AMD Ryzen 7000X3D Processors Prone to Physical Damage with Voltage-assisted Overclocking, Motherboard Vendors Rush BIOS Updates with Voltage Limiters

[AusWolf]

I wonder if there's been any more news about this, or any more cases confirmed.

 

[Zubasa]

I wonder if there's been any more news about this, or any more cases confirmed.

Wait for GN's video on the topic, it is going to be spicy.

 

[AusWolf]

Wait for GN's video on the topic, it is going to be spicy.
View attachment 293295

Nice. I gather it's actually Asus at fault here.

 

[Zubasa]

Nice. I gather it's actually Asus at fault here.

It is pretty clear from the start it is a board vendor fault, they have been sneaking in Auto OCs to make themselves look good in reviews, and it is nothing new.
By constantly trying to one up each-other, It is only a matter of time until one of them went too far.

The CPU by itself is just an inert object. It cannot generate any of the voltages it requires to function.
I have never heard of CPUs killing motherboards, only the other way around.

 

[Dirt Chip]

It is pretty clear from the start it is a board vendor fault, they have been sneaking in Auto OCs to make themselves look good in reviews, and it is nothing new.
By constantly trying to one up each-other, It is only a matter of time until one of them went too far.

The CPU by itself is just an inert object. It cannot generate any of the voltages it requires to function.
I have never heard of CPUs killing motherboards, only the other way around.

Thing is, even at extreme voltage and temp the CPU power\voltage\temp protection should kick in and shutdown before any motherboard damage accrue.
Unless all vendors bypass those protection (and then it will be an amusing shit-show to watch) than the CPU protection design is responsible as well.
The bus is long, plenty of room for all under it..

Also, it will be interesting debate about whos fault is when you get out-of-the-box OC from the mobo settings that practically cancel your warranty before even using the product.

 

[AusWolf]

It is pretty clear from the start it is a board vendor fault, they have been sneaking in Auto OCs to make themselves look good in reviews, and it is nothing new.
By constantly trying to one up each-other, It is only a matter of time until one of them went too far.

The CPU by itself is just an inert object. It cannot generate any of the voltages it requires to function.
I have never heard of CPUs killing motherboards, only the other way around.

Exactly my thoughts! Yet, here we are with a headline article titled "Ryzen 7000X3D CPUs have a voltage problem". No, they don't. Certain Asus boards with certain BIOS versions do. This is some Der8auer level of flaming and shaming. At least TPU should be more technically literate than this.

Thing is, even at extreme voltage and temp the CPU power\voltage\temp protection should kick in and shutdown before any motherboard damage accrue.
Unless all vendors bypass those protection (and then it will be an amusing shit-show to watch) than the CPU protection design is responsible as well.
The bus is long, plenty of room for all under it..

I'm wondering how much of that protection is built into the CPU itself, and how much of it is part of the motherboard BIOS.

Also, it will be interesting debate about whos fault is when you get out-of-the-box OC from the mobo settings that practically cancel your warranty before even using the product.

That's been an ongoing debate ever since boost clocks were invented. Interesting question, nonetheless.

 

[Zubasa]

Thing is, even at extreme voltage and temp the CPU power\voltage\temp protection should kick in and shutdown before any motherboard damage accrue.
Unless all vendors bypass those protection (and then it will be an amusing shit-show to watch) than the CPU protection design is responsible as well.
The bus is long, plenty of room for all under it..

Also, it will be interesting debate about whos fault is when you get out-of-the-box OC from the mobo settings that practically cancel your warranty before even using the product.

I'm wondering how much of that protection is built into the CPU itself, and how much of it is part of the motherboard BIOS.


That's been an ongoing debate ever since boost clocks were invented. Interesting question, nonetheless.

I am leaning towards the motherboard bios, in most case.
For example on Intel's page tjmax on my 12700k is 100C. Yet my motherboard has the option to increase that to 118C and I did tested it with Prime95 smallFFT, the cpu stopped throttling @100C.
It shows that the mobo at least can override the protection to a certain degree.

 

[Dirt Chip]

I'm wondering how much of that protection is built into the CPU itself, and how much of it is part of the motherboard BIOS.


That's been an ongoing debate ever since boost clocks were invented. Interesting question, nonetheless.

If the problem accurse on mobo`s from different vendors than AMD is the root source.
Also, the design of the CPU sensors is solely AMD so if they over-look a (rare) case that enable the CPU to reach extreme temp than it`s their head on the spike (read as: just RMA and move on).
I guess that AMD make a basic bios and then each vendor modify to their suite so it may be that the CPU hardware\design\sensors is spotless and the problem is the basic bios from AMD that doesn't use the data in the right way (and then easy fix).

