9800X3D Failures/Deaths in the UK?

[mikemod]

My 9800X3D and Asrock B850 Pro RS have now been running over 100 days now with no issues.
I put on bios 3.25 via flashback before putting the processor in and have kept up to date to the current version 3.4

However as I'm running 3440x1440 on a 5080FE GPU, it seems to be asleep most of the time, the 1% lows in games with everything maxed aren't much different from the frame rate.

Of all the other forums I am on the only failures I see are on Reddit and it seems to be a race to see who can put the "Another one bites the dust" Meme up.

 

[sastusbulbas]

It sort of has, there are multiple videos and articles on it from big names.

I think the difference is that if someone is asking for CPU recommendations its easier to say "buy AMD" as they're fast, have future upgrade paths and as long as you avoid combination X then it'll be fine vs Intel 13th/14th where it's a dead platform with an unknown combination that *might* one day kill your CPU

Big names lol, no just idiots.

The latest basically saying oh noes best not buy Asrock waaawaaa.

When someone can produce a video but apparently lacks the IQ to Google the subject matter it can say a lot about the YouTube content creator.

It's easy to work out it's a rare problem when you simply look at figures online. From over 2 million AM5 Asro motherboards and a 46% majority for many AM5 sales being X3D, but just over 100 CPU X3D failures all related to bios 3.15 that was being worked on when Gamers Nexus went to see Asrock, around the time Asrock released 3.25 to address the bios issue which was again updated later.

It literally works out to be an error in the decimals figure wise.

Meanhile 99.999 % of Asrock AM5 users with 9000 series X3D CPUs have no clue there is an issue. (Figure may not reflect actual percentage)

I think it was Kleo Klan on YouTube who created a settings video for Asrock users with X3D before 3.25 bios, and that has been the go to and possibly why Asrock themselves tweaked bios after 3.30.

Sadly we may never know, but we do know 3.15 was too aggressive.

 

[Rroff]

Meanhile 99.999 % of Asrock AM5 users with 9000 series X3D CPUs have no clue there is an issue. (Figure may not reflect actual percentage)

Return rates at retailers who publish details vary between 0.56% and 1.1% for the 9800X3D, I know one retailer has nearly 300 pieces (out of well into the 10s of thousands sold) returned with the majority having symptoms consistent with this failure.

 

[LtMatt]

When you consider the number of CPUs sold both, so far, are very much edge cases, the evidence suggests that it is happening on other boards not just ASRock but with a far lower chance of happening - I think someone mentioned something like 1% and 4% of the failures consistent with this issue are on Gigabyte and MSI boards. Which then leads to the bit I find amusing in that the Intel issue has generated a lot of stink but the AMD one hasn't.

The Intel one created stink because we have actual game developers and PC builders writing articles about the reports they get of an high number of Intel CPU failures. That’s how this first came to light.

 

[Rroff]

The Intel one created stink because we have actual game developers and PC builders writing articles about the reports they get of an high number of Intel CPU failures. That’s how this first came to light.

One of the original sources was RAD Game Tools - who has only ever seen small numbers of CPUs showing these symptoms and due to the nature of what they do and deep presence in the PC market are more on the pulse of it than anyone. The other main source was someone anonymous on a game dev forum who was coming out with hardware technical information which only could have come internally from Intel (or maybe a competitor) - there is zero chance anyone external to Intel did that level of debugging without a LOT of resources/money and ready access to a failure lab.

The couple of game developers and 1-2 related companies coming out of the woodwork with mass failures were never verified, not replicated by other companies doing similar stuff and when the likes of Gamers Nexus started investigating they came up short of any hard evidence - which should be pretty easy to find if companies were seeing 20% let alone 50-100% failures.

There was some PC builders reporting high return rates but that was only after the stink got kicked up and people started panic returning them due to the hysteria.

IMO, unless longer term effects kick in, it is a pretty sure bet disgruntled ex-Intel employees from the layoffs around that time have taken some real issues but blown them up out of all proportions to try and get back at Intel and a load of people have [willingly] fallen for it. And a good number of the CPUs which have exhibited the known degradation weakness have likely done so due to aggressive Asus motherboard optimisations - which isn't exactly a new story.

 

[Rroff]

GN investigating:

Weren't able to replicate it themselves.

 

[G J]

GN investigating:

(snipped the video)

Weren't able to replicate it themselves.

