• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.
  • The forums have been upgraded with support for dark mode. By default it will follow the setting on your system/browser. You may override it by scrolling to the end of the page and clicking the gears icon.

Intel Will Not Recall Failing 13th and 14th Gen CPUs

Who knows, the CPU might kill you right?.
The broken English might kill you sooner there lol

FTR I mean the broken english in that UE5 engine message :)
 
Last edited:
WOW!!!!! This is way too much. I could understand a message saying "Intel CPUs are unstable with UE5 and could crush the system, so save any unfinished work before continuing", but saying "Shut down your PC immediately"? Just WOW!!!! People with no tech experience could panicking seeing this message.

Someone is very very angry at Intel and I can understand them, having to deal for months with problems and bugs in the UE5 engine that in the end weren't problems and bugs of the UE5 engine. I bet people have spend a huge number of work hour and lost many nights of good night sleep to troubleshoot what in the end is Intel's failure.

I read later in the topic it popups only on known error, shader compile...
 
If Intel were truly being honest, they would create a detection tool and push it through Windows Update, as well as pushing it through their home page, and all tech media outlets. The tool would tell the user how to RMA the chip, and even start a cross ship for them.
 
Last edited:
If Intel were truly being honest, they would create a detection tool and push it through Windows Update, as well as pushing it through their home page, and all tech media outlets. The tool would tell the user how to RMA the chip, and even start a cross ship for them.

I don't think this is feasible, you see, the processors affected aren't defective, they are damaged. Anyone who overclocked a CPU balls to the wall and degraded them knows the pattern of behavior from these crash happy chips - electromigration damage makes the CPUs extremely unstable. Once electromigration damage reaches a certain extent, the CPU will stop POSTing and on boards with a debug code reader, it'll likely read 00 as if there was no processor installed. By that point, the chip is completely wasted at a physical level.
 
Things arnt looking to good for intel lately........

 
If Intel were truly being honest, they would create a detection tool and push it through Windows Update, as well as pushing it through their home page, and all tech media outlets. The tool would tell the user how to RMA the chip, and even start a cross ship for them.
That would work only if they shipped out 15th Gen as the replacement, not 14th Gen with the gimped micro code update.
 
That would work only if they shipped out 15th Gen as the replacement, not 14th Gen with the gimped micro code update.

If they can't have enough 13-14th gen, so they may league with AMD to ake users switch to AMD :p , instead.
 
That would work only if they shipped out 15th Gen as the replacement, not 14th Gen with the gimped micro code update.
Then Intel needs a new stepping, and offer a non-gimped CPU, or offer a free upgrade to the 15th gen. It's their mess, and they need to step up, not brush it under the rug, which is what they are currently doing. This will destroy their reputation if this isn't done right.

Yes it will cost them a fortune, but it's all their fault. They insist on pretending a new CPU needs a new MB chipset, that's the same as the last.

The very interesting thing is that it's also affecting Laptops, and OEM systems. This is going to be very expensive.
 
Last edited:
Seems likely that this is due to overbinning.

They probably thought they could get thier wafer cost down by getting more out of the lower bins by pumping absurd voltage through them to hit clocks. There are informal reports of the replaced chips are coming down from 1.45-1.5 vcore to mid 1.2s idling the bios, probably lower VID tables on the proper bins all around.
 
Last edited:
Things arnt looking to good for intel lately........

Intel is definitely on my financial health watch list. It is very bad for them right now and even though I'm not in the board room, I get this gut feeling that Intel has no idea how bad off they really are.
 
Intel is definitely on my financial health watch list. It is very bad for them right now and even though I'm not in the board room, I get this gut feeling that Intel has no idea how bad off they really are.
But they're also THE LAST high-end fab that isn't in Asia. I would bet they're going to get bailed out if anything happens.
 
A re-review is called for, then.


Exactly...


You never know. Only time will tell. They can't just let go of benchmarks though, they MUST appear faster than AMD's offerings even if it's just by (less than) 1% at any cost.

