Support NeoGAF

[Gavin Stevens]

Just a heads up but blur is my indie studio and no, we don’t have any kits, as I’ve said a few times before. I wouldn’t be dropping any hints at all if I did.

But I’ve worked with several people over the years, we chat, we game, we talk bollocks; friends. And the they have shared a few tidbits, nothing amazing but likewise nothing incredibly out there. Some of which has come true, some of which has yet to pass.

But no, myself and Blur are nothing to do with development on next gen machines. What I share I share as rumblings, not as insider knowledge. As I said, if I did have access to that sort of stuff myself personally you can bet your arse I wouldn’t be sharing it. That’s not very professional to share it in the public if I signed an nda or something.

I made a large post a few days ago with all I did know, and haven’t heard a thing since for a few weeks. To be fair, I was one of the first to suggest rdna might not be 2.0 but a sort of 1.9, bolted one version, and that’s looking highly likely. I said about the sound stuff being RT on SX too and a point in Sonys favour before they confirmed it as ray traced hardware, ages ago. And I’ve even said a while back about how the SSD gap isn’t as big as some people may be hoping for (1.0 to 1.8gb difference, likely lower end of that, which is again nothing much in the scheme of things when purely looking at these speeds).

but the biggest thing I’ve been told, was that the target specs on a near final revision for ps5 was 10.5-11.5, and they were having trouble hitting that upper limit due to heat being a concern. It was then reported after that, that Sony had invested in a higher cooking cost per bom, and I would bet my bottom dollar on both heat issues being present later. But yeah, the values I keep hearing there are the same, and those are, let’s be honest, totally realistic in all areas, with regard to comparing to the bom cost, the target that they are aiming for, and the fact that common sense would dictate that some things being said around here are highly unlikely without melting some marshmallows 9tf? No, just no, it was an older target they were working on ages back, year+, they had worked on and tested but then changed plans at some point hence why there is evidence for it but it won’t be the final design. 9tf is just as silly as 14 lol

I’m also standing by everything I’m saying and have been pretty consistent now for the last month, despite the specs dropping for the SX and tommy coming out with these quite frankly incredible figures.

But don’t hang on what I say, because while i do know a few first party devs for both sides who ive worked with, and a few third party ones, I could still be spun a yarn, as can anybody. So when I talk I talk from common sense, rather than just wanting one box to beat the other (which as ive said so many times now, would be pretty shit if it was by a big margin).

xxx

 

[jonnyp]

No, you need huge coolings for running a chip at 2 GHz. And that might have been the only way for Sony to get to 9 TF, aside from a completely new chip design, which would've set them back two years. The devkit design is strong evidence that that the 9.2 TF APU at 2 GHz is real.

Not that big of a cooling solution,. Nope, the devkit seems to suggest it's more than 9 TF. Sorry.

 

[Ralf]

what else do you expect from a hardcore Xbox fanboy? He's gonna project the lowest numbers.

what else do you expect from a hardcore Sony fanboy? He's gonna project the highest numbers.

 

[FranXico]

But don’t hang on what I say, because while i do know a few first party devs for both sides who ive worked with, and a few third party ones, I could still be spun a yarn, as can anybody. So when I talk I talk from common sense, rather than just wanting one box to beat the other (which as ive said so many times now, would be pretty shit if it was by a big margin).

xxx

Thanks.

 

[Burrito Bandito]

Yo it's so funny when you see a huge Xbox fanboy giving credibility to other Xbox fanboys (the most delusional ones like this one) and believe in what they say in their delusional private universe..... YET, guys like Kleegamefan (verified person that has been in the industry for more than 20 years) and guys like Jason Schreier saying both consoles are more powerful than RTX 2080, guys like that get shitted on because they said PS5 is as powerful or more powerful than XSX.

So funny

the funnier circle jerk is the one here on gaf.

some people are taking hard data as a starting point and others are taking a game of telephone. Klee has zero track record so he should be taken with a massive grain of salt. Schreier? perhaps but its still word of mouth.

they may very well be right but there is a lot of conflicting stuff out there and we will only find out when Sony finally releases some tidbits.

