Matt at Era is also in denial that GitHub got everything right
Just because Sony made a last minute over clock on the GPU doesn't mean that the GitHub 2Ghz leaks from June 2019 were wrong. Yet Matt keeps insisting that GitHub was wrong.
Dudes in straight denial.
Matt basically controlled the conversation about the PS5 leaks that were actually the real specs.
Matt didn't want people using Github "as evidence of anything." Yet here we are today, by far Github, the Verge and Windows Central were the most accurate sources of information.
I knew something was off with Matt when he immediately said "disregard it" in relation to the Github leaks. I mean, to not even consider it as a possibility seemed odd to me, the wording of how he mentioned something with actual data to it (whether it was relevant to the moment he made the comment or not) seemed almost standoff-ish, like its mere presence irritated him beyond the possibility of it just being another fake rumor leak.
Combined that with how quickly that leak was removed and the possibility that it would've likely been Sony who were behind the removal (MS would have no reason since it was both right about XSX (in insinuating a 12 TF system) plus they let the media run wild with 12 TF speculation post-TGAs that they could've stopped right then and there if it wasn't going to hit that number; AMD would have no incentive to damage control out of preference for one major client over the other over something as petty as console warring on gaming forums) and I had a good feeling there was at least some truth to it.
The question, of course, was always the pertinence of the leak and the testing data, but when we kept seeing the same PS5 chips being tested over a consistent timeline over and over, info on the testing going from early 2019 (Ariel) to June 2019 (first Oberon chip) to December 2019 and later (Oberon E0 revision, which might've just been the 2nd revision :S)...I don't see why anyone would have wanted to deny that level of evidence. Not to mention, testing data for other AMD chips was starting to be verified through the existence of actual cards and chips matching a lot of that same testing data.
If I had to venture a guess, since we never got updated clockspeeds on Oberon E0 (C0?), then chances are that is the revision where Sony made the clock adjustment to 2.23 GHz. Obviously, they also increased the bandwidth for that one to 512 GB/s, but since the final system has a lower bandwidth I'm guessing they had to make the choice between that 2.23 GHz clock or higher bandwidth while keeping performance stable, and they went with the higher GPU clock. And that's basically the Oberon that we now know is the PS5.
Like some of you all have already said, Heisenberg was closest to PS5's actual spec. I think O'dium's number could've been from some interim dev kit or maybe just an inaccurate source. Tommy's PS5 specs was a troll but they did actually get the active CU count for XSX right and pretty much the same with the clock, so he might've been a MS guy or someone connected to them (some people've even said it was TimDog. Hell maybe it could've been). Schreir never had any solid idea on what the systems were, he just rode on what others were saying.
As for Klee and OsirisBlack, I don't know what turned up with their sources. Maybe those sources had XSX target specs and threw in bogus PS5 specs before handing them off to them, can't say. WhispersInTheWind...I actually can't recall what their PS5 speculation was atm. So that said, even though most of the insiders were completely off-base with PS5's specs, there's also a reason I was careful to still consider things they brought up, because all of them were right about PS5's SSD, we can't deny that part. Some were also right in predicting certain event announcements, IIRC Klee mentioned something about XSX at the TGAs a couple days before it happened? Well, it happened.
So I give credit where it's due. That all said, there's a lot of people out there (they know who they are) who need to give Github and the testing data, and people like Komachi, Rogame, Cartman (no not South Park Cartman, the data mining Cartman guy on Twitter) etc. their props. Data is data, and those guys simply went searching around for it. I never suspected them altering the data to make Sony and PS5 look bad, I never considered Github compromised because Microsoft owns some portion of it, etc. Those were just completely baseless and idiotic concern trolling from console warriors who were dead-set on next-gen being a complete blowout for their favorite brand before the game even started. They were never interested in actual parity between the consoles and the only way they could be satisfied with XSX was if it were absolutely weaker than PS5 in every area and took itself out of the running before the systems even launched.
