What's your view of Next-Gen Rumor Religions? +Poll! *excommunicado thread*

[PSman1700]

98% of 2.23GHz is 2.185GHz, so 10.1 TFLOPs.

The github option clearly says "around 9.2 TFLOPs".
With a suggested delta of 2% for "worst case game" meaning between 10.07 and 10.28 TFLOPs, the winning option is number 3:

Your taking that 2%, but that's never clarified as in that number. Between 9 and 10TF is what you will get in heavy workloads.
If it would be 2%, they wouldnt have needed to go this route.

 

[AlNom]

You might see top end GPU clock and reduced CPU clock when running BC. Otherwise, we'll have to see how Navi performs in the latest production runs to gauge how much power it sucks at 2.2GHz.

 

[BRiT]

This lol-dynamic-clocks is always why I was more keen on CU counts being the measurable quantity to use instead of TFs.

 

[BRiT]

This took way too damn long for all those pretenders to be outed. So how many of them are scrambling with their "things changed"?

 

[iroboto]

so were there any bets on RDNA 1.9 ?

 

[Proelite]

I wonder if Xsx can get a ~10% upclock too hmm. Still amazed at the late clock adjustment. That's better than Xb1's 6.6%.

 

[Silent_Buddha]

So, wait, I was right to be skeptical of all the insiders? Did any of them get the "leaks" correct? I mean, someone must have gotten it correct just by throwing darts at a dart board right?

Also, right to be skeptical of GitHub, while the information there was likely correct "at the time of testing" things can change and/or data is incomplete.

Regards,
SB

 

[BRiT]

I wonder if Xsx can get a 10% upclock too hmm.

Like one of those "and another thing..." moments?

 

[Deleted member 13524]

New This took way too damn long for all those pretenders to be outed. So how many of them are scrambling with their "things changed"?

You mean the ones claiming a 2GHz GPU was impossible since the almighty gospel was just testing the limits of the silicon and we'd be looking at a 8 TF PS5 in the end?
Like the one who made a $100 000 bet on specs that ended up false and is probably not paying that amount in the end?

There's a bunch of those on B3D.

There's no need to dogpile on people who don't post here and can't defend or explain themselves.
Firstly because it's in rather poor character to do so (and I thought that practice had been contained to one stinky thread, silly me).
Secondly, because you don't need to go beyond this forum to find these fake insiders.



Suggestion: why don't you create a thread dedicated to dogpiling on the people who didn't hit on the specs that you personally find important (which seems like kind of a moving target, but whatever)?
That way all this bullshit doesn't have to be repeated throughout several threads.

 

[BRiT]

Suggestion: why don't you create a thread dedicated to dogpiling on the people who didn't hit on the specs

Isn't that this thread entirely?

 

[Esrever]

Isn't that this thread entirely?

who we dogpiling first?

 

[Proelite]

I remember being dogged on, along with others, for believing in the relevancy of Github.

 

[DavidGraham]

New so were there any bets on RDNA 1.9 ?

If PS5 truly lacks VRS and ML, then some people might consider it RDNA 1.9 indeed.

 

[BRiT]

That high? Surely VRS and ML amount to more than just 0.1 ...

 

[HBRU]

The GitHub Gospel followers win. 36 CUs at 2.2GHz Boost clock will be around 2 GHz more realistically most of the time. The tested hardware, 36 CUs at 2 GHz, was PS5's SoC. Klee, OsirisBlack, etc.were false religions.

Well I even said 10.2 TF since one year or even more (just for marketing purposes)....

That 2ghz mark won't even be hit under most realistic workloads.

Edit: PS5 was going to be double the PS4 Pro, no idea what happened afterwards, but guess they had to upclock that thing alot to even come close to 10TF. Now we get 'max turbo clocks' just to be able to claim 10TF. Most of the time it will be abit over 9TF.

Yes. Sure 9.2 was the initial target but seeing what MS did they decided to upclock as much as possible just to reach that 10.2 number. I assure everyone I don't work for Sony.... [emoji28]... .

 

[AbsoluteBeginner]

You mean the ones claiming a 2GHz GPU was impossible since the almighty gospel was just testing the limits of the silicon and we'd be looking at a 8 TF PS5 in the end?
Like the one who made a $100 000 bet on specs that ended up false and is probably not paying that amount in the end?

There's a bunch of those on B3D.

There's no need to dogpile on people who don't post here and can't defend or explain themselves.
Firstly because it's in rather poor character to do so (and I thought that practice had been contained to one stinky thread, silly me).
Secondly, because you don't need to go beyond this forum to find these fake insiders.



Suggestion: why don't you create a thread dedicated to dogpiling on the people who didn't hit on the specs that you personally find important (which seems like kind of a moving target, but whatever)?
That way all this bullshit doesn't have to be repeated throughout several threads.

Amm...that would be you, no? Oberon_Native_results seem pretty indicative that AMD and Sony thought its possible. Gospel clearly said :

• Oberon codename
• 256 bit bus 14Gbps (although alot of us speculated on higher clocks for memory - config .ini file specified 1750MHz GDDR6)
• 36CU @2.0GHz
• 64ROPs

So Sony clocked above this, and all these chip revisions now make sense, with AMD's Smart Shifts they were able to push clocks even higher, but full chip was there - in Github.

In comparison to alot of you guys who were sharing role playing insider posts from pastebin and GAF, and saying how "Cerny is not idiot, he would never go narrow and fast like Github says". Guess what, he went even faster.

I like the machine, I think its superb in design but ain't no way Github was wrong. You can say X360 diagram that leaked in 2003 was wrong because it had clocks at 3.5GHz instead of 3.2GHz, when it painstakingly described entire design of system.

 

[chris1515]

I was in no religion part.

 

[Proelite]

I am just shocked that Github proved to be more accurate than 4chan, Reddit, OsirisBlack and Tommy Fisher.

 

[PSman1700]

Well I even said 10.2 TF since one year or even more (just for marketing purposes)....

Yes. Sure 9.2 was the initial target but seeing what MS did they decided to upclock as much as possible just to reach that 10.2 number. I assure everyone I don't work for Sony.... [emoji28]... .

Intresting part now will be what the sustained clocks are, i assume we will learn more about that later on.

 

[Shifty Geezer]

This is clarified by a quote from Mark Cerny:

98% of 2.23GHz is 2.185GHz, so 10.1 TFLOPs.

The github option clearly says "around 9.2 TFLOPs".

In that, you are right. This poll picked TFs as the metric. However ,that's somewhat skewed by dynamic clocks.

The main take home was a leaked spec for an SOC of 36 CUs, and a bunch of would-be insiders making claims of notably faster, and whether people believed the leak was final hardware (regardless what it could be clocked at) or whether the insiders were actual insiders providing real info. The insiders were, AFAICS, by and large, a bunch of fibbers. So the lesson here is don't trust GAF/Era insiders.

Although if one of the would-be insiders like Osiris did state 10.x TFs and stuff that makes it look legitimate, that could be recorded for future prosperity come PS6 predictions. But the other poll options have OsirisBlack saying 12 TF and Klee saying 12 to 13. None of the prophets said 10 TF, so the 'data' versus 'insiders' argument is completely in the data's favour.