 

[darkinstinct]

but the biggest thing I’ve been told, was that the target specs on a near final revision for pa5 was 10.5-11.5,

See, where I think you were told wrong info (or to be precise: information out of context) is exactly this target spec. Why is it a range? A target spec is a target spec. It's not a range of target specs. You target 11.5 - and then maybe end up at 10.5 or 11 or 12. But giving a 10 % range seems highly unlikely. Do I build a game for that 11.5 TF? Do I build it for 10.5? And of course we have Klee that says the target specs for both consoles were always 12 TF. So if I believe your source, then Klee is lying. And some others as well. Unless: We aren't talking about the same things. Surely not every dev is treated the same way. First party gets dev kits first. Important third parties like EA follow. But you don't hand out those dev kits to every developer out there that wants to develop for PS5. For some, you will most certainly tell them a target spec that is a range - because that target spec is not the same architecture. There is no 1:1 translation from GCN Vega to RDNA. You could tell them a target range of 10.5 to 11.5 Vega. Which translates to around 9.2 TF RDNA. Curiously enough, the official efficiency gains of RDNA1 over GCN Vega as communicated by AMD are 25 %. 9.2 TF * 1.25 = 11.5. The upper limit of that range you were told, which is tough to reach because of heat issues. Which would certainly arise from a 2GHz 9.2 TF APU.

 

[darkinstinct]

Not that big of a cooling solution,, especially iIf it's RDNA 2 and 7nm EUV.

Nothing suggests that PS5 is on RDNA2 or 7N+. Absolutely nothing.

 

[Panajev2001a]

Nothing suggests that PS5 is on RDNA2 or 7N+. Absolutely nothing.

Compact-ish size of DevKit, Sony always picking latest and upcoming features from AMD’s roadmap (see PS4 and advanced async compute support and other customisations or Vega’s RPM/Double rate FP16 and other Polaris enhancements for PS4 Pro)?

 

[Panajev2001a]

See, where I think you were told wrong info (or to be precise: information out of context) is exactly this target spec. Why is it a range? A target spec is a target spec. It's not a range of target specs. You target 11.5 - and then maybe end up at 10.5 or 11 or 12. But giving a 10 % range seems highly unlikely. Do I build a game for that 11.5 TF? Do I build it for 10.5? And of course we have Klee that says the target specs for both consoles were always 12 TF. So if I believe your source, then Klee is lying. And some others as well. Unless: We aren't talking about the same things. Surely not every dev is treated the same way. First party gets dev kits first. Important third parties like EA follow. But you don't hand out those dev kits to every developer out there that wants to develop for PS5. For some, you will most certainly tell them a target spec that is a range - because that target spec is not the same architecture. There is no 1:1 translation from GCN Vega to RDNA. You could tell them a target range of 10.5 to 11.5 Vega. Which translates to around 9.2 TF RDNA. Curiously enough, the official efficiency gains of RDNA1 over GCN Vega as communicated by AMD are 25 %. 9.2 TF * 1.25 = 11.5. The upper limit of that range you were told, which is tough to reach because of heat issues. Which would certainly arise from a 2GHz 9.2 TF APU.

You are assuming to know the exact CU count as given and inflexible. Not sure why...

 

[FranXico]

See, where I think you were told wrong info (or to be precise: information out of context) is exactly this target spec. Why is it a range? A target spec is a target spec. It's not a range of target specs. You target 11.5 - and then maybe end up at 10.5 or 11 or 12. But giving a 10 % range seems highly unlikely. Do I build a game for that 11.5 TF? Do I build it for 10.5? And of course we have Klee that says the target specs for both consoles were always 12 TF. So if I believe your source, then Klee is lying. And some others as well. Unless: We aren't talking about the same things. Surely not every dev is treated the same way. First party gets dev kits first. Important third parties like EA follow. But you don't hand out those dev kits to every developer out there that wants to develop for PS5. For some, you will most certainly tell them a target spec that is a range - because that target spec is not the same architecture. There is no 1:1 translation from GCN Vega to RDNA. You could tell them a target range of 10.5 to 11.5 Vega. Which translates to around 9.2 TF RDNA. Curiously enough, the official efficiency gains of RDNA1 over GCN Vega as communicated by AMD are 25 %. 9.2 TF * 1.25 = 11.5. The upper limit of that range you were told, which is tough to reach because of heat issues. Which would certainly arise from a 2GHz 9.2 TF APU.

