AMD Asks Motherboard Makers to Remove Overclocking Options for Ryzen 7 5800X3D

[Zareek]

Sweet, no overclocking... That means they'll be less competition and I might get my hands on one. That is, if the performance is great and the price is reasonable.

 

[OneMoar]

I mean most of us are already tuning PBO down anyway to get better performace
so what is the issue
the 5800x runs hot more cache on top of that is going to make that worse

I assume they are talking about Manual/Allcore overclocks here which I have no problem with because that is pretty pointless

 

[qlum]

A shame if true, even if it's largely irrelevant for the vast majority of users. Unless this also affects the memory controller which I doubt.

 

[Deleted member 24505]

let's hope it is a stonker or it will be a double kick to the nuts.

 

[Deleted member 218758]

I guess it's time to backup every 1204(A)/1205/1206 bios that exists for 30+ boards
~ as SMU is identical in size and perfectly transplant-able between these versions

 

[TheLostSwede]

@TheLostSwede you made the front page of TH today along with a few other tech sites.

That's nice.

We will in fact see it this very generation of Intel CPUs, if ever somebody releases a cheap DDR4 B660/Z690 motherboard with an external clock generator.

The i3-12100f is a hundred eurobucks, can be cranked over 20% - easily more, depending on how lucky you get with the memory controller on locked System Agent (memory controller) voltage.
Really cool stuff.. At some indeterminate point in time!

No we won't. 20% isn't 50% or more as it was sometimes back then.

Yup everyone has to wait for an actual product review instead of just hearsay.

I mean this isn't hearsay, I confirmed it with one of my sources.

That said, as we don't know the reason for this, or if it's a temporary or permanent thing, it's most definitely too early to draw any conclusions.

 

[Bwaze]

"The information suggests that this happened back at the end of January, although it's no surprise that this information took some time to leak, as it's not the kind of information that would normally make its way outside of the motherboard manufacturers."

But it's an information reviewers would surely point out, and customers notice?

Not necessarily, just remember der8auer's polls about Ryzen 3000 frequencies back in 2019 - he clearly showed most customers were not getting the advertised frequencies, a problem no reviewer touched.

Or the latest AMD fTPM debacle. Users were reporting stuttering for a year now. Now we finally have an official response the bug exists, but for a year no reviewer noticed anything.

 

[seth1911]

Not necessarily, just remember der8auer's polls about Ryzen 3000 frequencies back in 2019 - he clearly showed most customers were not getting the advertised frequencies, a problem no reviewer touched.

Or the latest AMD fTPM debacle. Users were reporting stuttering for a year now. Now we finally have an official response the bug exists, but for a year no reviewer noticed anything.

U cant write this, Amd is the holy grail of the tech companys

If Amd make a bad move to consumer = Good
If intel make that = get in hell

 

[Bwaze]

I don't think AMD vs. Intel has anything to do with it. Intel also got away scot free with all the mess in the last few generations of processors regarding security flaws and mitigation impact, and absurd power draw figures - at least if you only look review sites.

Independent tech journalism is more or less dead, the sites that tried to go against the flow were very publicly punished (HardOCP, Anandtech... ), and we consumers have applauded at the spectacle and are now getting our prize - instead of journalism we're getting commercials, not even wrapped into something resembling independent reviewer experiences. Like fanboy Youtuber praises, in hope of getting another free product in the next cycle of releases.

 

[phanbuey]

I don't think AMD vs. Intel has anything to do with it. Intel also got away scot free with all the mess in the last few generations of processors regarding security flaws and mitigation impact, and absurd power draw figures - at least if you only look review sites.

Independent tech journalism is more or less dead, the sites that tried to go against the flow were very publicly punished (HardOCP, Anandtech... ), and we consumers have applauded at the spectacle and are now getting our prize - instead of journalism we're getting commercials, not even wrapped into something resembling independent reviewer experiences. Like fanboy Youtuber praises, in hope of getting another free product in the next cycle of releases.

