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

[Serhend]

Everything works the best just out of the box.
The time and effort spent to achieve a stable OC is not worth it especially the first 3 years of having a new gen GPU or CPU. Only after that it might make sense to extract that 5 to 15% at best extra. By that time the product is also worth less than what it is on day 1 anyways.

 

[Zubasa]

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

 

[Dirt Chip]

- Fallowing reset, the Asus mobo keep feeding the CPU with AMP`s even after the CPU was killed- above 400w being feed to a dead CPU...
This is dispite many extra fancy PR protection capabilities that this spacific mobo hac. That`s way the mobo was damaged as well, not just the CPU.
- Gigabyte mobo was able to stop the above as to not damage the CPU socket (as oppose to the Asus board). The CPU was still killed.
- It is not directly EXPO related, but turning EXPO on is one way of triggeingr of the CPU Suicide.
- The CPU death probabilityIt is very much silicon loterry depended- on weak silicon the probability is much higher. In the short term stronger silicon can cope with the too high voltage being feed but may give up at some point in the futere. Talk about 'slow cooking' of the CPU.
- Verious bugs (will be an additional video on that) in the bios were discoverd and wrong\un-related output of voltage is measured, above what is displayed and set in the bios screen.
- All in all, a rare case that coming from mix neglectance of both AMD and vendors.
- Do keep on choosing ZEN4 if it`s the right product for you, but Do double check and be yourself in control\monitoring over voltages.
- More tests are coming with in-depth insights.
- AMD RMA any CPU case, even if EXPO was applay. Mobo vendors RMA- yet to be seen...

My take- a cluster of many small things coming together to a catastrophic error. Like in Airplane investigation you see in National Geographic. This is the hidden cost of a new platform and with AM5 it appers there still a lot to irone out. Unless you are into some adventures, wait to ZEN5 if you can. ZEN4 is still a work in progress and I suspect more weird cases will show up as the root couse of the problem still hasent attended- capping the voltage is the easy immidiate (panic if you will) solustion.

 

[Klemc]

RMA mass will follow this informations !!

 

[Nam-A-M-D]

This is also effecting the 7950X as mine was fluctuating voltages from 1.390 - 1.492v set on auto. Even after the updated bios on the Asus x670 Gaming WIFI BIOS 1401 the voltage was still 1.350v and slightly higher under load on auto and standard. I have manually set the voltage to 1.225v and this is stable at 1.270v - 1.279v. the amount of damage to anyones CPU might not even show until later in its life, but the damage is irreversable and excessive strain has already been put on the CPU. It is nothing to do with overclocking or anything like that as it is an error on AMD's side and also the Motherboard Vendors lack of communication with each other. If you choose just running standard then you are still at risk of something from the excessive voltages which are set too high. Obviously we are going to see more come to light in the near future with more tests which could include other voltages not set correctly. Only time will tell. RMA process will only work if you have a broken CPU then you are in your right to do the Process as AMD stated they will honor.

 

[R0H1T]

Not everyone watches Steve, also you can avoid it by checking the voltages on boot. It's still poor from the manufacturers, especially MB like Asus, as even at default or "optimized" settings this type of behaviour is unacceptable!

 

[Nam-A-M-D]

With the price of the X670 platform you expect better from them.

 

[Klemc]

B650(E) too. It's 350 euro the board !

 

[AusWolf]

7000X6D? Wow!

 

[Nam-A-M-D]

With the price of the X670 platform you expect better from them.

The whole range for the 7000 series is costly i should have said. cheers

View attachment 293811

7000X6D? Wow!

Even the flustration is effecting the Web team now. That processor must be the Dimentional edition for aliens

 

[AusWolf]

Even the flustration is effecting the Web team now. That processor must be the Dimentional edition for aliens

That would make sense. I thought it's just a CPU with two layers of 3D cache.

Right... I've just updated my BIOS to version 1.82 whith the voltage limiter. Everything is exactly as it was before, 1.2 VSoC with EXPO, but my idle power consumption went from 20-22 W to 24-25 W and my idle CPU temp rose by about 5-6 °C, and I can't figure out why. So far, I'm not happy.