They do not want another 7800X3D situation. They cannot afford their shiny new Core Ultra 9 285K (not official naming) to be beaten by the upcoming 9800X3D (with improved V Cache, mind you).

We truly need both of them strong for a thriving CPU market.


Maybe Bring Cyrix/UMC back ;)

Seems likely that this is due to overbinning.

They probably thought they could get thier wafer cost down by getting more out of the lower bins by pumping absurd voltage through them to hit clocks. There are informal reports of the replaced chips are coming down from 1.45-1.5 vcore to mid 1.2s idling the bios, probably lower VID tables on the proper bins all around.
So basically smoke and mirrors, intel being caught red handed, their arrogance is making them fall.
 
I mean i've had no stability issues with a day 1 13700KF and MSI mobo -- easily one of the most stable platforms / nicest platforms i've had so far.
Good for you but dont use your one sample to belittle the real problem. Besides unless you have a crystal ball and can see the future you dont know if your CPU is going to be fine for one, two or more years.
Really curious to see what the actual numbers of the issue are after the OMFG ZOMG 100% FAIL RATE techtuber clickbaits die down.
Who has said there's a 100% fail rate? This is not a "few youtubers" making noise. This is a real issue.
I doubt these issues are as widespread as they are made out to be.
Based on what? Your one sample or wishful thinking?
Don't be so dramatic. While the whole situation sucks, you could always buy an entry level CPU as a replacement for this time.
He could but he would be stuck with the same problem. Especially if he did not want to lose performance and goes with another 14900 as a replacement.
You can even go to 12 series or some after market CPU to just boot PC. Not sure if you need the power of 14900hf for work reasons.
Yes that's a viable quick fix and avoids the issue popping up in a replacement.
I wonder how many of those having instability issues did overclock their CPUs over the insanely high factory OC (Turbo 2.0, TVB, eTVB, Turbo 3.0), and how many of them used the third party contact frame to mount their CPU. Because you know Deubouer says it's good. :D
Probably even without ESD equipment. ;)
People running stock intel profile have issues. Server motherboards that have zero OC and even have lowered memory speeds have issues. Laptops have issues. Even low power T series chips are having issues. It's irrelevant how many users are facing this problem due to overclock because right now it seems to be a matter of luck more than anything else. Besides the number of users OC'ing these days is very small regardless.

And what does contact frame have to do with this? Contact frame improves the mounting pressure and temps but this issues does not seem temperature related based on Wendell's data as none of the server CPU's exceeded 83c. If ESD occurred it would kill the motherboard first, not the CPU.
 
Good for you but dont use your one sample to belittle the real problem. Besides unless you have a crystal ball and can see the future you dont know if your CPU is going to be fine for one, two or more years.

Who has said there's a 100% fail rate? This is not a "few youtubers" making noise. This is a real issue.

Based on what? Your one sample or wishful thinking?
1722179331956.png


1722179362189.png

1722179400550.png



We have a bunch of CPUs on 13th gen deployed in our fleet - over roughly 200 laptops on 13th gen i7s that have been deployed for a while now, plus the 2 14700k dev boxes, and my own custom builds for friends - none of them are crashing (so far). Which is why I think alot of this issue is being used to drive clicks - it's a overbinning (greed) problem at intel for sure, but it's obvious that it's being sensationalized. Same thing happened with AMD 7900XTX coolers, Nvidia power connectors, 7800x3ds exploding... it's "THE END OF THE WORLD" and then everyone forgets about it and moves on and in the end the issue ended up being a fraction of the apocalypse that it was made out to be.
 
Last edited:
Good for you but dont use your one sample to belittle the real problem. Besides unless you have a crystal ball and can see the future you dont know if your CPU is going to be fine for one, two or more years.

Who has said there's a 100% fail rate? This is not a "few youtubers" making noise. This is a real issue.
It's not really just his sample though. Most Raptor Lake owners never had any issues. This would be a lot bigger if every CPU was crashing or worse, which would be a given for a lot of 13th gen i7 and above, in particular the 13900K and KS, which based on their age would already have degraded enough to show issues. Naturally that doesn't mean the issue doesn't exist, or that it isn't big, but it isn't the apocalypse you would think reading the comments in the news about the issue.