So, when MS PR spokespeople say 12TF, there must not be any ambiguity, it is RDNA TeraFlops, but first party developers who would surely know the difference would confuse another peer in their profession with GCN TeraFlops?

 

[Neo Blaster]

---
Sony: GAF, you thirsty?

GAF: Oh yeah.

Sony: How Thirsty?

GAF:

Switch that to Diet Coke, and Sony throw some Menthos. That will be GAF reaction when that happens.

 

[Panajev2001a]

So, when MS PR spokespeople say 12TF, there must not be any ambiguity, it is RDNA TeraFlops, but first party developers who would surely know the difference would confuse another peer in their profession with GCN TeraFlops?

Well, what is wrong with believing only what a company PR’s spokesperson tells you and interpret it only under the best possible light and best of any other assumptions ?

 

[Gavin Stevens]

Because a target range is just that, a target range. I don’t know why it was this range, but this is a first party dev. To be fair, I would expect a range as a target, and this target has been that way for a good way now, aiming for 11.5 as their upper target heat depending. That’s kinda how all dev kits work, you have different versions with a target, and as things progress and get more refined you get closer to that target. You don’t just make a first revision of a dev kit and go right to the target spec.

 

[FranXico]

Well, what is wrong with believing only what a company PR’s spokesperson tell you and interpret it only under the best possible light and best of any other assumptions ?

Indeed.

 

[Neo Blaster]

If Sony keeps delaying announcement only to reveal a much weaker PS5, this will be a bigger dumpster fire than fyre festival.

If it's the opposite, Sony would be like:

 

[darkinstinct]

Here, AMD at last year's E3.

PS5:

Custom 7nm AMD Ryzen Zen 2 and Radeon "Navi" Architectures -> no word of 7N+, the exact same opposite. No word of next-gen RDNA (also known as RDNA2)

Xbox Series X:

"The next amazing chapter in our partnership", "custom processor co-engineered between our team and the Microsoft team, with one goal in mind: To deliver the ultimate gaming experience.", "and it uses our Zen 2 core and our next-generation RDNA architecture"

Just hearing how different she talks about those two products shows that there is a huge gap between them. I am almost certain that AMD doesn't even have their own raytracing solution, they are sourcing it from Microsoft (there was a job offer for AMD two weeks ago that talked about working on future products using the Microsoft raytracing cores powering the next Xbox).

 

[darkinstinct]

So, when MS PR spokespeople say 12TF, there must not be any ambiguity, it is RDNA TeraFlops, but first party developers who would surely know the difference would confuse another peer in their profession with GCN TeraFlops?

MS gave out a press release stating 12 TF RDNA2 based GPU performance. There is no ambiguity. It is what it is.

 

[PocoJoe]

Nothing suggests that PS5 is on RDNA2 or 7N+. Absolutely nothing.

And what suggests it isnt? Your ass, github "leak" whom anyone can make?

 

[PocoJoe]

Here, AMD at last year's E3.

PS5:

Custom 7nm AMD Ryzen Zen 2 and Radeon "Navi" Architectures -> no word of 7N+, the exact same opposite. No word of next-gen RDNA (also known as RDNA2)

Xbox Series X:

"The next amazing chapter in our partnership", "custom processor co-engineered between our team and the Microsoft team, with one goal in mind: To deliver the ultimate gaming experience.", "and it uses our Zen 2 core and our next-generation RDNA architecture"

Just hearing how different she talks about those two products shows that there is a huge gap between them. I am almost certain that AMD doesn't even have their own raytracing solution, they are sourcing it from Microsoft (there was a job offer for AMD two weeks ago that talked about working on future products using the Microsoft raytracing cores powering the next Xbox).