 

[Klemc]

That would make sense. I thought it's just a CPU with two layers of 3D cache.

Right... I've just updated my BIOS to version 1.82 whith the voltage limiter. Everything is exactly as it was before, 1.2 VSoC with EXPO, but my idle power consumption went from 20-22 W to 24-25 W and my idle CPU temp rose by about 5-6 °C, and I can't figure out why. So far, I'm not happy.

I readIt's chaotics BIOS updates !!

 

[trparky]

Well, at least when it comes to my system, I'm... mostly in the clear since I don't have an ASUS board. (Whew!) That's a load off my mind. However, even Gigabyte (which I have a Gigabyte board) isn't without sin here since, at least according to Steve, they had some kind of bug in BIOS version F5a that caused the SOC voltage to not be reset to default settings even if you reset the whole enchilada back to recommended settings in BIOS (oopsies). BIOS version F5a has since been pulled from their web site and has been replaced by version F5b; hopefully that SOC voltage reset bug that Steve mentioned, and I alluded to earlier in this sentence, was fixed (crossing fingers here) in version F5b.

I concur with Steve here, the AM5 platform when it comes to AMD communicating with their motherboard vendors on what's safe and what's not safe has resulted in an epic clusterfuck. There's a lot of room for blame here on AMD and their board partners.

I look forward to their side piece that Steve suggested will be coming soon because as he said, the video that was released recently was getting to be too long. I'm also looking forward to the full failure analysis that apparently will be weeks away due to them having to send the chip and motherboard off to a third-party failure analysis lab.

 

[wheresmycar]

View attachment 293811

7000X6D? Wow!

Don't be surprised, ASUS used to run with that slogan "in search of incredible".... looks like they've found it: 7600X6D (special edition: made in Jonney Shihs back garden lol)

View attachment 293799
View attachment 293800

At some point i was strongly considering moving up to AM5 but held off for a pending 5800X3D on a free-handed spare AM4 build. Although had i opted for AM5, i would have waited for some time to avoid early adopter snags and hopefully discounted platform costs. This ones more than just a "snag" - hope none of you AM5 subscribers have been affected.

As always, a big thumbs up to Steve@GN, great work!!

As for AMD/partners in crime... this level of imprudence is unacceptable. There's too much focus on tiny incremental advances to come out on top and less focus on stock fine-tuning. 7000-series and its efficiency was/is a big winner in my eyes but this mishap somewhat leaves a sour taste in my mouth. Although, optimistically speaking, i'm sure these concerns will be quickly resolved with upcoming mobo patches.,

 

[Klemc]

View attachment 293811

7000X6D? Wow!

Asus guys will also apply double to all BIOS and DOCP settings ???

 

[AusWolf]

Well, at least when it comes to my system, I'm... mostly in the clear since I don't have an ASUS board. (Whew!) That's a load off my mind. However, even Gigabyte (which I have a Gigabyte board) isn't without sin here since, at least according to Steve, they had some kind of bug in BIOS version F5a that caused the SOC voltage to not be reset to default settings even if you reset the whole enchilada back to recommended settings in BIOS (oopsies). BIOS version F5a has since been pulled from their web site and has been replaced by version F5b; hopefully that SOC voltage reset bug that Steve mentioned, and I alluded to earlier in this sentence, was fixed (crossing fingers here) in version F5b.

I concur with Steve here, the AM5 platform when it comes to AMD communicating with their motherboard vendors on what's safe and what's not safe has resulted in an epic clusterfuck. There's a lot of room for blame here on AMD and their board partners.

I look forward to their side piece that Steve suggested will be coming soon because as he said, the video that was released recently was getting to be too long. I'm also looking forward to the full failure analysis that apparently will be weeks away due to them having to send the chip and motherboard off to a third-party failure analysis lab.

