Alderon Games claims that substantial numbers of Intel 13th Gen and 14th Gen chips are defective

[OneSorcerer]

Looks like we have an official culprit, apparently it's a microcode issue causing the CPU voltage curve to be elevated beyond intended range. Similar to the AGESA issue that caused Ryzen to catch fire I reckon

Intel uncovers cause of 13/14th Gen Raptor Lake CPU instability - OC3D

Intel has announced that a fix is coming for the instability issues affecting their 13th/14th gen Raptor Lake CPUs.

overclock3d.net

C/c @btarunr

I need receipts. Another long time of observation in the wild before I consider an upgrade from 12th gen

 

[DAPUNISHER]

I need receipts. Another long time of observation in the wild before I consider an upgrade from 12th gen

Smart. We have no idea what the final outcome will be as of yet. Will performance be impacted negatively? What's going to happen when flaky CPUs with the new mc patch are getting less voltage? ARL will be here before this all shakes out?

 

[thesmokingman]

This aged worse than warm milk in this central Florida sun.

That's what you call denial.

 

[ShrimpBrime]

Erm...no..that comment of yours has zero bearing to the thread and only questions his credibility. You do realise that in the topic of degradation, Wendell has been going into far more depth than what a mere VID table which, if you know about their conversation, is something found in certain desktop boards Ai tweaker menu or something like that while Wendell mostly deals with server CPU's. I really don't see the point of your post but I'll stop at that.

You live in a parallel universe or are in some sort of unhealthy denial it seems. A major system integrator says there's 12% defect rate, a massive number of server farms and gave devs are all reporting issues and a few people even in this thread posted about their issues. He'll, even Intel has acknowledged that there are issues but in your books there's no issue still because you haven't seen any.

Maybe wake the f up?

You've missed every point I've laid onto the table and focused solely on your own.

FIRST and foremost, the mislead title, and I quote...

"DEFECTIVE"

 

[N/A]

Yeah right, more likely kiss goodbye to the 6.2 GHz boost on 1-4 t. And the 5-16 t as well. Go back to 12900 KS boost of 5.2-5.5 and maybe add 300 MHz that one 7nmm refinement can probably deliver safely at the same voltage, but for more than that you'll need more voltage and that is as intended.

 

[DAPUNISHER]

The plot thickens - https://www.reddit.com/r/intel/comments/1e9mf04

Short answer: We can confirm there was a via Oxidation manufacturing issue (addressed back in 2023) but it is not related to the instability issue.

Long answer: We can confirm that the via Oxidation manufacturing issue affected some early Intel Core 13th Gen desktop processors. However, the issue was root caused and addressed with manufacturing improvements and screens in 2023. We have also looked at it from the instability reports on Intel Core 13th Gen desktop processors and the analysis to-date has determined that only a small number of instability reports can be connected to the manufacturing issue.

For the Instability issue, we are delivering a microcode patch which addresses exposure to elevated voltages which is a key element of the Instability issue. We are currently validating the microcode patch to ensure the instability issues for 13th/14th Gen are addressed.

I'll be on the thread for the next couple of hours trying to address any questions you folks might have. Please keep in mind that I won't be able to answer every question but I'll do my best to address most of them.

Thanks

Lex H. - Intel

I suspect this is to get out in front of GN having sent CPUs off to the FA lab. Just in case.

 

[MikeSnow]

Unless YOU (and others) have definitive PROOF of said "defects" causing RMA, then STFU.

OK, go and ask Intel for proof then. Because they just admitted it:

Based on extensive analysis of Intel Core 13th/14th Gen desktop processors returned to us due to instability issues, we have determined that elevated operating voltage is causing instability issues in some 13th/14th Gen desktop processors. Our analysis of returned processors confirms that the elevated operating voltage is stemming from a microcode algorithm resulting in incorrect voltage requests to the processor.

The only question that remains is if that "incorrect voltage" (which in my opinion qualifies as a "defect") was intentional, or an error. My guess is that their yields were not high enough for the high end parts, so they pushed the voltages a bit higher to be able to reach higher frequencies that they could not have achieved with safer voltages, so that they could sell even inferior silicon as high-end. Basically, I think that they gambled that nothing would go wrong by overvolting a bit, or that few people would be affected, and that gamble didn't pay off.

