Gigabyte claims 12900k 8GHz OC record; CPU-Z, Der8auer disputes claim

[OneOfYas]

9 hours ago, CarlBar said:

 

That fact no one else has been able to get within more than around 400mhz of the claimed overclock however makes it exceedingly unlikely in the extreme. My guess is HiCookie himself just hadn't heard of the bug when he posted the results and Gigabyte then ran with it without waiting for the kind of feedback and intercommunication that goes on with somthing like this amongst the OC community.

And is not evidence.

[Dracarris]

7 hours ago, HenrySalayne said:

It can be easily blamed on CPU-Z and is not at all and never has been the fault of Gigabyte.

It can easily be blamed on blatantly missing common sense and a complete lack of even sanity checking the power drawn from the wall that would accompany such clocks.

 

It's an issue in CPU-Z that has happened before where not the actually effective FLL multiplicator is read back. Any serious overclocker monitors the power draw on the 12V lines to the CPU (like Roman) and immediately recognizes that sth doesn't compute.

 

And Gigabyte did exactly this in the past, not the first time exactly this BS happens on their side. So yes, 100% their fault, not really debatable at all.

[Beerzerker]

Guys, being that I'm a competitor at the bot with my own 8GHz+ results I already know what goes at the bot so I'm speaking from direct experience. Been doing this for well over a decade so I at least have a good idea of what's up.

Just know there are a few things in play here.

First:
Being that Hicookie is a long time competitor at the bot, he knows the rules so I seriously doubt there was intention to fake anything considering the consequenses for such. If there was and if proven as such, he'll pay the price for it, esp when you consider Der8auer has already called him out on this.
I'm not going to say or accuse him or anyone else of cheating, that's really not my place to do except in the case of it being blatantly obvious - And that's what the report button for questionable subs is for over there.
It's the staff's job to say if it's really valid or not and letting them take care of it.

Second:
This is a brand new architecture so any defects/bugs and so on (Errata) is not uncommon to see and can affect things - Remember the errata with first gen AM2/AM2+ Phenom chips?
These issues take time to get sorted and now that it's come to light Intel is working to get these problems fixed - Or at least they should be.

Third:
It's possible, since this is a new architecture the board maker(s) have some work to do for getting things right too in the BIOS to properly monitor and report things.
Most of that shoudn't be much of a problem since it all works in the same basic way but bear in mind this is an entirely new thing with a setup unlike anything that's been seen before from either AMD or Intel, that undoubtedly being a factor with it regardless.

Fourth:
It's really not Intel's fault Gigabyte chose to start crowing about it when they saw it before confirming things. That fault is squarely on Gigabyte because they "Jumped the gun" before making sure it was indeed 100% legit.
The marketing/sales guys most likely went nuts when they saw it and it all went from there. Have to admit there may or may not have been some nudging from Intel but the ultimate decision to do it was still Gigabyte's, so there it is.

Give it time guys, the real stuff about it will come to light eventually.

[RejZoR]

Whoever buys processor based on entirely unrealistic extreme overclocks that absolutely don't apply to day to day use is an idiot. Starting with the fact you pour liters of liquid nitrogen into the pot for just 2 hours of such use, a lot of extreme overclockers turn off cores, turn off HT and do all sorts of things just for the sake of higher number. All you need to do is validate it. Which means it has to do certain load without crashing. Unless required by the guidelines, all of the above is allowed. I still remember how back in the day quad cores had just 1 core functional and others disabled which helped reaching much higher clocks. Hell, people have done that to achieve higher clocks on water too. But it's entirely useless because it was just 1 core then. Imagine using 12900K with just 1 functioning core. It would be 8GHz with LN, but it would be plain useless.

[leadeater]

17 hours ago, OneOfYas said:

And is not evidence.

The existence of a known bug is however, so unless it can be repeated along with actual proof of real operating clocks this isn't really going to go anywhere. If it can be done once it can be done again so basically HiCookie needs to do it again or it didn't happen.

[OneOfYas]

6 hours ago, leadeater said:

The existence of a known bug is however, so unless it can be repeated along with actual proof of real operating clocks this isn't really going to go anywhere. If it can be done once it can be done again so basically HiCookie needs to do it again or it didn't happen.

There is no evidence, yet this community would rather tarnish a company based on logical fallacy....

 

[HenrySalayne]

8 minutes ago, OneOfYas said:

There is no evidence, yet this community would rather tarnish a company based on logical fallacy....

