Intel Statement on 13th and 14th Gen Core Instability: Faulty Microcode Causes Excessive Voltages, Fix Out Soon

[Dr. Dro]

You can apply this to every tech produced, most PC related hardware is "trying something new" when it first appears.

If I didnt buy from a company again after I (or friend) had problems, I would have blacklisted the following companies. There may be more, these are the ones I most easily remember.

AMD - FX degradation making stock unstable
Asus - Failed capacitors and unstable voltages in their "asus optimised" bios.
Asrock - Unsafe voltages in their bios when activating XMP and stock settings exceeding tjmax spec.
BenQ - Monitor with flawed display port (display wont wake up if turned off whilst using display port, system needs to be rebooted to wake up the port on the monitor).
Crucial - SSD shipped with flawed firmware.
EVGA - GPU shipped with unstable v/f curve out of the box.
Gigabyte - GPU shipped with unstable v/f curve out of the box on performance bios.
Kingston - Numerous SSDs failing.
Viewsonic - Failed monitor, and Monitor RMA switcheroo was a replacement with same fault.

I agree. Speaking for myself, I have positively horrible, terrible luck with Seagate. It's a 100% failure rate, every single Seagate HDD I have ever owned has malfunctioned and died within a year of ownership. Meanwhile, I've got at least two Western Digital drives that now exceed 50,000 hours of operation with SMART reporting perfect condition, the newest of these is a WD Blue WD10EZEX that is about 10 years old and has been on 24/7 on my server box.

Does that make Seagate horribad, avoid at all costs? I don't think so... I'm in an overwhelming minority, and I'm sure some people have had the same ill luck with WD drives dying on them. People just take their clubism to the extreme sometimes.

 

[OkieDan]

I can also offer a worse example: Sony.
I bought their Linkbuds S wireless earbuds in 2022 only for these to develop a battery discharge issue a week after my two year warranty period ended (both go from 100% to empty withing 15-30 minutes instead of usual ~8 hours).
Reading reviews and comments online there are many people facing the same issue with both the Linkbuds S and WF-1000 XM4 models produced and bought in 2022.

Yet Sony has not even acknowledged the issue nor provided any replacements for customers because in their eyes the warranty period for both products (for 2022 buyers at least) has ended and thus they feel they dont have to do anything.

A product failing a week after warranty ended feels like planned obsolescence...

I've avoided everything Sony since their DRM root kit.

 

[Sunny and 75]

Even if it wasn't the intention, my confidence has been shaken.

Hope they can restore some of that with the release of ARL and BTL for a thriving CPU market.

 

[Dr. Dro]

Hope they can restore some of that with the release of ARL and BTL for a thriving CPU market.

I am really looking forward to Bartlett Lake. I know it's a NEX product but I hope Intel will go ahead with a full-scale release of it. I'll buy one at launch.

 

[Sunny and 75]

I am really looking forward to Bartlett Lake.

The same.

 

[trparky]

Hope they can restore some of that with the release of ARL and BTL for a thriving CPU market.

I hope so as well, mainly because I don't want to see AMD become complacent in the market. When there's competition, the consumer wins.

 

[Sunny and 75]

When there's competition, the consumer wins.

That's the definition of a thriving market.

 

[ikjadoon]

There is no fix for Intel’s crashing 13th and 14th Gen CPUs — any damage is permanent

We got some answers from Intel, and more are on the way.

www.theverge.com

Ugly to disgusting. This needs a proper watchdog / regulatory investigation at this point.

• Intel confirms ANY 65W or higher 13th/14th CPU is susceptible to accelerated degradation.
• Intel will not recall any CPUs.
• Intel will not provide any manufacturing dates / serial number ranges for via oxidation.
• Intel (obviously) confirms the damage is permanent. It will never get better.

 

[close]

Probably not, or they wouldn't have done it in the first place.

"Quietly drop the voltage and performance when reviewers stop looking" would be actual fraud.

Noooo, it's fixing a bug in microcode and improving reliability. The way they're approaching this whole scandal reeks of fraud, especially since the message they passed on to the press and consumers ("just some microcode fixes") was very different to what they were already discussing internally about voltages and excessive oxidation prior to that. They withheld that information and if even one consumer's CPU ran out of warranty while Intel was yet again trying to bury the truth, that's already fraud enough for me. They now have a long tradition of trying to hide issues for just long enough for the CEO to dump some shares. But as I said elsewhere, as long as enough people and companies have more brand loyalty than sense, they will be treated by Intel accordingly.