I dont know if they "Got away with" -- Intel got absolutely raked over the coals. Their stock got comparatively destroyed and even when the 3000 series were out, which were objectively worse processors for gaming even than the cheaper 8700ks at the time, people were recommending them over intel.

 

[Bwaze]

But mostly by the users, and some more independent Youtube reviewers. Some "trusued" review sites even used completely dishonest reviews, like using power cap when measuring power draw ("because it is stock Intel configuration"), and then using full available power draw when benchmarking ("because it is stock setting on most motherboards").

The site that in my opinion most honestly reviewed the stated power draw figures by Intel, Anandtech, has ended on the "black list" by most vendors - they aren't receiving their free samples any more. Gotta send them to Youtube "influencers" (who should at this point perhaps generate more hatred than interest, but who am I kidding)...

 

[TheLostSwede]

"The information suggests that this happened back at the end of January, although it's no surprise that this information took some time to leak, as it's not the kind of information that would normally make its way outside of the motherboard manufacturers."

But it's an information reviewers would surely point out, and customers notice?

Not necessarily, just remember der8auer's polls about Ryzen 3000 frequencies back in 2019 - he clearly showed most customers were not getting the advertised frequencies, a problem no reviewer touched.

Or the latest AMD fTPM debacle. Users were reporting stuttering for a year now. Now we finally have an official response the bug exists, but for a year no reviewer noticed anything.

How can reviewers point out something they don't know about an unreleased product?
I got additional information from one of my sources that very few people are aware of.
He's raging on Twitter about this, but we don't even know if this is a temporary thing or not. Maybe AMD simply did it so people can't overclock engineering samples of the CPU. We simply don't know.

 

[Mussels]

Theres two reasons


1. They want to do an intel and lock it down

2. They have a new fancy overclocking method they want to force people to use



If it's 2, i bet its a preview of how AM5 overclocking/PBO works

Do Zen 3 overclock at all, they seem already at their limit.

Yes - it's just that all core and boost are different values, that need to be clocked differently.

Few boards let you manually dial that in, so people choose a lower all core, or get worse multi threaded peformance to chase the low threaded gainz
(Again, if AMD bring something new that lets us tweak things better....)

 

[Bwaze]

More likely they are forced to lock it down, because there isn't any available headroom left in the chip, and manual overclocking would only result in degradation or some other damage.

 

[Bloax]

No we won't. 20% isn't 50% or more as it was sometimes back then.

if you want to be that picky about numbers in a multicore at reasonable stock frequencies world, then the Cellery G6900 (yes, Cellery once again) can do +50-60% and happens to be a not-core2duo with 2c/2t

so really if you look at it properly, you get twice the cellery, decades of selective breeding for taste and crunch, for even cheaper
crazy!

 

[TheLostSwede]

if you want to be that picky about numbers in a multicore at reasonable stock frequencies world, then the Cellery G6900 (yes, Cellery once again) can do +50-60% and happens to be a not-core2duo with 2c/2t

Pentium II thankyouverymuch. You just sound jealous that you weren't playing with fun stuff like tha back then.

so really if you look at it properly, you get twice the cellery, decades of selective breeding for taste and crunch, for even cheaper
crazy!

I hate celery, just the smell of it makes me want to vomit.

 

[looniam]

i'll just leave these here:

https://twitter.com/x/status/1501889826848755714

https://twitter.com/x/status/1501890684206432257

 

[phanbuey]

i'll just leave these here:

https://twitter.com/x/status/1501889826848755714

https://twitter.com/x/status/1501890684206432257

Would make sense.

 

[RJARRRPCGP]

Theres two reasons


1. They want to do an intel and lock it down

I fear that AMD is going to treat them like a non-K Intel!

 

[ghazi]

I don't think anyone's lost enthusiasm... it's just that new tech > overclocking. I can run my chip at 5.4Ghz, but I don't see any benefits from 5.2 to 5.4 in games because I am memory bottlenecked. Faster cache memory gives huge gains, more core clock almost nothing.

if I had to choose between a heavily clocked 5800x and a 5800x with 96 mb of 3d cache would be an easy choice.