And now, since that fantasy has died, some of them who were only thinking about teraflops literally days ago are now trying to turn SSDs into miracle sauce that'll wipe away every perceived technical disadvantage, and usher in the greatest game-changing design shift while conveniently ignoring the system with greater resources for an actual paradigm programming shift, GPGPU. They're trying to spread FUD about one system being an elegant and highly optimized piece of kit while the other is seemingly some brutish hulking monstrosity with no elegance to its design. They're trying to spread FUD about XSX being held back by the XBO, as if cross-gen has never been a thing for the first year of next-gen, and as if MS would invest billions into next-gen R&D and manufacturing to simply not target their next-gen high end offering to justify the tech inside the machine.
I'm not saying that's only happening here, or that it's only out of intent for being disingenuous. It's happening at lots of spots and at least some of it comes down to people who are simply excited for next-gen but may not truly understand what certain tech is really capable of. In those instances it's forgivable. But if over the course of the next few weeks I guess we will see how truly fair and balanced the gaming media shows itself to be. There's zero problem in clearing up misconceptions regarding PS5 or talking about its strengths; that's needed TBH. But if you see a pattern of them doing that for only PS5 and failing (or refusing) to do the same for XSX, that should give some pause.
Also yeah; I get that a lot of people are probably mad with insiders for whatever reason, but honestly I don't feel upset or anything in that regard. Things just played out the way they did. The insiders were just messengers at the end of the day, and like I said before they did get some things correct even if they were wrong on other stuff. At the end of the day they're just interested in and excited for next-gen like the rest of us, so it should all be in good fun. Other people who clung to rumors and speculation to the point of antagonizing those who gave any of the hard data from Github or the testing info, though, aught to reconsider their quickness for dismissal of such on the next go-round.
I can tell you how I got things wrong and I am still around to take whatever people need to do to me in the process for missing things.
I had mentioned several times I had source saying that 12 number was accurate for the XsX. They would also tell me XsX was more powerful then the PS5 but they were super close.
I then brought up this crazy birdie friend of mine at SIE who told me in early Jan there would be no Feb reveal when everyone was expecting that. He also told me the PS5 was 10.5 but also that both systems were super close but the XsX was slightly more powerful.
I didn't think both people had that system right just were off on the other system on how close they were.
The one thing that both of my people said were the XsX was more powerful but the PS5 had the faster SSD.
I am without a doubt in the wrong for saying the PS5 was within 10% of the XsX.
It's cool dude; turns out in the end you were closest to guessing PS5's specs, but I can understand why you may've been reluctant to stick with that 10.5 at the time, given the environment in the thread. FWIW I think that might've also been part of the reason o'dium went with a higher number on PS5 while still insisting it was a bit weaker. Might've had a feeling the actual number was lower but seeing the attacks some people in that thread (and on Twitter) put them through a bit later, I can only imagine what'd of happened if they speculated a lower number.
But one thing I think people should give you and other insiders credit for is the PS5 SSD stuff. That was on-the-money speculation, even if not particularly exact. We started getting some speculation on actual SSD numbers when the Phison memory controller leaks came out, tho.
And reading through the thread some, just made me remember the bullshit R600 and VFXVeteran were put through in that next-gen thread, too. TBH R600 was one of the ones who helped get some early understanding on the nature of the testing data within the Github leak and the data mines. They gave clarification on the block structuring for enabling certain functions of the chip on/off for example. Some of my other understanding on that data such as the Ariel iGPU testing profiles, I had to search out lurking Beyond3D for, I can only imagine the shit some of those posters would've gotten in the next-gen thread if they posted there during the hot and bothersome periods xD.
At the end of the day for myself tho, I can't feel any ill will towards you or the other insiders. Even the shit I gave Tommy Fischer I apologized for after the XSX exact GPU specs came out, tho I can understand if PS diehards are still mad with them maybe. They shouldn't be IMHO, but I can understand why. I don't think anyone should be mad with the likes of yourself or guys like O'dium, BGs, or even Osiris or Klee, either. Same with Schreir, Matt etc.
That might be in big part tho because I always took insider speculation and rumors with a pinch of salt. Did the same with the Github leak and testing data, too, that's why I didn't turn away any of the big sources but tried speculating something in the middle. That's a big reason why I felt like sticking with 10.4 - 11.05 after that ITMedia article came out. The higher end of that just being optimistic if Oberon E0/C0 or another Oberon revision happened to be a bigger chip.
Eh, I'm just probably rambling now lol.