 

[Zubasa]

If the problem accurse on mobo`s from different vendors than AMD is the root source.
Also, the design of the CPU sensors is solely AMD so if they over-look a (rear) case that enable the CPU to reach extreme temp than it`s there head on the spike (read as: just RMA and move on).
I guess that AMD make a basic bios and then each vendor modify to their suite so it may be that the CPU hardware\design\sensors is spotless and the problem is the basic bios from AMD that doesn't use the data the right way (and then easy fix).

It is not as simple as you think, on my old X399 board there is actually a physical switch labeled ProcHot that disable the thermal shut down.
It is one thing for the CPU to signal the mobo to shutdown, it is another thing if the board keeps going or not.

 

[Dirt Chip]

It is not as simple as you think, on my old X399 board there is actually a physical switch labeled ProcHot that disable the thermal shut down.
It is one thing for the CPU to signal the mobo to shutdown, it is another thing if the board keeps going or not.
View attachment 293304

Yep, and if this is the case- a power play (with and without quotes ) between mobo vendors to AMD about who controls critical CPU protection, all done on the consumer back- than as I said it will be a very good show to watch.

It can put off many mainstream consumer that just want a solid, reliable build. Hope it will never come to that.

 

[Zubasa]

Yep, and if this is the case- a power play (with and without quotes between mobo vendors to AMD about who controls critical CPU protection- than as I said it will be a very good show to watch.

It can put off many mainstream consumer that just want a solid, reliable build. Hope it will never come to that.

TBH the option already exists, for people who just want something reliable.
Those are the machines from the major OEMs like Dell / HP / Lenovo / Acer.
These machines tends to have almost no options in their bios thus very little points of failure. No XMP / EXPO / MCE / PBO whatsoever.
Also no finger pointing with who is at fault when things goes south, you just call the brand you brought it from.
There are many reasons why these are the machines you see in businesses.

I would assume even Asus' own non-ROG OEM machines are much more failsafe than their gaming stuff.

 

[Dirt Chip]

TBH the option already exists, for people who just want something reliable.
Those are the machines from the major OEMs like Dell / HP / Lenovo / Acer.
These machines tends to have almost no options in their bios thus very little points of failure. No XMP / EXPO / MCE / PBO whatsoever.
Also no finger pointing with who is at fault when things goes south, you just call the brand you brought it from.
There are many reasons why these are the machines you see in businesses.

I would assume even Asus' own non-ROG OEM machines are much more failsafe than their gaming stuff.

This is a very bad advice, to opt those OEM even for a normal day-to-day PC..
better go Intel hand-build instead, isn't it?

 

[Zubasa]

This is a very bad advice, to opt those OEM even for a normal day-to-day PC..
better go Intel hand-build instead, isn't it?

FYI, DIY boards are doing all kinds of hackery on the Intel side as well.
You seems to be under the impression that this can only happen on AMD machines.

There is one difference on the Intel side, on Intel there are the Non-K CPUs and B / H series boards that are much more lock down.
Until 12th gen you cannot even do memory OC on B series boards, those effectively performs like OEM machines.

 

[Dirt Chip]

I am leaning towards the motherboard bios, in most case.

Isn`t that phenomena was confirm multiple vendors?
So all of them, separately, doing the same hidden off-spec overvoltageing and\or bypass fundamental CPU protection?
Because all of them removed all old bios I suspect that AMD contact them and instruct what to do (as a new restrictive bios) until further info emerge.

 

[Zubasa]

Isn`t that phenomena was confirm multiple vendors?
So all of them, separately, doing the same hidden off-spec overvoltageing and\or bypass fundamental CPU protection?
Because all of them removed all old bios I suspect that AMD contact them and instruct what to do (as a new restrictive bios) until further info emerge.

Yes, you can look up different articles by likes of GN and Hardware Unboxed over the years.
There is specifally a "normal" setting instead of "auto" in each voltage setting on Gigabyte boards which were implemeneted up on reviewers request.
That setting means the board actually respects the Intel / AMD voltage spec instead of the presets "optimized" by the board vendor.
Additionally there is an "Intel POR" setting that sets both PL1 and PL2 to the Intel spec (on the 12700K @190W) instead of auto which is 4095W / unlimited.
This was added so that reviewers have an easier time testing the CPU actully at stock.

Correct, I remember this one vividly

Don't Run Z490 Motherboards with Default Settings: Thermals, Power, Boosting, & MCE for 10th Gen CPUs

It’s difficult to differentiate motherboards, at least from a marketing perspective. There are definitely better and worse boards, and you can check any of the roundups or reviews Buildzoid has produced for this channel for explanations as to why, but “better” doesn’t mean “higher FPS in games”...

www.gamersnexus.net

I could easily fish for another handful of generations/sockets where this has occurred at one or more mobo mfgrs.