A lot of people said/implied the fail rate is 100% or close to it, in this news there are multiple people claiming things of the kind, some youtubers and "devs" have also claimed similar things. The issue is real, but how widespread the degradation is, continues to be unknown for the moment.
 
View attachment 356654

View attachment 356655
View attachment 356656


We have a bunch of CPUs on 13th gen deployed in our fleet - over roughly 200 laptops on 13th gen i7s that have been deployed for a while now, plus the 2 14700k dev boxes, and my own custom builds for friends - none of them are crashing (so far). Which is why I think alot of this issue is being used to drive clicks - it's a overbinning (greed) problem at intel for sure, but it's obvious that it's being sensationalized. Same thing happened with AMD 7900XTX coolers, Nvidia power connectors, 7800x3ds exploding... it's "THE END OF THE WORLD" and then everyone forgets about it and moves on.
You're lucky then that not single one of yours has failed. Others who have hundreds deployed have double digit failure rate and that's not acceptable.
And not just one studio or developer is claiming this. Multiple are claiming this.

Also the return rate for 13th and 14th gen remains 3-4x higher that other models. I very much doubt all these people returning them are doing so because of the news.

The issues you brought up are all very different. One is a stock HSF defect on one model that was solved quickly. The other is a power connector issue that is going on even today but still mainly affects 4090. The third was an issue again with one CPU model that affected mostly ASUS boards and was fixed quickly by AMD with clear communication to affected users. All of these issues mainly affected a single model. Intel's problem - whatever it turns up to be affects multiple series of models. Not just some balls to the wall enthusiast models.

I agree that 100% failure rate seems very unlikely. Based on current reporting it seems to be 25-50% in some instances.
 
It's going to be really hard to figure out the actual failure rate. Intel is going to have the best idea of that, and they aren't going to divulge that willingly unless they can say with some certainty that it's a really low percentage. I don't see them using language like "a small number affected" or anything like that. Instead, the patch covers most of the 13/14 series lineup. The original speculation was that this affected the K-series only, but now they are extending the patch to anything 65W and above, which is pretty much the whole desktop lineup. That's not a good sign.

The microcode patch is going to be revealing, IMO. If we see some drastic changes in the voltage and boost behavior, then I think that's a telling sign that these chips were indeed pushed way too hard. If they really need to hobble TVB, it's literally damage control, as they needed to do something drastic to reduce the failure rates. Their first step is trying to cut down on replacements and warranty service. The BYO crowd isn't even the biggest problem--they need to worry about upset OEM partners that are dealing with a lot of warranty claims. The US isn't the issue, but in many other countries, warranties have a 2-3 year minimum. And now, Intel can lose OEM volume to not just AMD, but now Qualcomm.
 
You're lucky then that not single one of yours has failed. Others who have hundreds deployed have double digit failure rate and that's not acceptable.
And not just one studio or developer is claiming this. Multiple are claiming this.

Also the return rate for 13th and 14th gen remains 3-4x higher that other models. I very much doubt all these people returning them are doing so because of the news.

The issues you brought up are all very different. One is a stock HSF defect on one model that was solved quickly. The other is a power connector issue that is going on even today but still mainly affects 4090. The third was an issue again with one CPU model that affected mostly ASUS boards and was fixed quickly by AMD with clear communication to affected users. All of these issues mainly affected a single model. Intel's problem - whatever it turns up to be affects multiple series of models. Not just some balls to the wall enthusiast models.

I agree that 100% failure rate seems very unlikely. Based on current reporting it seems to be 25-50% in some instances.

Those issues when they arose were "ALL 7900XT and XTX have cooler flaw", and "ALL 12VPHR CONNECTORS WILL MELT" and "7800X3Ds are already starting to explode, AMD blames motheroards etc." Once everyone's panties were unwadded, the true size of the issues were much smaller, as you pointed out above.