Amds tech pdf from last autumn had next gen consoles(plural) under next gen rdna(aka 2).

No own RT solution. Dude you are delusional or troll

 

[Neo Blaster]

@6:57

Larry: "Wait a minute, wait a minute... Wa..wa.. Ray tracing audio, that's the first time I've heard that, that's interesting."

Spoiler

Why do everybody just ignores what Cerny says?

 

[Mendou]

Read statements from 2012/2013 about PS4, they sound exactly the same. And we got a low end mobile CPU and a medium mobile GPU for it. I am confused as to why people think that PS5 will be a beast based on what they know about PS4. The current gen consoles were underpowered as fuck. What we would've gotten if this was a typical console cycle is the Pro. But at launch, not three years later. It was a generation about doing away with subsidies. And Sony went and sold 105 million so far and probably 140 million when PS4 is going out of the market. Their strategy worked.

So why, just why do you assume that Sony would have any interest at all in going back to the PS3 generation, where they had huge subsidies for fresh hardware and sold 80 million? All this trust in Cerny BS is coming from the wrong assumption that PS4 is powerful. It's not. Even PS4 Pro was not, Sony didn't go for native 4K, they went for checkerboard rendering. They wanted the cheapest way to get to 4K. Again, why do you expect a company with this track record to completely change it for PS5 after the success they had with PS4?

The talk over the last years has always been that for next gen we will not see the huge TF wars anymore, everybody will talk about other things because it's hard to sell a constantly lowering increase gen over gen. Yet here we are and everybody thinks that it's all about TF. Sony doesn't seem to think so. That's why they talk about the controller and the SSD. Just as people predicted three years ago. The odd one out is Microsoft, who just talk about power. Why? Because they have to. They are not the market leader, the only way to regain market share is through power. Sony has a completely different background right now. They don't have to be powerful. They can just release PS5 and if they use checkerboard rendering to close a 30 % gap like they did with Pro and X, which had a 50 % gap (and people don't end telling that Pro mostly even surpasses X with framerate stability) that's totally ok. And not the end of the world.

The only verified information today is that AMD leak that was actually taken down by lawyers and which is 3/5 for now. And everything regarding Sony and how they approached the current gen points to that leak being the actual story. The important thing for success is not power, but price. All that talk from Sony about wanting an even faster adoption of PS5 than PS4 (which was the fastest adoption of any console ever) and here we are having people imagine a 12 TF machine from Sony. Which leads to one thing only, a higher price.

PS5 is a console built the same way as PS4. A medium mobile GPU but with a more current mobile Zen 2. For the same 399 price point. And unless Sony now wants to take a huge loss on hardware (if they sell it at a $100 loss they will lose 2 billion in the first year, based on 20 million consoles sold).

Microsoft had their most success with a 399 SKU. Sony must've expected them to target that price point again for the next generation. And built a console to compete with it. A 9.2 TF console (probably 8.6 retail).

We already have a confirmed 12tf RDNA 2 console from Microsoft. I was among the pessimists in believing it wasn't going to be that, but lo and behold we got a 12tf console for presumably $500-600 USD. No insider on here or outside of here has stated that the PS5 has 8-9tf of power. Have some optimism.

 

[LED Guy?]

the funnier circle jerk is the one here on gaf.

some people are taking hard data as a starting point and others are taking a game of telephone. Klee has zero track record so he should be taken with a massive grain of salt. Schreier? perhaps but its still word of mouth.

they may very well be right but there is a lot of conflicting stuff out there and we will only find out when Sony finally releases some tidbits.

Except Klee got verified twice by ZHuge Daniel Ahmed and came off as a credible source.

As for Jason, almost everything that guy reports become true, like almost everything.

 

[R600]

Flute benchmark has OPN number. That OPN number matches to OPN number found in Github repo in a file called :

Oberon_100-000000004-20_4_GPUID_ThermTripLimit_Feb-10

Inside this file there is PCI ID - 13F9 which matches Flute PCI ID (OPN + PCI ID).

Notice 13F9:OBR?

Check Flute benchmark OPN and PCI ID..

...

Look at RAM test results of.