 

[BoggledBeagle]

• Intel (obviously) confirms the damage is permanent. It will never get better.

We need the silicon healer!

 

[RatusNatus]

Realistically it won't. Worst case scenario it drops what, 100mhz on the pcores? Let's say 200? Heck let's make it 300 just for the sake. That's a 3% drop on the mt performance of the chip. It's so irrelevant you won't even notice.

Some Minecraft servers (14900k) were dying in 2 months of usage.
They reduced 500mhz and that made it last 6 to 8 months before failing.

This is the source...

 

[MikeSnow]

I was quite negative here about Intel this week, but my recent RMA experience has been flowless. I had stability issues with my 13900KF CPU for over an year, but I refrained from asking an RMA until this week, as I needed my computer for work and didn't want to wait a few weeks for a replacement.

Anyway, as this Saturday I'm leaving on vacation for two weeks, I decided that this week would be the best time to ask an RMA for it, especially in light of the latest revelations. I was hoping to get a replacement when I got back from vacation.

So, on Tuesday I made the RMA request, telling them that my CPU is unstable even with the latest BIOSes and the Intel defaults.

- 8 hours later, they were asking for the serial number, as I have given them an incorrect serial number initially.
- 4 hours later the RMA was approved, and I received an estimation of 5-7 business days to receive the replacement, from the time I submit the faulty unit.
- 1 hour later I received the DHL return label. The collection was scheduled for Thursday, 2 days later, but I decided not to wait for that.
- 3 hours later I handed over the faulty CPU at a local DHL service point
- 15 hours later the CPU started its travel towards Intel
- 22 hours later the CPU arrived at Intel
- 7 hours later they initiated shipping for the replacement CPU
- 2 hours later the CPU was on its way to me
- 3 hours later I received an email confirming they received the CPU and they need to validate it which would take 1 to 3 business days. Not sure what that email was about, since they already sent the replacement, probably just an automated message.
- 16 hours later I had a brand new 13900KF CPU on my desk.

Overall, the process took just a bit over 3 days, which I find almost unbelievable. I didn't get any pushback from Intel, and apparently when you send them a CPU that they know is very likely to be affected they don't actually do any validation, except that you actually sent them the CPU, and almost immediately send you a replacement.

So, while they might have been reluctant to accept RMAs in the past, it seems that right now they are actually doing the right thing regarding RMAs, and they are trying to make things as painless as possible for their customers.

While I can't totally forgive them for their past behavior, I have to give them a 10/10 for my recent experience.

The new CPU uses just 1.279 V in the BIOS, compared to the defective CPU, which used 1.447V. I left the BIOS on the Intel defaults with no XMP, and so far I don't have any stability issues and it reaches the maximum frequencies properly. But I won't do more extensive testing or stress it more until the microcode updates are released.

 

[Dr. Dro]

Some Minecraft servers (14900k) were dying in 2 months of usage.
They reduced 500mhz and that made it last 6 to 8 months before failing.

This is the source...

I set up a Minecraft server on my personal server (C2Q Q9505) last year and basically forgot about it. I just logged in there yesterday and everything was as I left, talk about time machine effect. Not good. A 14900K must be so outrageously powerful for Minecraft hosting that you probably need to host a 500-player server before it begins to challenge the chip

I was quite negative here about Intel this week, but my recent RMA experience has been flowless. I had stability issues with my 13900KF CPU for over an year, but I refrained from asking an RMA until this week, as I needed my computer for work and didn't want to wait a few weeks for a replacement.

Anyway, as this Saturday I'm leaving on vacation for two weeks, I decided that this week would be the best time to ask an RMA for it, especially in light of the latest revelations. I was hoping to get a replacement when I got back from vacation.

So, on Tuesday I made the RMA request, telling them that my CPU is unstable even with the latest BIOSes and the Intel defaults.

- 8 hours later, they were asking for the serial number, as I have given them an incorrect serial number initially.
- 4 hours later the RMA was approved, and I received an estimation of 5-7 business days to receive the replacement, from the time I submit the faulty unit.
- 1 hour later I received the DHL return label. The collection was scheduled for Thursday, 2 days later, but I decided not to wait for that.
- 3 hours later I handed over the faulty CPU at a local DHL service point
- 15 hours later the CPU started its travel towards Intel
- 22 hours later the CPU arrived at Intel
- 7 hours later they initiated shipping for the replacement CPU
- 2 hours later the CPU was on its way to me
- 3 hours later I received an email confirming they received the CPU and they need to validate it which would take 1 to 3 business days. Not sure what that email was about, since they already sent the replacement, probably just an automated message.
- 16 hours later I had a brand new 13900KF CPU on my desk.