The developer that's claiming a 50% failure rate is using a farm that put 14900ks and 13900ks in blades that were designed for Alder Lake Xeons and had to be flashed with a custom bios to support RPL-S, the desktop variant. And when those blades started chewing through 13900ks and 14900ks - they put in more 13900ks and 14900ks to replace them. If a system is consistently having problems with a certain model of processor, and you keep putting said processor in that system, then you can reach a 50% failure rate fairly easily.

For sure there is a massive problem - but I would bet based on past issues that it's unlikely to be even in the double digits. 3-4x return rate on Intel processors when return rates ar 1-2% for processors as a whole means there is a 3-8% return rate, which is MASSIVE -- but no where near the 25% - 100% being pushed. For the overbinned i9s (14900ks,13900ks) are probably the most affected.
 
Last edited:
You're lucky then that not single one of yours has failed. Others who have hundreds deployed have double digit failure rate and that's not acceptable.
And not just one studio or developer is claiming this. Multiple are claiming this.

Also the return rate for 13th and 14th gen remains 3-4x higher that other models. I very much doubt all these people returning them are doing so because of the news.

The issues you brought up are all very different. One is a stock HSF defect on one model that was solved quickly. The other is a power connector issue that is going on even today but still mainly affects 4090. The third was an issue again with one CPU model that affected mostly ASUS boards and was fixed quickly by AMD with clear communication to affected users. All of these issues mainly affected a single model. Intel's problem - whatever it turns up to be affects multiple series of models. Not just some balls to the wall enthusiast models.

I agree that 100% failure rate seems very unlikely. Based on current reporting it seems to be 25-50% in some instances.
As far as I've seen it's one studio and one developer claiming huge failure rates. I've seen others claiming to have issues, but none of the ones I've seen claimed similar failure rates.
Return rate for CPUs are probably really low normally, so even a 4x increase might not be that much in absolute numbers. Still a major issue but without knowing the number it's getting compared against, it's hard to understand how big the issue is.
 
Those issues when they arose were "ALL 7900XT and XTX have cooler flaw", and "ALL 12VPHR CONNECTORS WILL MELT" and "7800X3Ds are already starting to explode, AMD blames motheroards etc." Once everyone's panties were unwadded, the true size of the issues were much smaller, as you pointed out above.

The developer that's claiming a 50% failure rate is using a farm that put 14900ks and 13900ks in blades that were designed for Alder Lake Xeons and had to be flashed with a custom bios to support RPL-S, the desktop variant. And when those blades started chewing through 13900ks and 14900ks - they put in more 13900ks and 14900ks to replace them. If a system is consistently having problems with a certain model of processor, and you keep putting said processor in that system, then you can reach a 50% failure rate fairly easily.

For sure there is a massive problem - but I would bet based on past issues that it's unlikely to be even in the double digits. 3-4x return rate on Intel processors when return rates ar 1-2% for processors as a whole means there is a 3-8% return rate, which is MASSIVE -- but no where near the 25% - 100% being pushed. For the overbinned i9s (14900ks,13900ks) are probably the most affected.
According to some people on Twitter that keep track of return rates RMA and stuff you are indeed correct, return rates seem to be between 3 and 7%.

On 12th and amd cpus return rates are around 1.5 to 3.

As far as I've seen it's one studio and one developer claiming huge failure rates. I've seen others claiming to have issues, but none of the ones I've seen claimed similar failure rates.
Return rate for CPUs are probably really low normally, so even a 4x increase might not be that much in absolute numbers. Still a major issue but without knowing the number it's getting compared against, it's hard to understand how big the issue is.
It's one of those "the boy who cried wolf" situations. The massive intel hatred leads a lot of content creators to over dramatize the situation for the clicks. If there was less intel hatred in general I might take these reports more seriously, but as is, I just don't believe the numbers are anywhere near those 50 and 100% claims.