16GB
528GB/s (16.5Gbps chips)
256bit bus

What is there left to say? It all fits perfectly. 528GB/s shared between CPU and GPU is not enough for 13TF chip, not even close, especially with RT thrown in there.

Results, hard data...its starring you in a face. "Insiders" like Klee have left. They are nowhere to be found after Github, and for a good reason.

 

[xool]

Important message from our leaders :

 

[Gavin Stevens]

I’ll say again that if you have 1-2tf difference between a console it all depends on the base range you are using.

1tf and 3tf consoles? That’s a HUGE difference. 11 and 13tf? That’s a much smaller difference.

 

[xool]

Flute benchmark has OPN number. That OPN number matches to OPN number found in Github repo in a file called :

[..snip..]

Inside this file there is PCI ID - 13F9 which matches Flute PCI ID (OPN + PCI ID).

[..snip...]

Notice 13F9:OBR?

How did I miss this first time.. Don't remember this at the time of the leak (bench is down isn't it) [hope you're not photoshopping these LOL]

I mean at the time I thought people weren't sure what it was .. and there it is in black and white

Latencies (154ns) are GDDR levels too, slower than DDR and HBM too..

Well. There it is. What the future hold, who really knows.

[edit -wasn't this the thing with the cut down L3 and possible Zen (not Zen2) levels of FP performance .. I think I thought it was a RavenRidge part .. ]

 

[Snake29]

Here, AMD at last year's E3.

PS5:

Custom 7nm AMD Ryzen Zen 2 and Radeon "Navi" Architectures -> no word of 7N+, the exact same opposite. No word of next-gen RDNA (also known as RDNA2)

Xbox Series X:

"The next amazing chapter in our partnership", "custom processor co-engineered between our team and the Microsoft team, with one goal in mind: To deliver the ultimate gaming experience.", "and it uses our Zen 2 core and our next-generation RDNA architecture"

Just hearing how different she talks about those two products shows that there is a huge gap between them. I am almost certain that AMD doesn't even have their own raytracing solution, they are sourcing it from Microsoft (there was a job offer for AMD two weeks ago that talked about working on future products using the Microsoft raytracing cores powering the next Xbox).

Are you the same dude on Tweakers.nl saying that the Xbox One X has dedicated RT cores? I had a stupid discussion with you.

I will translate it here for the people:

drakiesoft
@Recon24 - 29 February 2020 09:28
I never said full rt. But real rt for shadow. That's only possible with dedicated cores that has the onex and soon to a large extent the series x. Anyway, what Phil Spencer didn't tell. The series x also gets ai cores.
For direct ml and super res filtering gaming ai. Also for processing neural ai and Azure ai. Microsoft has been working on their Microsoft research program for a long time with the development of ai applications. Now that Xbox series x ism with AMD will soon also have ai cores at their disposal, can only lead to a unique game experience. Aren't you curious what the ps5 will do? It already has a ps5 logo and fast SSD, so booting shouldn't be a problem .

Edit: Well i see you are from Switzerland, but maybe it's your twin brother, because you both are reading to much into those marketing talk.

 

[Panajev2001a]

Flute benchmark has OPN number. That OPN number matches to OPN number found in Github repo in a file called :


Inside this file there is PCI ID - 13F9 which matches Flute PCI ID (OPN + PCI ID).

Notice 13F9:OBR?

Check Flute benchmark OPN and PCI ID..

...

Look at RAM test results of.

16GB
528GB/s (16.5Gbps chips)
256bit bus

What is there left to say? It all fits perfectly. 528GB/s shared between CPU and GPU is not enough for 13TF chip, not even close, especially with RT thrown in there.

Results, hard data...its starring you in a face. "Insiders" like Klee have left. They are nowhere to be found after Github, and for a good reason.

It is coherent with itself and seems decently assembled set of data (the GitHub leak), but it still has nothing that really ties it to PS5 still.

The thought seems to be that since it cannot be proven it is not PS5 it must be PS5... .