If I'm right they could not replicate it themselves when it was Intel but that did not stop him though I knew just from the video title that this was going to be a lukewarm kid gloves approach and GN did not disappoint. Seems to have also missed that this is not just a 9800X3D issue with Asrock, its a number of Zen 4 and Zen 5 cpus also.

 

[Epburpee]

I just bought an Asus ROG board for my new AMD build, yikes

 

[deviation]

So am I safe to assume that whatever I buy now, will be fine, given the BIOS updates since? Was it mainly AsRock and a few MSI's, and another brand? I forget who?
Was the issue purely down to a BIOS issue with giving PBO too much juice, or what exactly?
I'll be honest, I didn't look too much into it, as I figured a BIOS update would have fixed it by the time I got round to building a new system anyway

Anyone?

 

[welshrat]

Anyone?

Honestly at this stage we don't know. There does seem to be a lot less reports on 3.40 onwards, currently none on 3.50 yet, so fingers crossed we may just be down to normal failure rates. If it was me now I would buy MSI carbon board or even wait for what appears with Zen 6. However as I own an Asrock Nova X870E I just keep my fingers crossed, update the BIOS and enjoy my PC.

 

[jigger]

One of the original sources was RAD Game Tools - who has only ever seen small numbers of CPUs showing these symptoms and due to the nature of what they do and deep presence in the PC market are more on the pulse of it than anyone. The other main source was someone anonymous on a game dev forum who was coming out with hardware technical information which only could have come internally from Intel (or maybe a competitor) - there is zero chance anyone external to Intel did that level of debugging without a LOT of resources/money and ready access to a failure lab.

The couple of game developers and 1-2 related companies coming out of the woodwork with mass failures were never verified, not replicated by other companies doing similar stuff and when the likes of Gamers Nexus started investigating they came up short of any hard evidence - which should be pretty easy to find if companies were seeing 20% let alone 50-100% failures.

There was some PC builders reporting high return rates but that was only after the stink got kicked up and people started panic returning them due to the hysteria.

IMO, unless longer term effects kick in, it is a pretty sure bet disgruntled ex-Intel employees from the layoffs around that time have taken some real issues but blown them up out of all proportions to try and get back at Intel and a load of people have [willingly] fallen for it. And a good number of the CPUs which have exhibited the known degradation weakness have likely done so due to aggressive Asus motherboard optimisations - which isn't exactly a new story.

Nope, Intel just dropped the mother of all changers ( actually two in succession) and dealt with the issues in a typical monopolistic corporation way… Intel is rightly coughing up the cake now though and making people square, even paying for platform upgrades in some cases.

 

[welshrat]

Nope, Intel just dropped the mother of all changers ( actually two in succession) and dealt with the issues in a typical monopolistic corporation way… Intel is rightly coughing up the cake now though and making people square, even paying for platform upgrades in some cases.

Yeah I agree, this is very much an Asrock issue rather than AMD so very much apples and oranges. (typing this as I cross my legs and fingers)

 

[Rroff]

Honestly at this stage we don't know. There does seem to be a lot less reports on 3.40 onwards, currently none on 3.50 yet, so fingers crossed we may just be down to normal failure rates. If it was me now I would buy MSI carbon board or even wait for what appears with Zen 6. However as I own an Asrock Nova X870E I just keep my fingers crossed, update the BIOS and enjoy my PC.

It is hard to be sure with what little information is out there but there does seem to be a small number of 9000 series CPUs, potential a problem batch or batches, which have less safety margin than designed, which is then exacerbated on ASRock boards - it isn't just ASRock boards this failure is being seen on but it is relatively rare on other boards.

 

[Rroff]

Nope, Intel just dropped the mother of all changers ( actually two in succession) and dealt with the issues in a typical monopolistic corporation way… Intel is rightly coughing up the cake now though and making people square, even paying for platform upgrades in some cases.

Nope what? Fact is there are more documented cases of these 9000 series failures than there is 13th/14th gen ones and Asus motherboards seem to be a significant factor in exacerbating the issue with the 13th/14th gen as well.

 

[welshrat]

It is hard to be sure with what little information is out there but there does seem to be a small number of 9000 series CPUs, potential a problem batch or batches, which have less safety margin than designed, which is then exacerbated on ASRock boards - it isn't just ASRock boards this failure is being seen on but it is relatively rare on other boards.

Agreed it's not just ASRock but I think we are in the realms of acceptable failure rates on other boards, especially now nodes are so small and we have all the high-κ dielectrics etc in silicon to mitigate all the issues at this size, I imagine intel, Qualcomm etc must have similar rates. Also a huge amount of X3D chips sold for obvious reasons.