As an MSi owner (thank heavens I didn't wait for that lovely-looking Asus TUF board, and pulled the trigger on the MSi Pro for faster delivery), I'm gonna be waiting for further BIOS updates. Everything was fine with version 1.50, but I'm not happy with the increased idle power consumption and heat with 1.82.

 

[MarsM4N]

Just watched GamersNexus's video, and holy moly, it's even worse than I guesstimated. Just pure unprofessionality on every end!

I mean, if they don't reach every customer out there to update their BIOS dead CPU's and some dead boards will just pile up. Esp. the part about "slow degradation" would be frightening if I had one of these. Even if you update your BIOS, your chip could be already taken a hit. Which could be even the reason for the strange reported idle power spikes.

View attachment 293811

7000X6D? Wow!

I bet their BIOS'es look the same. Full of typo's.

As an MSi owner (thank heavens I didn't wait for that lovely-looking Asus TUF board, and pulled the trigger on the MSi Pro for faster delivery), I'm gonna be waiting for further BIOS updates. Everything was fine with version 1.50, but I'm not happy with the increased idle power consumption and heat with 1.82.

Lower the SOC voltage manually and keep the lowest stable setting. Some folks reported they could drop the power consumption by up to 20W.

 

[AusWolf]

Lower the SOC voltage manually and keep the lowest stable setting. Some folks reported they could drop the power consumption by up to 20W.

My VSoC is exactly the same as it was before: 1.2 V, but my idle power and temp are still a bit higher than they were, god knows why. Thankfully, nothing has changed under load, so I'm not gonna bother it for now.

 

[Klemc]

I'm not sure about assumption, but, each time i see somebody with problems on AM5, it lists GSkill RAM in the specs !!

 

[corners]

How? AMD has been insistent that X3D chips shouldn't be OC'd, and even included a warning that OC related damage wasn't covered under warranty (at least, with the 5800X3D).

So it's not really a surprise that OC'ing an X3D chip hard would end up frying the more voltage sensitive CPU.

Don't all cpus and gpus have that same warning about oc damage not being covered?

 

[trparky]

Don't all cpus and gpus have that same warning about oc damage not being covered?

Yes, but the crux of the problem that we're facing here is that EXPO (and XMP) memory profiles have never been pushed so heavily as part of the primary product advertising. AMD even features it front and center in much of their core marketing material so it's only natural that one would think that AMD would be OK with this. Why market something so heavily and in your face if they're just going to turn around and tell you that if you do what they're shoving in your face (via marketing material) they're going to basically tell you that you're shit out of luck if something goes wrong?

I can understand overclocking the chip itself and making it go faster than the base clock but memory? It's so damn easy to do these days that even the biggest PC building n00b will know how to enable it.

I imagine that things are going to be changing from here on out when it comes to warranties.

 

[mikehcl]

Hi, I upgraded to the latest F10C bios my Gigabyte X670E Aorus Master, Ryzen 7950x3d and EXPO enabled, yesterday after 5 minutes of testing on Cyberpunk and 15 minutes of gaming on world of warships I checked Hwinfo64 finding these crazy voltages! But that's not possible come on, the CPU should be dead instantly and anyway even the average voltages are strange, if sffectively or reached such high peaks shouldn't the averages be higher?

 

[danc]

Asus clears up gamernexus SOC measurement FUD

auto translate english.

A pity youtube does not redact the past views, GN ran away with a big income using lazy research and too smug to apologise i betcha

 

[Klemc]

So, 1616 with AGESA 1.0.0.7A (alpha, not even beta now !?) is verry good or should i use 1413 first ?.. i will mount my PC friday and use flashback !

 

[rv8000]

Asus clears up gamernexus SOC measurement FUD

auto translate english.

A pity youtube does not redact the past views, GN ran away with a big income using lazy research and too smug to apologise i betcha

That’s GN for ya, specifically Steve. That man takes any chance he gets to be a sarcastic wisecrack.

ASUS definitely was in the wrong with how things were handled, but Steve could’ve been a lot more helpful about how they approached the content and recommendations to users.