 

[R600]

How did I miss this first time.. Don't remember this at the time of the leak (bench is down isn't it) [hope you're not photoshopping these LOL]

I mean at the time I thought people weren't sure what it was .. and there it is in black and white

Latencies (154ns) are GDDR levels too, slower than DDR and HBM too..

Well. There it is. What the future hold, who really knows.

13F9 benchmark on userbenchmark was unearthed few months ago, was not part of same Flute leak. We knew 13F9 is PS5 related, we just didnt have it black on white its Oberon.

No it was not RavenRidge. Its Zen2 with cut L3 cache (which laptop based Zen2s will have as well). It saves ton of space for little performance loss duo to being on APU. Flute was removed the same day it leaked. Look at socket BL5 (not PC/laptop therefore).

It is coherent in itself, but this does not tie it to PS5 still. The thought seems to be that since it cannot be proven it is not PS5 it must be PS5... .

But it is PS5. You had literal commit on Github Oberon data where it mentions SIE. And you have Flute, Oberon and Prospero all being characters from Shakespear...

 

[tflopsmatter]

13F9 benchmark on userbenchmark was unearthed few months ago, was not part of same Flute leak. We knew 13F9 is PS5 related, we just didnt have it black on white its Oberon.

The Github leak has NOTHING to do with the PS5. Sony wouldn't launch a next gen console with less than 12TF, period. Obviously, the PS5 chip simply didn't leak out.

 

[xool]

13F9 benchmark on userbenchmark was unearthed few months ago, was not part of same Flute leak. We knew 13F9 is PS5 related, we just didnt have it black on white its Oberon.

Interesting - so this thing is still fairly current, not from early 2019 or whatever. (?)

tbh I want a PS5 better than this, small L3, mid-range bandwidth, not as good as the competition .. getting OG XboxOne vibes from all this ..

All they need to do is bundle it with compulsory PS-VR and we're done /jk

 

[it_wasn't_me]

the funnier circle jerk is the one here on gaf.

some people are taking hard data as a starting point and others are taking a game of telephone. Klee has zero track record so he should be taken with a massive grain of salt. Schreier? perhaps but its still word of mouth.

they may very well be right but there is a lot of conflicting stuff out there and we will only find out when Sony finally releases some tidbits.

"I am Kelly Rickards, former editor from Die Hard Gamefan magazine(yes, I was there during the "Jap Bastard" incident), former west coast editor of EGM. I was on the GAF forums since 1998 and joined era during the mass exodus here. Got married, got out of the gaming industry in '04 but still maintain lots of contacts in the industry. "

Next-gen PS5 and next Xbox speculation launch thread |OT6| - Mostly Fan Noise and Hot Air OT

I'm not joking. I still remember the lengthy debates. With one person especially... I think the name was I-Low or something like that. Dude legit thought Sony was going to launch the PS4 that was just a few Gigaflops more powerful and maybe 1-2GB or RAM. I'd have to search through the old place...

www.resetera.com

So, you joined GAF 20 years later. What's your track record?

 

[01011001]

The Github leak has NOTHING to do with the PS5. Sony wouldn't launch a next gen console with less than 12TF, period. Obviously, the PS5 chip simply didn't leak out.

man will you be disappointed soon

 

[Panajev2001a]

13F9 benchmark on userbenchmark was unearthed few months ago, was not part of same Flute leak. We knew 13F9 is PS5 related, we just didnt have it black on white its Oberon.

No it was not RavenRidge. Its Zen2 with cut L3 cache (which laptop based Zen2s will have as well). It saves ton of space for little performance loss duo to being on APU. Flute was removed the same day it leaked. Look at socket BL5 (not PC/laptop therefore).


But it is PS5. You had literal commit on Github Oberon data where it mentions SIE. And you have Flute, Oberon and Prospero all being characters from Shakespear...

Those bits seem to be tied together, but the tie to PS5 is the flimsiest yet the one you want to hold on to the most .

 

[R600]

Interesting - so this thing is still fairly current, not from early 2019 or whatever. (?)

tbh I want a PS5 better than this, small L3, mid-range bandwidth, not as good as the competition .. getting OG XboxOne vibes from all this ..