We know:

- everbody except Gigabyte got only up to the 7.5 GHz region

- there is a known bug that will allow you to verify false clock numbers.

Conclusion? You be the judge!

 

Listen to DerBauer's reasoning.

And here some more:

These two sources are literally in the first post of this topic as well.

[CarlBar]

18 minutes ago, OneOfYas said:

There is no evidence, yet this community would rather tarnish a company based on logical fallacy....

 

 

Other OC'ers are indicating lack of belief. They do in fact get to make decisions like that. As noted previously, it's the kind of thing they take real seriously, (for obvious reasons).

[leadeater]

3 hours ago, OneOfYas said:

There is no evidence, yet this community would rather tarnish a company based on logical fallacy....

 

What do you mean no evidence, there is literally a confirmed and known bug with these CPUs not taking multiplier changes and misreporting CPU clocks. What more evidence do you need?

 

Dubious OC result + known bug existence = likely not legitimate result.

 

Repeat OC again with knew knowledge to confirm to yourself and prove that it's legitimate. Like I said if it can be done once it can be done again.

 

Out of a single OCer and the CPU-Z dev who looks after the validation I'd trust the latter's judgement not the former's claim if the result is disputed.

 

This is a situation where the result is contested and there is a lack of evidence proving it happened beyond reasonable doubt.

[Beerzerker]

4 hours ago, OneOfYas said:

There is no evidence, yet this community would rather tarnish a company based on logical fallacy....

 

No evidence?
Logical fallacy?
The errata it has is evidence enough to at least bring things into question and the errata itself has been proven to exist too.
The last part of that sounds a little fan-ish to me too.

 

Do you have real evidence to the contrary?
Just show it to us already, in all fairness we can't ask for more than that.

 

[Blademaster91]

4 hours ago, OneOfYas said:

yet this community would rather tarnish a company based on logical fallacy...

Well to be honest if this was Asus I feel the reactions would be a lot different.

Anyway I like Gigabyte boards, I think its stupid they did this, although I don't see how this really matters for most users that aren't pro OC'ers.

[Beerzerker]

2 minutes ago, Blademaster91 said:

Well to be honest if this was Asus I feel the reactions would be a lot different.

Anyway I like Gigabyte boards, I think its stupid they did this, although I don't see how this really matters for most users that aren't pro OC'ers.

TBH it's only matters to those (Like me) that chase WR's all the time, to anyone else that's not pouring Ln2 into a pot on the chip it doesn't really mean squat because it's not like you'll be gaming at 8 or even 7GHz on a daily basis... Or at all really.

[CT854]

On 11/7/2021 at 8:22 AM, Spotty said:

The question is did Gigabyte know that its overclocking result was likely a bogus result or were they not aware of the issue? I'd wager they weren't aware of the issue and the moment the overclocker announced their impressive overclocking results Gigabyte's PR team ran with it.

So funny to me how so many of the controversies in the PC gaming space is pretty much just "backend makes a mistake, PR shits the bed by doing PR things"

[HenrySalayne]

7 minutes ago, OneOfYas said:

I am judging that we have no solid evidence, would you do this in court with someone's life?

 

These are also competitive overclockers, of course they will get sour and emotional.

No, claiming a record works the opposite way around: the person who wants to be acknowledged as the record holder has to prove they've achieved it. Nobody has to prove Gigabyte hasn't done it.

So either they do it again or they present more evidence. And this wouldn't be such a big topic if Gigabyte wouldn't have used an unconfirmed record for their advertisement campaign.

[OneOfYas]

1 minute ago, HenrySalayne said:

No, claiming a record works the opposite way around: the person who wants to be acknowledged as the record holder has to prove they've achieved it. Nobody has to prove Gigabyte hasn't done it.

So either they do it again or they present more evidence. And this wouldn't be such a big topic if Gigabyte wouldn't have used an unconfirmed record for their advertisement campaign.

That works well.

Gigabyte should not be pushing endorsements until this is reattempted.

At least that seems a logical outcome of the situation.

[WkdPaul]

- thread cleaned -

 

The rules outlined in the Community Standards have to be followed at all time.