All the hysteria up until now about any other intel issue has been a big nothing burger (needs W11 to work, ecores cause megastutters in games, background task performance drops like a rock and what have you). It's really hard to trust the techtubers on this one because up until now they haven't earned my trust. Intel has. Maybe less clickbaiting and more facts would go a long way towards actually protecting me and the rest of the consumers. Cause if this is an actual big issue I'd really like to know and avoid intel but alas - it's about the clicks.
 
Sure they won't have to change their name, but if it was 100% fail it could become Intel Outside.

25% would be good enough, money and reputation wise, but 50% i doubt it can be called good.
 
The cope is intense, and i'm not surprised given how much mindshare Intel has. Trusting intel over reputable tech youtubers bringing factual information here? Without tech youtubers bringing forth the corrosion issue, and the rapid degradation of i9's I doubt intel would have even admitted the 13th gen cpus had a degradation problem. Sure there are a bunch of people jumping on the topic just repeating what GN and Wendell has said but that doesn't dismiss cpus failing being a problem.
IMO not offering a recall isn't enough even if the failure rate is only around 7%, who knows how damaged these cpus are and how affected the lifespan is even with the supposed microcode fix, I definitely wouldn't buy used 13th or 14th gen.
 
Could all of this be the reason why we are getting a 1700 CPU with just P cores and n HT. I think it is called Bartlett. As much as some people think that this is being inflated did not watch Wendlell's video completely.
 
it's a overbinning (greed) problem at intel for sure, but it's obvious that it's being sensationalized.
Just for clarity, no chips are "more binned" or "overbinned". Binning is simply the process of sorting a mass-produced product by qualities for product segmentation. All microchips are "binned", an i9-14900K is no more "binned" than a i5-14600.
Whether Intel makes lots of small bins that goes into a single SKU, or they make one large bin with a lot of variance ranging from lower quality samples to golden samples, I don't know.
But they do most certainly have this data for every single chip, and they would also know if they changed the binning at some time and it had unintentional effects on RMAs. So if it were the case that the lower quality xx% of i9-14900Ks were overrepresented in RMAs, but the rest were normal return rates, they would be able to identify this problem very quickly and recall those serial numbers. Similarly, if there were certain batches or production lines which had extremely high failure rates, they could have identified this a long time ago. Remember, these chips were run through qualification two years ago, but the widespread problems only showed up fairly recently (this year?).

I don't think the problem is really that simple. While silicon quality probably plays a role in when/if the symptoms arrive, their binning is likely not the main cause, but rather the aggressive application of voltage combined with all the various design considerations.

Same thing happened with AMD 7900XTX coolers, Nvidia power connectors, 7800x3ds exploding... it's "THE END OF THE WORLD" and then everyone forgets about it and moves on and in the end the issue ended up being a fraction of the apocalypse that it was made out to be.
Most have long forgotten that the great Sandy Bridge was a horrible disaster in the beginning. Bad chips, defective chipsets and even many motherboards with major issues, and it kept on going for like six months or so. So many swore not to buy Intel ever again…
When they eventually got the issues sorted out, the platform turned out pretty great.

Also the return rate for 13th and 14th gen remains 3-4x higher that other models. I very much doubt all these people returning them are doing so because of the news.
Considering how low RMA rates for CPUs normally are compared to other PC parts, it wouldn't take much to "dominate" those statistics. If the return rates were like 20% or 50% like some have suggested, the big systems integrators would have stopped selling these a long time ago, that overhead would simply just kill their margins.
 
The cope is intense, and i'm not surprised given how much mindshare Intel has. Trusting intel over reputable tech youtubers bringing factual information here? Without tech youtubers bringing forth the corrosion issue, and the rapid degradation of i9's I doubt intel would have even admitted the 13th gen cpus had a degradation problem. Sure there are a bunch of people jumping on the topic just repeating what GN and Wendell but that doesn't dismiss cpus failing being a problem.
IMO not offering a recall isn't enough even if the failure rate is only around 7%, who knows how damaged these cpus are and how affected the lifespan is even with the supposed microcode fix, I definitely wouldn't buy used 13th or 14th gen.
One thing doesnt exclude the other; drama makes people rich
 
Back
Top