All they need to do is bundle it with compulsory PS-VR and we're done /jk

Its from July (Flute), but V type dev kits are also from summer so they must have had APU ready (at least in small samples) by then.

Even AquariusZi (well known Taiwanese leaker) confirmed Oberon is PS5 chip and went through another revision in November (and we know it to be true because Komachi unearthed Oberon E0 in January).

My guess they are fixing and optimizing it because such high clocks are hard to pull of in consoles.

 

[Panajev2001a]

man will you be disappointed soon

Whishful thinking or insider knowledge ?

 

[Gavin Stevens]

The Github leak has NOTHING to do with the PS5. Sony wouldn't launch a next gen console with less than 12TF, period. Obviously, the PS5 chip simply didn't leak out.

An 11.5tf PS5 machine with a faster HDD, possible better memory and RT would bridge them gap in performance and be on par with a 12tf machine.

So yes, they could. Tf are part of the picture, not the whole picture.

 

[01011001]

Whishful thinking or insider knowledge ?

neither, just being objective.

 

[deccangamer]

Hey Gavin Stevens does your indie studio make only playstation exclusive games ?

 

[tflopsmatter]

I’ll say again that if you have 1-2tf difference between a console it all depends on the base range you are using.

1tf and 3tf consoles? That’s a HUGE difference. 11 and 13tf? That’s a much smaller difference.

What really matters is if the new console fits the description of a generational leap. If you multiply 1.84 times 7, you get 12.88TF. Sony would NOT launch with less than this due to the fact that a generational leap is defined as a minimal 7-10 time improvement in TFLOPs. If Sony couldn't produce a system with at least close to seven times the TFLOPS as the PS4, they would have simply waited another year. For Sony to utilize the chip from the Github leak would mean they would be utilizing an 8TF chip because a sustained 2000mhz frequency isn't going to be used. That's simply ridiculous. Instead of pushing a cheap where it's operating outside of it's most efficient zone, they would add CUs. To say otherwise is to call the engineers at Sony complete blathering idiots.

 

[jon3sy9]

What really matters is if the new console fits the description of a generational leap. If you multiply 1.84 times 7, you get 12.88TF. Sony would NOT launch with less than this due to the fact that a generational leap is defined as a minimal 7-10 time improvement in TFLOPs. If Sony couldn't produce a system with at least close to seven times the TFLOPS as the PS4, they would have simply waited another year. For Sony to utilize the chip from the Github leak would mean they would be utilizing an 8TF chip because a sustained 2000mhz frequency isn't going to be used. That's simply ridiculous. Instead of pushing a cheap where it's operating outside of it's most efficient zone, they would add CUs. To say otherwise is to call the engineers at Sony complete blathering idiots.

wow i cant wait for those 84-96tf machines next gen if thats the case

 

[Panajev2001a]

13F9 benchmark on userbenchmark was unearthed few months ago, was not part of same Flute leak. We knew 13F9 is PS5 related, we just didnt have it black on white its Oberon.

No it was not RavenRidge. Its Zen2 with cut L3 cache (which laptop based Zen2s will have as well). It saves ton of space for little performance loss duo to being on APU. Flute was removed the same day it leaked. Look at socket BL5 (not PC/laptop therefore).


But it is PS5. You had literal commit on Github Oberon data where it mentions SIE. And you have Flute, Oberon and Prospero all being characters from Shakespear...

You are selling it as fact, but you know you have an unverified leak with a three characters string in there and have built a giant castle upon it. This could be a set of APU revisions, it could be taking some semi-custom enhancements Sony has contributed (think PS4’s GPU semi custom enhancements) or yes it could be tied to PS5. You are selling it as proof on the basis that the reference to SIE is there and the data on the GPU are not fanboy badly written incoherent crap.