[Lightwreather]

Update: https://www.tomshardware.com/news/cpu-z-invalidates-gigabyte-8ghz-world-record-amid-lack-of-evidence

 

Quote

Apparently, this wasn't Gigabyte or HiCookie's first rodeo. When AMD launched the Ryzen 5 5950X, Gigabyte was avidly promoting how its X570 Aorus Master motherboard had set a new world record with the Zen 3 chip reportedly at 6,362.16 MHz. The submission was rejected due to a known bug with Ryzen 5000 chips at launch. Doc TB utilized the new analysis algorithm on the CPU-Z submission and has confirmed that the 16-core processor was really running at 5,683.94 MHz. Guess who had submitted that dubious result to HWBot?

 

The CPU-Z team is cooking up a new algorithm to identify the bogus submissions from the real ones and will reprocess all the Alder Lake entries to remove the fakes. At the time of writing, the current record for the Core i9-12900K is at 7,543.95 MHz, achieved by overclocker Jon "Elmor" Sandström with liquid helium on Asus' ROG Maximus Z690 Apex motherboard. This puts an end to the overlocking soap opera, and Gigabyte would have gotten away with it too, if it weren't for a meddling CPU-Z Validator developer.

 

[tikker]

7 hours ago, J-from-Nucleon said:

with liquid helium

Since when are we OC-ing near absolute zero?

[Stahlmann]

On 11/7/2021 at 3:23 PM, MultiGamerClub said:

Gigabyte shooting itself in the foot..

 

Why they does this move when their rep is already bad.. Don't understand that company

At this point the have everything to gain and nothing to lose

[BetteBalterZen]

13 minutes ago, Stahlmann said:

At this point the have everything to gain and nothing to lose

True dat. I'm never buying anything from Gigabyte again. 

[Stahlmann]

7 minutes ago, BetteBalterZen said:

True dat. I'm never buying anything from Gigabyte again. 

I'm not at this point yet. From my experience you'll always run into issue no matter the brand. My current Asus support experience is also terrible. But that doesn't mean i won't buy Asus again. Just not from them directly, but through Amazon where i won't have to deal with them myself in case i need support again.

[BetteBalterZen]

1 minute ago, Stahlmann said:

I'm not at this point yet. From my experience you'll always run into issue no matter the brand. My current Asus support experience is also terrible. But that doesn't mean i won't buy Asus again. Just not from them directly, but through Amazon where i won't have to deal with them myself in case i need support again.

True, my experience with ASUS haven't really been flawless either but lately with Gigabyte, haha, total joke IMO. They're dead to me

[Stahlmann]

Just now, BetteBalterZen said:

True, my experience with ASUS haven't really been flawless either but lately with Gigabyte, haha, total joke IMO. They're dead to me

Just like with other brands there are also competent engineers at Gigabyte. Sadly their competence is undermined by stupid and incompetent marketing or another engineering team responsible for other products. Just because their PSUs are terrible doesn't mean their monitors are terrible. These different products are engineered by different people and manufactured by different factories. So in my opinion avoiding brands because you don't like their PR also means avoiding some good products. In the monitor market Gigabyte is currently offering some of the best products, even when compared to Asus, Dell, etc. If you decide to avoid them, all power to you. But just know that you also avoid good products then.

[BetteBalterZen]

26 minutes ago, Stahlmann said:

Just like with other brands there are also competent engineers at Gigabyte. Sadly their competence is undermined by stupid and incompetent marketing or another engineering team responsible for other products. Just because their PSUs are terrible doesn't mean their monitors are terrible. These different products are engineered by different people and manufactured by different factories. So in my opinion avoiding brands because you don't like their PR also means avoiding some good products. In the monitor market Gigabyte is currently offering some of the best products, even when compared to Asus, Dell, etc. If you decide to avoid them, all power to you. But just know that you also avoid good products then.

True. But I also avoid good products by not going with MSI hehe.

I do feel bad for the competent people at Gigabyte though, but that still won't make me buy their products. That ship has sailed. 

[tikker]

1 hour ago, James Evens said:

@tikker Very long but only a few can do it:

Liquid Helium is expensive. Very expensive especilly if you want to cool that kind of heatload without a recovery system.

 

How long will people do that? Don't know until the environment hypetrain noticed how wasteful it is and make big cooperation's no longer want to spend there PR money on that.

Cool, didn't know people did that. I thought it would be too cold. I knew Helium is expensive. During my masters we needed to cool stuff. LN2 was basically the equivalent of tap water, you could just walk in and fill up a 5L dewar without people caring much. Helium needed explicit purpose permission to tap