Be nice to get an end to this from the ASrock side though as its a constant fear for me.

 

[jigger]

Nope what? Fact is there are more documented cases of these 9000 series failures than there is 13th/14th gen ones and Asus motherboards seem to be a significant factor in exacerbating the issue with the 13th/14th gen as well.

It’s not disgruntled former Intel employees spreading rumours about Intels chips failing. The failures are down to Intel cocking up very badly and trying to hide it. It was well understood within the industry that Intel had a serious issue with its chips. Even now happy to use strong arm tactics against its customers and even its own employees that speak about it. The former Nvidia guys have really changed the culture within Intel for the worse.

 

[Rroff]

Agreed it's not just ASRock but I think we are in the realms of acceptable failure rates on other boards, especially now nodes are so small and we have all the high-κ dielectrics etc in silicon to mitigate all the issues at this size, I imagine intel, Qualcomm etc must have similar rates. Also a huge amount of X3D chips sold for obvious reasons.

Be nice to get an end to this from the ASrock side though as its a constant fear for me.

The failure rates overall are very low, even at the peak of the reports, but it isn't a normal failure mechanism like you might get within "acceptable failure rates" if you get what I mean. It is pretty rare, outside of the 7000 series SoC voltage issue and this problem, for CPUs to malfunction in this fashion. It is actually pretty rare for CPUs, when run at stock settings, to fail at all.

 

[Rroff]

It’s not disgruntled former Intel employees spreading rumours about Intels chips failing. The failures are down to Intel cocking up very badly and trying to hide it. It was well understood within the industry that Intel had a serious issue with its chips. Even now happy to use strong arm tactics against its customers and even its own employees that speak about it. The former Nvidia guys have really changed the culture within Intel for the worse.

Maybe maybe not on the ex-Intel employees, the timing is a bit suspicious and too many facts tend to lean that way when you look at it objectively. The fact is the failures in the real world are nothing like a small number of sources were claiming - most of which were unable or unwilling to back it up when people like GN went investigating and we have categorically not seen that scale of failures in the wild including in other companies doing the exact same thing.

Reality is if you look at the relative megathreads for these issues on various forums and Reddit there are far more documented cases and/or reported cases consistent with this AMD 9000 series issue than there is 13th/14th gen degredation - you have to go back 3 months for the last confirmed case on the Reddit megathread and then 8 months after that (and that has been the trend since the issue came to light), whereas the AMD megathread is getting reported cases weekly, sometimes close to daily. I've not kept up with it on Reddit but just noticed there are some interesting graphs of reported issues here https://www.reddit.com/r/ASRock/comments/1mvgndh/9000series_cpu_failuresdeaths_megathread_2/ far too small numbers to say anything definite but seems to be a high batch correlation which is also held up by information I have from elsewhere.

 

[jigger]

The failure rates overall are very low, even at the peak of the reports, but it isn't a normal failure mechanism like you might get within "acceptable failure rates" if you get what I mean. It is pretty rare, outside of the 7000 series SoC voltage issue and this problem, for CPUs to malfunction in this fashion. It is actually pretty rare for CPUs, when run at stock settings, to fail at all.

The failure rates of 13 and 14 gen Intel have been massive. Massive to the point Intels lead time on replacements was two quarters or more and Intel had to offer complete platform upgrades or settle financially with customers no longer prepared to take the risk or manage the increase hosting/servicing costs on the systems.

Anyway, I think we’ve been here before.

 

[Rroff]

The failure rates of 13 and 14 gen Intel have been massive. Massive to the point Intels lead time on replacements was two quarters or more and Intel had to offer complete platform upgrades or settle financially with customers no longer prepared to take the risk or manage the increase hosting/servicing costs on the systems.

Anyway, I think we’ve been here before.

Except there is no evidence that the degradation failures have been massive - the vast majority of replacements were people panic RMAing their CPUs and/or out of an abundance of caution because they were concerned about second hand value when they came to upgrading, often switching to AMD, in reaction to the news not because of their CPU actually experiencing failures. It was largely customers unwilling to take a perceived risk rather than actual failures. That RMA activity didn't exist at scale before the news broke and quickly returned to normal after the news broke - not indicative of a long term problem.

As I've said before there are sources with their finger on the pulse like RAD Game Tools who'd be the first ones reporting if these failures were at the magnitude a small number of sources were claiming to Level1Techs - which was never verified or seen in the wild by other companies doing similar.

You can keep saying these issues are massive all you want but the evidence does not support your claims.