But, even assuming that they didn't intentionally do this, and it was just some kind of microcode bug causing the increased voltage, the problem now is that many of us have degraded CPUs that have been exposed to that excessive voltage for many months. That makes them unstable even with the current increased voltages, sometimes unusable without underclocking, and we have to manually reduce the peak frequencies to make them work properly. If Intel pushes a microcode update that reduces the voltages to safe levels, we'll probably have to reduce the peak frequency even further to keep them stable. Worst case, we might not be even able to reach the BIOS anymore to be able to underclock them.

In conclusion, I doubt the microcode update will make things more stable for those of us that are already affected by the degradation and stability issues. Possibly the other way around. But it will probably slow down the degradation, and fewer CPUs will degrade enough to become unstable, which is good news for those that currently do not experience the issues.

 

[Hecate91]

Support homie is like - keep upping the voltage until it stabilizes.

His employer - " microcode patch which addresses the root cause of exposure to elevated voltages"

Big OOF! Anyone that followed that support advice needs to RMA, your CPU is broke

I guess proof from Nvidia, a major SI having problems with cpu's failing, game devs having data on which CPU's are failing, game servers at Intel baseline spec with way under XMP DRAM settings still crashing isn't enough for them.
It's either denial or paid damage control, i've seen comments on another forum some oc'ers think they know better than GN and Wendell who have asked these companies firsthand on what their issues are.
And applying more voltage into it isn't going to fix it when the voltage might be part of the reason why 14th gen i9's have degraded to the point of failure after just a year and a half. Maybe more voltage fixes it for someone doing quick OC benchmark runs, but it isn't a real fix for anyone really using their system or leaving it running 24/7.

I need receipts. Another long time of observation in the wild before I consider an upgrade from 12th gen

I agree there needs to be additional proof, although I doubt microcode is going to fix degradation, if the microcode is lowering clocks then IMO a class action needs to happen as the cpu's with new microcode applied won't run at the performance Intel have claimed.

 

[Klemc]

Intel finally announces a solution for CPU crashing and instability problems — claims elevated voltages are the root cause; patch coming by mid-August [Updated]

Patch is on the way.

www.tomshardware.com

Never would have been so fast without all the talks.

It's like a day1 patch, but 1.5 year later.

 

[Dr. Dro]

I need receipts. Another long time of observation in the wild before I consider an upgrade from 12th gen

Keep an eye on my i9 KS owners club thread. I'll be sure to let you guys know if something wrong happens

 

[Klemc]

Keep an eye on my i9 KS owners club thread. I'll be sure to let you guys know if something wrong happens

You'll have to buy another one brand new to compare with the latests BIOS fix results. Please.

Just to know if it aged badly tho.

 

[Dr. Dro]

You'll have to buy another one brand new to compare with the latests BIOS fix results. Please.

Just to know if it aged badly tho.

No need, my CPU is not defective and has never exhibited any symptoms, probably because it's a KS and I have always run it undervolted.

 

[DAPUNISHER]

I guess proof from Nvidia, a major SI having problems withcpu's failing, game devs having data on which CPU's are failing, game servers at Intel baseline spec with way under XMP DRAM settings still crashing isn't enough for them.
It's either denial or paid damage control,

Not a chance. Anyone working for or with Intel, even a 3rd party social media agent, would conduct themselves with far better decorum. It has went beyond the backfire effect to witnessing someone's cheese slide of the cracker. It is rather cringe.

We should respect they are having a bad mental health day and ignore them.

As I observed elsewhere; this is becoming the old commercial - how many licks does it take to get to the tootsie roll center of a tootsie pop? The world may never know.

No need, my CPU is not defective and has never exhibited any symptoms, probably because it's a KS and I have always run it undervolted.

Everyone's data is that Alder Lake is rock solid.

 

[Sunny and 75]

On avg 3D-chips can get an avg boost of around 5% with ram tuning. Some games get 10%+. Others 0%. Non-3D get 10-12%.

I would advice all to tune ram anyways. 80% of the gain is from tuning RFC and REFI alone. Just set ram to 6000, REFI to 50000 (65535 is posdible but can lead to file corruption if temps on dimms rise above 50C) and RFC to 500 if you have Hynix M or 400 if you have A.

Does ram tuning increase the Vcore?

 

[Vario]

Does ram tuning increase the Vcore?