So again... user due diligence is needed, has been needed and will always be needed with DIY building. Its why we call it DIY. People need to get a life and a good bit of common sense. And AMD/mobo partners need to lock this down on X3D's just as well, because there is literally no purpose to leave it unlocked. And if unlockable, it needs to come with warning on the UEFI toggle.

Frankly if you fried an X3D, you're just a n00b. And you just learned a thing. RMA it / get new one
/thread

Just in case you missed this post from another user.

 

[Dirt Chip]

FYI, DIY boards are doing all kinds of hackery on the Intel side as well.
You seems to be under the impression that this can only happen on AMD machines.

There is one difference on the Intel side, on Intel there are the Non-K CPUs and B / H series boards that are much more lock down.
Until 12th gen you cannot even do memory OC on B series boards, those effectively performs like OEM machines.

I`m talking about possible 'hot potato' of responsibly that will be thrown between AMD and vendors (as steve from GN suggested) as the cause of consumer to choose 'less risky option' just as a matter of the bad-product-image that this fight will make.
That thing is not a far fetched scenario as to with Intel side.
Shit can and have happen on Intel side more than once, but now it is AMD side who take the heat.
If you see a fight , you probably go elsewhere to buy ice-cream (except us, tech nerds, that pray on each yellowish detail and debate it to death..).

And on top of all, just as you said- going non-K Intel is pretty safe. And just like that, the restrictiveness of the platform ('safer' non-K) became it`s best merit...

 

[Zubasa]

I`m talking about possible 'hot potato' of responsibly that will be thrown between AMD and vendors (as steve from GN suggested) as the cause of consumer to choose 'less risky option' just as a matter of the bad-product-image that this fight will make.
That thing is not a far fetched scenario as with Intel side.
Shit can and have happen on Intel side more than once, but now it is AMD side who take the heat.
If you see a fight , you probably go elsewhere to buy ice-cream (except us, tech nerds, that pray on each yellowish detail and debate it to death..).

And on top of all, just as you said- going non-K Intel is pretty safe. And just like that, the restrictiveness of the platform ('safer' non-K) became it`s best merit...

FYI, they figured out ways to overclock non-k CPUs too on 12th gen....
On the older gen, as soon as you increase BCLK over 102 the system will refuse to boot, on 12th gen somehow it is no longer the case.
Intel Alder Lake non-K Overclocking via BCLK – der8auer

That Vcore though....

 

[Dirt Chip]

Yes, you can look up different articles by likes of GN and Hardware Unboxed over the years.
There is specifally a "normal" setting instead of "auto" in each voltage setting on Gigabyte boards which were implemeneted up on reviewers request.
That setting means the board actually respects the Intel / AMD voltage spec instead of the presets "optimized" by the board vendor.
Additionally there is an "Intel POR" setting that sets both PL1 and PL2 to the Intel spec (on the 12700K @190W) instead of auto which is 4095W / unlimited.
This was added so that reviewers have an easier time testing the CPU actully at stock.



Just in case you missed this post from another user.

I know about vendors changing turbo limits, on both Intel and AMD boards, to gain more pref and that it is much more pronounce on Intel boards (13900k to 350+w, I know, I have one with the same AERO GB board you have) but here we talk about something else - fundamental CPU protection that some how didn't work even with extreme off-spec values.
As if somehow the hierarchy of protection changed (or just plain cancelled) and the CPU was allowed to first go out-of-the-roof voltage (as 4095w analog) and only then apply the temp control (95-105 degree) as opposite to "go to unlimited W as long as under x temp".

FYI, they figured out ways to overclock non-k CPUs too on 12th gen....
On the older gen, as soon as you increase BCLK over 102 the system will refuse to boot, on 12th gen somehow it is no longer the case.
Intel Alder Lake non-K Overclocking via BCLK – der8auer

That Vcore though....

A nice fun fact (and limited compering real OC I might add) I know about, thank you. and so, it`s just strengthen your own argue that non-K is 'safer'. You need not-so-little manual bios changes to do that BCLK OC. No ordinary consumer will do that plus no vendor will ship an non-k BCLK-OC PC out of the box (if they have any sense in them..)

 

[AusWolf]

How hard would it be for Intel and AMD to have stricter specs and enforce motherboard manufacturers to use them as the default setting?

 

[Bomby569]

It's not just ASUS. Gigabyte, Asrock, Biostar, so something else is broken.