Overall, the process took just a bit over 3 days, which I find almost unbelievable. I didn't get any pushback from Intel, and apparently when you send them a CPU that they know is very likely to be affected they don't actually do any validation, except that you actually sent them the CPU, and almost immediately send you a replacement.

So, while they might have been reluctant to accept RMAs in the past, it seems that right now they are actually doing the right thing regarding RMAs, and they are trying to make things as painless as possible for their customers.

While I can't totally forgive them for their past behavior, I have to give them a 10/10 for my recent experience.

The new CPU uses just 1.279 V in the BIOS, compared to the defective CPU, which used 1.447V. I left the BIOS on the Intel defaults with no XMP, and so far I don't have any stability issues and it reaches the maximum frequencies properly. But I won't do more extensive testing or stress it more until the microcode updates are released.

Dang! That's a good improvement. Do you have an ASUS ROG motherboard? Curious to see the SP prediction for your chip

 

[MikeSnow]

Yes, I have a ASUS ROG STRIX Z790-F GAMING WIFI motherboard.

 

[Dr. Dro]

Yes, I have a ASUS ROG STRIX Z790-F GAMING WIFI motherboard.

Mmm I don't know if the Strix series have SP prediction readout capability like the Maximus series, but if so you can find it on "AI tweaking" section. Make sure to update your BIOS and use the Intel "Performance" profile if you want to keep your chip safe, though. You'll lose some perf but temps will be great and volts nice and low

 

[trparky]

The new CPU uses just 1.279 V in the BIOS, compared to the defective CPU, which used 1.447V. I left the BIOS on the Intel defaults with no XMP, and so far I don't have any stability issues and it reaches the maximum frequencies properly. But I won't do more extensive testing or stress it more until the microcode updates are released.

According to a video by Gamer Meld on YouTube that went over an article by Igor's Lab and The Verge, all new replacement processors already have the new microcode loaded.

Will replacement / RMA’d chips ship with the microcode update preapplied beginning in August? Is Intel still shipping replacement chips ahead of that update?
Intel will be applying to microcode to 13th/14th Gen desktop processors that are not yet shipped once the production patch is released to OEM/ODM partners (targeting mid-August or sooner). For 13th /14th Gen desktop processors already in service, users will need to apply the patch via BIOS update once available.

 

[Sunny and 75]

I was quite negative here about Intel this week, but my recent RMA experience has been flowless. I had stability issues with my 13900KF CPU for over an year, but I refrained from asking an RMA until this week, as I needed my computer for work and didn't want to wait a few weeks for a replacement.

Anyway, as this Saturday I'm leaving on vacation for two weeks, I decided that this week would be the best time to ask an RMA for it, especially in light of the latest revelations. I was hoping to get a replacement when I got back from vacation.

So, on Tuesday I made the RMA request, telling them that my CPU is unstable even with the latest BIOSes and the Intel defaults.

- 8 hours later, they were asking for the serial number, as I have given them an incorrect serial number initially.
- 4 hours later the RMA was approved, and I received an estimation of 5-7 business days to receive the replacement, from the time I submit the faulty unit.
- 1 hour later I received the DHL return label. The collection was scheduled for Thursday, 2 days later, but I decided not to wait for that.
- 3 hours later I handed over the faulty CPU at a local DHL service point
- 15 hours later the CPU started its travel towards Intel
- 22 hours later the CPU arrived at Intel
- 7 hours later they initiated shipping for the replacement CPU
- 2 hours later the CPU was on its way to me
- 3 hours later I received an email confirming they received the CPU and they need to validate it which would take 1 to 3 business days. Not sure what that email was about, since they already sent the replacement, probably just an automated message.
- 16 hours later I had a brand new 13900KF CPU on my desk.

Overall, the process took just a bit over 3 days, which I find almost unbelievable. I didn't get any pushback from Intel, and apparently when you send them a CPU that they know is very likely to be affected they don't actually do any validation, except that you actually sent them the CPU, and almost immediately send you a replacement.

So, while they might have been reluctant to accept RMAs in the past, it seems that right now they are actually doing the right thing regarding RMAs, and they are trying to make things as painless as possible for their customers.

While I can't totally forgive them for their past behavior, I have to give them a 10/10 for my recent experience.