Shouldn't, usually relevant to the CPU are stuff like VTT or VCCIO or VCSSA.

 

[ShrimpBrime]

OK, go and ask Intel for proof then. Because they just admitted it:



The only question that remains is if that "incorrect voltage" (which in my opinion qualifies as a "defect") was intentional, or an error. My guess is that their yields were not high enough for the high end parts, so they pushed the voltages a bit higher to be able to reach higher frequencies that they could not have achieved with safer voltages, so that they could sell even inferior silicon as high-end. Basically, I think that they gambled that nothing would go wrong by overvolting a bit, or that few people would be affected, and that gamble didn't pay off.

But, even assuming that they didn't intentionally do this, and it was just some kind of microcode bug causing the increased voltage, the problem now is that many of us have degraded CPUs that have been exposed to that excessive voltage for many months. That makes them unstable even with the current increased voltages, sometimes unusable without underclocking, and we have to manually reduce the peak frequencies to make them work properly. If Intel pushes a microcode update that reduces the voltages to safe levels, we'll probably have to reduce the peak frequency even further to keep them stable. Worst case, we might not be even able to reach the BIOS anymore to be able to underclock them.

In conclusion, I doubt the microcode update will make things more stable for those of us that are already affected by the degradation and stability issues. Possibly the other way around. But it will probably slow down the degradation, and fewer CPUs will degrade enough to become unstable, which is good news for those that currently do not experience the issues.

Yes, and didn't I say could be a ME issue, bios setting and other? Didn't Intel also exclaim to blame MB manufacturers for this? I also gave an example of 2 similar issues concerning of 1.5v reported by bios? Did I not also say the wrong micro code installed posted the cpu over 1.7v.

Pretty sure I mentioned a lot of these things as a possibility not so much it being the physical hardware it's self, but the settings being used.

I also mentioned that Raptor Lake has been out since 2022. And this was a fairly "newer" issue, but not to condemn the processors themselves? Did I also not mention there may very possibly be a fix for this in the future??

I just wanted to have these things presented on the table to be viewed, looked at and respected as a possibility.

Cause ME firmware is a HUGE player in how the cpu is operated.

The software needs to match the firmware. They'll have to work with MS to force ME FIRMWARE updates to marry the software to. This is actually an issue in big class 8 trucking. Get a lot of funky things that can happen when certain powertrain modules don't have the correct version matching chassis modules. (As only an example)

I digress. Be well.

 

[Klemc]

Yes, and didn't I say could be a ME issue, bios setting and other? Didn't Intel also exclaim to blame MB manufacturers for this? I also gave an example of 2 similar issues concerning of 1.5v reported by bios? Did I not also say the wrong micro code installed posted the cpu over 1.7v.

Pretty sure I mentioned a lot of these things as a possibility not so much it being the physical hardware it's self, but the settings being used.

I also mentioned that Raptor Lake has been out since 2022. And this was a fairly "newer" issue, but not to condemn the processors themselves? Did I also not mention there may very possibly be a fix for this in the future??

I just wanted to have these things presented on the table to be viewed, looked at and respected as a possibility.

Cause ME firmware is a HUGE player in how the cpu is operated.

The software needs to match the firmware. They'll have to work with MS to force ME FIRMWARE updates to marry the software to. This is actually an issue in big class 8 trucking. Get a lot of funky things that can happen when certain powertrain modules don't have the correct version matching chassis modules. (As only an example)

I digress. Be well.

OKay but after 1.5 years it became an hardware failure anyway !.. software too late, hardware broken.

 

[ShrimpBrime]

Shouldn't, usually relevant to the CPU are stuff like VTT or VCCIO or VCSSA.

My B660-G with XMP enabled forces cache to 5ghz. Reported Vccsa as high as 1.40v pretty normal on an Asus board, but it sure does make it run hotter. So does the Asus Performance enhancement switch. First thing I do, even as an overclocker is turn that straight disabled. Let me fry my own Cpu please, thanks.

OKay but after 1.5 years it became an hardware failure anyway !.. software too late, hardware broken.

I guess you got me there!

 

[Hecate91]

Not a chance. Anyone working for or with Intel, even a 3rd party social media agent, would conduct themselves with far better decorum. It has went beyond the backfire effect to witnessing someone's cheese slide of the cracker. It is rather cringe.