It's a bit like the Nvidia cables, if they were better designed, clear rules on how to use, systems in place to prevent misuse, QA, this would not have happened.

Now it's blame shifting, but i blame AMD and the partners. AMD surely tests their CPU's on the actual mobos they will be working on. Come on. This is like the most basic QA. But just like in games this days, we do the QA and beta testing, and engineering control.

Cases and cases keep pilling up, these companies, software and hardware just don't do QA, PERIOD.

 

[Dirt Chip]

How hard would it be for Intel and AMD to have stricter specs and enforce motherboard manufacturers to use them as the default setting?

I guess very easy but not very welcome all the same. It might lead AMD to restrict OC just as Intel on non-K CPU`s.
We don`t need more restriction on every level of OC, just that the very basic to work rock solid (that is shutdown before any permanent and\or physical damage to the CPU\mobo).

My guess- a mix responsibly that start with AMD oversighting a rare case (as those incidents are very few so far) that caused by mobo vendors that changed the optimal settings (probably EXPO related) in order to get better out of the box pref for reviews, without telling it`s canceling warranty.
ZEN 4 is a new platform so, despite all the QA, a glitch found it's way out and make a mass.

 

[chrcoluk]

Why do people not just leave the CPUs as they are and as for all the bashing about rushed product and being the fault of AMD. That is not true. One of the things that people who have built many PCs knows is that Gigabyte especially likes to push more voltage than reported by software but MSI and AS Rock are not far behind. Only buy Asus TUF and above as Prime is treated like budget by Asus. I have not seen my 7900X3D go past 1.36 volts and I don't mind at all. I know that there is a narrative out there that OC is always a good thing but the current generation of chips are already tuned to perform as it is not like these chips are being released in Dead space. Thank Intel for taking up the gauntlet and challenging AMD. When I had my 5800X3D I lamented about the clock speed. Well looks like the 7800X3D can also boost over 5 GHZ and that is not something to diminish. DDR5 pricing is no longer a mitigating factor either but the only thing I would do for OC would be to use AMD Software and the one button choices like undervolting.

Good that AMD chips are not going to 1.5v at stock anymore?

Agree, the only time I tinker now would be undervolting to reduce power, modern CPUs and GPUs are basically pre binned to run at almost as fast as they can go out of the box. I remember trying on my 2700X, PBO was actually slower than stock as it hit temp throttling earlier, and when I tried manual o/c, even an extra 100mhz would freeze during cinebench.

5600G so fast out of the box, I didnt even bother trying, just disabled XFR for low voltage, power consumption and ran it that way since. (its been used mostly as a server type workload, so XFR not important).

Yes, you can look up different articles by likes of GN and Hardware Unboxed over the years.
There is specifally a "normal" setting instead of "auto" in each voltage setting on Gigabyte boards which were implemeneted up on reviewers request.
That setting means the board actually respects the Intel / AMD voltage spec instead of the presets "optimized" by the board vendor.
Additionally there is an "Intel POR" setting that sets both PL1 and PL2 to the Intel spec (on the 12700K @190W) instead of auto which is 4095W / unlimited.
This was added so that reviewers have an easier time testing the CPU actully at stock.
View attachment 293312View attachment 293313View attachment 293314View attachment 293315


Just in case you missed this post from another user.

An eye opener, on my ASRock boards I observe enabling XMP pumps excessive voltage to System Agent. Also voltage to CPU at default was above intel stock vcore.

 

[Redwoodz]

If the problem accurse on mobo`s from different vendors than AMD is the root source.
Also, the design of the CPU sensors is solely AMD so if they over-look a (rare) case that enable the CPU to reach extreme temp than it`s their head on the spike (read as: just RMA and move on).
I guess that AMD make a basic bios and then each vendor modify to their suite so it may be that the CPU hardware\design\sensors is spotless and the problem is the basic bios from AMD that doesn't use the data in the right way (and then easy fix).

So if a buyer buys a AM5 motherboard with an old bios(preX3D) and inserts an X3D chip without flashing the new bios with proper support, who's at fault?

 

[Dirt Chip]

So if a buyer buys a AM5 motherboard with an old bios(preX3D) and inserts an X3D chip without flashing the new bios with proper support, who's at fault?

I think It should be impossible to do (the PC shouldn't POST at all in that case) and it's first under AMD responsibly to set those guidelines and than under each vendor to make sure their bios is solid about it.

The bare minimum is to display a warning massage on POST\bios that requires bios update and alert from warranty cancel if not done.

And even if disregarding all of that by the user, the CPU shouldn't kill itself and take the mobo with it.

 

[Klemc]

I applied STRIX B650E-E GAMING BIOS 1413...