 

[01011001]

What really matters is if the new console fits the description of a generational leap. If you multiply 1.84 times 7, you get 12.88TF. Sony would NOT launch with less than this due to the fact that a generational leap is defined as a minimal 7-10 time improvement in TFLOPs. If Sony couldn't produce a system with at least close to seven times the TFLOPS as the PS4, they would have simply waited another year. For Sony to utilize the chip from the Github leak would mean they would be utilizing an 8TF chip because a sustained 2000mhz frequency isn't going to be used. That's simply ridiculous. Instead of pushing a cheap where it's operating outside of it's most efficient zone, they would add CUs. To say otherwise is to call the engineers at Sony complete blathering idiots.

so what you are saying is that sony never released the PS4... because of course they wouldn't, because it wasn't 7-10 times as powerful as the PS3...

the Xbox One also never released because of the same reason.

also who the hell defines this? and by what metric?

the CPU performance alone should be close to that "generational" leap you mention.
if you take both the GPU and CPU enhancements we are well above 10x as powerful even on a 9.2TF system

 

[tflopsmatter]

We already have a confirmed 12tf RDNA 2 console from Microsoft. I was among the pessimists in believing it wasn't going to be that, but lo and behold we got a 12tf console for presumably $500-600 USD. No insider on here or outside of here has stated that the PS5 has 8-9tf of power. Have some optimism.

We already have a confirmed 12tf RDNA 2 console from Microsoft. I was among the pessimists in believing it wasn't going to be that, but lo and behold we got a 12tf console for presumably $500-600 USD. No insider on here or outside of here has stated that the PS5 has 8-9tf of power. Have some optimism.

It's not about optimism but logic. Sony wouldn't launch the PS5 unless it was a true generational leap which would be approximately seven to ten times the TFLOPS of the PS4. If they couldn't launch such a system, they would wait until a better process node was able to be utilized. Moreover, the whole Github leak is meaningless. Sony would never clock the GPU in the PS5 at 2000mhz. So if that chip is the PS5 chip, it means that it would actually be an 8TF chip. This would not even be twice the competition of the XBox One X. In fact, it would only be 1.25 times the Tflops of the X Box One X.

The people who are pushing for the PS5 to be a gimped console are simply irrational.

 

[Gavin Stevens]

Hey Gavin Stevens does your indie studio make only playstation exclusive games ?

No, pc at the moment, and I want to stress again we have no direct access to any dev kit, and if I did first hand I wouldn’t be posting about it.

 

[Darklor01]

Flute benchmark has OPN number. That OPN number matches to OPN number found in Github repo in a file called :


Inside this file there is PCI ID - 13F9 which matches Flute PCI ID (OPN + PCI ID).

Notice 13F9:OBR?

Check Flute benchmark OPN and PCI ID..

...

Look at RAM test results of.

16GB
528GB/s (16.5Gbps chips)
256bit bus

What is there left to say? It all fits perfectly. 528GB/s shared between CPU and GPU is not enough for 13TF chip, not even close, especially with RT thrown in there.

Results, hard data...its starring you in a face. "Insiders" like Klee have left. They are nowhere to be found after Github, and for a good reason.

Hasn’t it been stated that Flute is not PS5?

 

[Panajev2001a]

neither, just being objective.

Ok, so where is the concrete evidence that makes your conclusion objective and not a subjective or emotionally derived one?

 

[Kusarigama]

The obsession of these XBox fanboys to see PS5 weaker than XSX is pathetic. They get an awesome box in Series X yet they only seem to care if it is more "powerful" than PS5. What's more pathetic is that XBox reps fuel these toxic party poopers.

 

[tflopsmatter]

You are selling it as fact, but you know you have an unverified leak with a three characters string in there and have built a giant castle upon it. This could be a set of APU revisions, it could be taking some semi-custom enhancements Sony has contributed (think PS4’s GPU semi custom enhancements) or yes it could be tied to PS5. You are selling it as proof on the basis that the reference to SIE is there and the data on the GPU are not fanboy badly written incoherent crap.

The Github GPU has nothing to do with the PS5. Obviously, the GPU in the PS5 didn't leak. I don't know why everyone assumes that the GPU of the PS5 has to be in that leak.

 

[Gamernyc78]

Just as the sun sets and rises, we are back full circle to the same regurgitated talking points lol

Good morning my fellow gaf brothers, you are all my nerd brethren whether rooting for green or blue