We should respect they are having a bad mental health day and ignore them.

As I observed elsewhere; this is becoming the old commercial - how many licks does it take to get to the tootsie roll center of a tootsie pop? The world may never know.


Everyone's data is that Alder Lake is rock solid.

Yeah,good point. I would expect someone working with Intel or at least getting free hardware to be acting more professional with their replies. The people still defending Intel isn't a good look.
There seems to be so many things going on at the same time with how 13th and 14th gen are degrading, I want to see proof this microcode works though IMO it took Intel way too long to come out with a PR statement after they allowed motherboard makers to push voltages over the limit in order to win benchmarks.
So far Alder Lake is trouble free, being on 10nm may help, the ring bus isn't being pushed nearly as hard with fewer e-cores.

OKay but after 1.5 years it became an hardware failure anyway !.. software too late, hardware broken.

It probably took 1.5 years for degradation to happen to the point of instability or failure, these cpu's should last well past the 3 year warranty without any significant degradation at default specs.
Finally a statement from Intel, but Intel didn't answer whether or not they will be honoring RMA's for degraded cpu's which IMO is of utmost importance as I've seen comments of Intel denying RMA's telling customers to run the baseline settings.
Also Intel knew about the oxidization issue since 2023, yet didn't make a statement until now so who knows how many cpu's are actually affected.

 

[pressing on]

Interesting comment by wccftech "...From what we have heard, 1/3rd of all Intel Raptor Lake CPUs that have been shipped are Core i9-13900K or 14900K units so that's roughly around 40-60 million units (estimates from Mike Bruzzone)." That would make total Raptor Lake shipments of 120-180 million to date. The August microcode update is going to have to work...

 

[Vya Domus]

It's very strange for more voltage to be causing instability, a lot of voltage might cause degradation which then leads to instability, strange explanation. I feel like this might be related to what's happening but not the actual culprit.

 

[Klemc]

It's very strange for more voltage to be causing instability, a lot of voltage might cause degradation which then leads to instability, strange explanation. I feel like this might be related to what's happening but not the actual culprit.

Fixed in 24/48H afyer 1.5 year of nothing, so...

 

[Darmok N Jalad]

It's very strange for more voltage to be causing instability, a lot of voltage might cause degradation which then leads to instability, strange explanation. I feel like this might be related to what's happening but not the actual culprit.

I think that the chip's exposure to excess voltage has permanently damaged it, and so what once was stable is no longer stable. This is often what we see with chips that are overvolted--when done to extreme levels, the max-overclock becomes unstable over time, and the overclock has to be reduced. Eventually that chip might not even run at factory settings. It's essentially burned up. This problem didn't come to a head until a year and a half into Raptor Lake. The 13 series took longer to start failing because they probably didn't have as extreme of a TVB setting as the 14 series. To me, this looks like reckless engineering from Intel, not some "oops, I got the decimal in the wrong place" mundane detail. I'm actually curious what this microcode update will do. I'd be surprised if we don't see a performance penalty. Even a 5% loss can make all the difference on "winning" and setting higher MSRPs.

What I suspect will happen is that Intel will roll out this microcode update, and it will successfully keep some chips from failing, like those that haven't been in service too long, or stressed too much. It might even "cure" the stability issues of some chips that aren't too far gone. However, I think some chips are still going to crash, and that's when Intel will "stand by its customers" and quietly offer warranty replacements. It's the lowest cost solution to them, and they can say they are honoring their warranties without having a full-blown recall.

 

[Vya Domus]

I think that the chip's exposure to excess voltage has permanently damaged it, and so what once was stable is no longer stable. This is often what we see with chips that are overvolted--when done to extreme levels

Yeah but that's the thing, extremes levels, how overvolted could these chips have been ? I doubt an extra 20mv would make any real difference, a lot of people on here overlock their CPUs 24/7, chip degradation is very rare.

I think there is still a manufacturing defect, made much worse by every increase in voltage, accelerating whatever it is that's wrong.

 

[Klemc]

Yeah but that's the thing, extremes levels, how overvolted could these chips have been ? I doubt an extra 20mv would make any real difference, a lot of people on here overlock their CPUs 24/7, chip degradation is very rare.

They are well cooled ?..! perhaps it changes the finality, users with airflow "or" letting the CPU go at higher temp have higher risk ?