You do have a point, but hasn't it always been that way? Discounting rare cases like the 3770K being slower than the 2600K OC vs OC. And yes, talking 5.4GHz on Intel you are bottlenecked, but with 5800X3D running 3.4 base / 4.5 boost, and having all that extra cache to feed it, the cores still have more potential.

New tech is also really expensive and assembling a PC takes 2-4 hours if you don't use a custom loop. But OCing and tuning gives many whole days of entertainment at practically no cost. And the enthusiasm is definitely down. Even though I suppose more people are going to extremes since returns are harder to get.

"The information suggests that this happened back at the end of January, although it's no surprise that this information took some time to leak, as it's not the kind of information that would normally make its way outside of the motherboard manufacturers."

But it's an information reviewers would surely point out, and customers notice?

Not necessarily, just remember der8auer's polls about Ryzen 3000 frequencies back in 2019 - he clearly showed most customers were not getting the advertised frequencies, a problem no reviewer touched.

Or the latest AMD fTPM debacle. Users were reporting stuttering for a year now. Now we finally have an official response the bug exists, but for a year no reviewer noticed anything.

der8auer actually misrepresented his OWN data and nobody pointed it out or noticed, and no reporters touched it. See the attached file, he intentionally deflated his averages by 100MHz and thought nobody would actually run his own data themselves. In the top left corner he writes average 4375MHz, and says so in his video, when the real average of the data he shows is 4475MHz. He is a shameless Intel shi!! through and through desperate to keep selling delid kits.

 

[RJARRRPCGP]

What's next? AMD suggesting that it's a waste to get more than an A520 chipset for that CPU?!

 

[OneMoar]

You do have a point, but hasn't it always been that way? Discounting rare cases like the 3770K being slower than the 2600K OC vs OC. And yes, talking 5.4GHz on Intel you are bottlenecked, but with 5800X3D running 3.4 base / 4.5 boost, and having all that extra cache to feed it, the cores still have more potential.

New tech is also really expensive and assembling a PC takes 2-4 hours if you don't use a custom loop. But OCing and tuning gives many whole days of entertainment at practically no cost. And the enthusiasm is definitely down. Even though I suppose more people are going to extremes since returns are harder to get.


der8auer actually misrepresented his OWN data and nobody pointed it out or noticed, and no reporters touched it. See the attached file, he intentionally deflated his averages by 100MHz and thought nobody would actually run his own data themselves. In the top left corner he writes average 4375MHz, and says so in his video, when the real average of the data he shows is 4475MHz. He is a shameless Intel shi!! through and through desperate to keep selling delid kits.

that's quite an accusation for 80 posts

also lol selling delidding kits that hasn't been relevant since both sides switched back to solder
take your rabid fanboyism elseware

 

[ghazi]

that's quite an accusation for 80 posts

also lol selling delidding kits that hasn't been relevant since both sides switched back to solder
take your rabid fanboyism elseware

The data speaks for itself, I don't have to rely on worthless forum clout, post counts or badges, and I have that elsewhere anyway. But apparently that's all you've got to rely on, what are you adding to the discussion?

The video is from 2019 and that's when I first called it out and what the delid comment is about. He kept trying to sell delid kits for 9900K and people did buy them, solder isn't as good as direct die. No idea what he does now. Take your wholly unqualified arrogance somewhere else, it only makes you look like more of a fool.

 

[OneMoar]

The data speaks for itself, I don't have to rely on worthless forum clout, post counts or badges, and I have that elsewhere anyway. But apparently that's all you've got to rely on, what are you adding to the discussion?

The video is from 2019 and that's when I first called it out and what the delid comment is about. He kept trying to sell delid kits for 9900K and people did buy them, solder isn't as good as direct die. No idea what he does now. Take your wholly unqualified arrogance somewhere else, it only makes you look like more of a fool.

pot meet kettle

 

[ghazi]

pot meet kettle

And yet what I said is irrefutably true and incontrovertibly substantiated regardless of your vacuous bile. But it is difficult to display as many negative traits as you have in such few words in a forum post, so I will give you credit. Now, der8auer falsified his claims about Zen 2 clock speeds, period.