on

 

[WhispersintheWind]

if what the devs tell you is so vague that you have to then form your own opinion and speculate on what this means... how the hell is this then at all helpful?

and how comes that these insiders happily tell people certain details but are seemingly always unable to give basic info about the hardware they have?
it's not like leaking target specs could be used to identify you any more than any other unreleased info could.

let's say some dev would tell you "it's currently targeting exactly 11.36TF"
how is that in any way more risky than talking to you in the first place?

I believe they can get by with saying we are targeting 11TF perf. Versus saying PS5 Retail console is.....

 

[thicc_girls_are_teh_best]

With the way RDN2A2 is more efficient, and that double side chip cooling patent, perhaps Sony could squeeze out a higher TF number, comparable to the XSX by increasing the clock speed?

I guess they could, but then there's nothing technically stopping MS from doing the same. Same architecture, same node process, same high-end components etc. But there's likely a logistical reason they've settled on 12 (for now) and there's only so far Sony can push the chip before thermals are negatively affected.

So if the PS5 is around the 36/40CU size (just going from Oberon stuff, and some speculation. It might be bigger than that), they can shrink the TF delta but there's probably going to be a limit.

My take: PS5 has lower TF number than XSX, but not by much.. around 11 maybe. Nevertheless.. Even though it ships with lower TF count, it makes up for it in smart architectural nuances i.e. the larger compute in PS4 APU compared to the norm at the time. Cerny is a smart man, he (as the opposition ofc) knows fairly well what's needed down the road. TF's isn't be-all here. I would rather have a SOC that's optimized towards the functions that are actually used by games instead of a raw TF benchmark monster. My bet is PS5 will perform better than XSX even tho' it calculates less TF's. Chew on that!

What is this idea that MS is simply throwing brute force at the problem? They're working with the same architecture and technology Sony more or less is, and they have access to the same GPU features , etc. Yes there're some things they'll both have to do custom API implementations on to provide access to them for developers (variable rate shading, for example), but that should likely come to around similar overhead for both sides (or maybe one has notably lower overhead on some things and the other has notably lower overhead over others).

You're right, TF's isn't the end-all-be-all and it's nice to see people actually keeping that in mind. But there isn't any magical implementation of APIs or optimizations that will somehow result in PS5 running rings around XSX in terms of efficiency of resources used. Sony has a great lead in someone like Cerny, but it's not like he has no equivalents on the MS side of things (and I'd suspect both companies are pulling engineers and architects from relative divisions to their PS/Xbox ones for next-gen console development).

 

[Faithless83]

For the Record. I can give my opinion what I think the PSV is. But it’s only Speculation. I have not factual data to verify my opinion.

My Opinion is based off what Devs have said with out going into details.

Thank you! Anything to stop the github nonsense. These guys are worse than vegans.

 

[01011001]

I believe they can get by with saying we are targeting 11TF perf. Versus saying PS5 Retail console is.....

well of course, but it would be nice to have exact details from a verified source.
and by exact I mean actual numbers on clock rates and CUs for the GPU, maybe CPU clocks, ram capacity and clocks etc.

this shouldn't be hard to get for developers, they have dev kits they can literally look this up on the dev kit's UI usually... right?

 

[Matreios Raven]

This discussion: “You spin me round right baby right round like a record baby,

Round, round, round, round.“

 

[01011001]

This discussion: “You spin me round right baby right round like a record baby,

Round, round, round, round.“

 

[Hatsuma]

Well he stated he sent mod of war evidence. I'm curious if it is legit. And if so, how much weight does it have

 

[WhispersintheWind]

I cannot verify this as fact. Only my Speculation based of conversations. I will explain my reasoning and why I have come to the conclusions that I have reached. So here we go.

I speculate the PS5 to be 11.5TF-12TF. 13 is a possibility slightly in my Opinion I will explain why.

It is rumored and I haven’t heard anything to counter the rumor as the time of me typing this. Is that Xbox Series X has more GDDR6 Memory Bandwidth. It’s GPU as of now is 12TF.
It take a lot of bandwidth and memory to fed and hold data for a GPU that high and also the CPU. It is also believed the memory bus of the Xbox Series X is larger then the PS5. Because of that I don’t think Higher TFlop count would be need. Because you would starve the GPU.

I believe because the The SSD On PS5 is rumored to have a higher bandwidth of 1GB higher transfer rate over The Xbox Series X you wouldn’t need higher GPU with a huge TFLOP count. Because you have an advantage with SSD Transfer rate. You can render and tile resource or mega texture ( What ever name you want to name it) data to the SSD. Which would lower the grunt work of your GPU and CPU. So you can get by with a lower clocked CPU and GPU.

With a lower clocked CPU and GPU. You create less thermals. Less heat means you can go with a smaller form factor.

Also when I say Devs say they are targeting 11TF doesn’t mean the console is only that. There is usually some overhead available that they use later in the optimization stage that they can leverage and polish up a game.

So in my Opinion that translates to PSV being around 11.5-12TF with 13 having a slight chance.

Now for the record no Dev has told me the exact TF count of the projected retail PSV.

 

[Roronoa Zoro]

GitHub was not a hoax.
GitHub was not wrong.
GitHub has real information.

However, the GPU in GitHub is not the GPU in the PS5.

What makes you so confident about github?

 

[WhispersintheWind]

Well he stated he sent mod of war evidence. I'm curious if it is legit. And if so, how much weight does it have

I have legitimately sent Mod of War evidence. But if he isn’t on and another mod is. I willing to send evidence and proof to them if the PM conversation me.