The new CPU uses just 1.279 V in the BIOS, compared to the defective CPU, which used 1.447V. I left the BIOS on the Intel defaults with no XMP, and so far I don't have any stability issues and it reaches the maximum frequencies properly. But I won't do more extensive testing or stress it more until the microcode updates are released.

That's nice to hear.

 

[Dredi]

According to a video by Gamer Meld on YouTube that went over an article by Igor's Lab and The Verge, all new replacement processors already have the new microcode loaded.

Yes, but only after it has been made available, in mid august. The now shipping replacements do not have it.

It even explicitly states that in the quote from verge you posted.

 

[trparky]

Now there's talk about how the problem might be with the ring bus. The reasoning is that for whatever stupid reason, the ring bus is on the same voltage domain as the CPU cores are. So as the cores request higher voltages to hit higher and higher clocks, the ring bus is being fed the same amount of voltage and getting cooked.

It even explicitly states that in the quote from verge you posted.

Yeah, I just realized that. Oops.

 

[Dredi]

Now there's talk about how the problem might be with the ring bus. The reasoning is that for whatever stupid reason, the ring bus is on the same voltage domain as the CPU cores are. So as the cores request higher voltages to hit higher and higher clocks, the ring bus is being fed the same amount of voltage and getting cooked.

That would be.. pretty bad. ”Fixing” it would mean that the top voltage of the cores must be lowered and the chips would no longer be able to hit the advertized clock speeds. Oof.

 

[Caring1]

all new replacement processors already have the new microcode loaded.

Pre gimped from the factory.

 

[Sunny and 75]

Pre gimped from the factory.

A re-review is called for, then.

 

[chrcoluk]

Now there's talk about how the problem might be with the ring bus. The reasoning is that for whatever stupid reason, the ring bus is on the same voltage domain as the CPU cores are. So as the cores request higher voltages to hit higher and higher clocks, the ring bus is being fed the same amount of voltage and getting cooked.


Yeah, I just realized that. Oops.

I linked information which says this isnt the case. I know Buildzoid has said they the same voltages but what other source has said it.

In my bios I have independent controls for core and ring voltage, which is an indication they different, but also Intel documentation says it itself.

Ring Interconnect - 002 - ID:743844 | 13th Generation Intel® Core™ Processors

edc.intel.com

I will quote it here as well

The Ring shares frequency and voltage with the Last Level Cache (LLC).

 

[trparky]

I linked information which says this isnt the case. I know Buildzoid has said they the same voltages but what other source has said it.

In my bios I have independent controls for core and ring voltage, which is an indication they different, but also Intel documentation says it itself.

Ring Interconnect - 002 - ID:743844 | 13th Generation Intel® Core™ Processors

edc.intel.com

I will quote it here as well

Then explain how in some cases people have "fixed" their issues by either running RAM at stock (no overclocking), disabling E-Cores thus reducing the possible extra voltage hit, or a combination of the two? The only thing that I can possibly think of is that that reduces the stress on an already compromised ring bus and that the ring bus has been overtaxed by higher than normal voltage.

Then let's take into account that the ring bus as we know it today is still the same ring bus that was used back when Intel was still pumping out quad-core CPUs. As Intel began to add more cores to their processors with the HEDT line, they found that adding more cores to the ring bus introduced too much stress and latency thus leading to reduced performance. Hence it leads to why Intel introduced a mesh-style bus to their X-series of chips that were part of their HEDT lineup.

Fast-forward to today and Intel is expecting the very same ring bus to handle far more cores than it ever was designed to handle.

 

[chrcoluk]

Then explain how in some cases people have "fixed" their issues by either running RAM at stock (no overclocking), disabling E-Cores thus reducing the possible extra voltage hit, or a combination of the two? The only thing that I can possibly think of is that that reduces the stress on an already compromised ring bus and that the ring bus has been overtaxed by higher than normal voltage.

Then let's take into account that the ring bus as we know it today is still the same ring bus that was used back when Intel was still pumping out quad-core CPUs. As Intel began to add more cores to their processors with the HEDT line, they found that adding more cores to the ring bus introduced too much stress and latency thus leading to reduced performance. Hence it leads to why Intel introduced a mesh-style bus to their X-series of chips that were part of their HEDT lineup.

Fast-forward to today and Intel is expecting the very same ring bus to handle far more cores than it ever was designed to handle.

I never said the ring voltage hasnt gone up, just I think its not the same as core voltage.

On RL the cache clock got a massive boost, so logically voltage will have gone up also to power that.