Intel's CPUs Are Failing, ft. Wendell of Level1 Techs
video description
Date: 2024-07-12
Related videos
Comments and reviews: 20
stevetheripper
I was already annoyed at Intel when 12th gen came out, and I had to cool them. There wasn't much above 100c cooling on the market back then that didn't cost an arm and a leg, i.e. custom loops, and they just dropped the CPU and said there ya go, knock yourselves out without giving case and cooling partners time to adapt. Considering doped silicone can handle about 120c, and I use the super scientific rule of thumb, 1watt = 1c, I thought to myself they better not increase the wattage. Of course, that's not to mention two different instruction sets for the E and P cores (AVX-512, I'm looking at you!), plus anti-cheat systems going nuts thinking the game was running on a VM or something, this has for me been three releases in a row that have been badly engineered. Just terrible engineering. Edit: Totally forgot about 11th gen. I wasn't building PCs at the time, but I heard it was the start of the unforced errors we've been seeing from Intel. Couple that with them needing to change their node names away from nanometres, because they couldn't keep up, and losing Apple...
reply
I was already annoyed at Intel when 12th gen came out, and I had to cool them. There wasn't much above 100c cooling on the market back then that didn't cost an arm and a leg, i.e. custom loops, and they just dropped the CPU and said there ya go, knock yourselves out without giving case and cooling partners time to adapt. Considering doped silicone can handle about 120c, and I use the super scientific rule of thumb, 1watt = 1c, I thought to myself they better not increase the wattage. Of course, that's not to mention two different instruction sets for the E and P cores (AVX-512, I'm looking at you!), plus anti-cheat systems going nuts thinking the game was running on a VM or something, this has for me been three releases in a row that have been badly engineered. Just terrible engineering. Edit: Totally forgot about 11th gen. I wasn't building PCs at the time, but I heard it was the start of the unforced errors we've been seeing from Intel. Couple that with them needing to change their node names away from nanometres, because they couldn't keep up, and losing Apple...
reply
MrHendrikje
Gosh, the one time I choose to run an Intel CPU and they have a rather big issue on their hands.
I do hope that my setup is irregular enough that it doesn't affect me.
I run an i7 13700K, on an Gigabyte Z690M AORUS ELITE AX DDR4.
My Ram is a DDR4 32GB Kit that runs at 3600MHZ.
I also applied a small undervolt of -100mv to -150mv I think its small anyhow, I forgot what I put down exactly..
All of this I hope means that my CPU isn't really being punished tooo much. the memory controller on the CPU also should have an easier time with DDR4 3600 compared to DDR5 speeds.
I haven't had issues my self, yet. But I will keep this in mind and see if I find anything that isn't out of the ordinary.
It likely doesnt matter too much, but my RTX 3080 10G is also undervolted with a locked speed.
Max voltage is set to 0.850mv
Max coreclock is set to 1830 MHZ.
(Tho.. it did create 2 more steps against my wishes, because.. Nvidia I guess. so at times it can go ip to 1840 or 1860.)
reply
Gosh, the one time I choose to run an Intel CPU and they have a rather big issue on their hands.
I do hope that my setup is irregular enough that it doesn't affect me.
I run an i7 13700K, on an Gigabyte Z690M AORUS ELITE AX DDR4.
My Ram is a DDR4 32GB Kit that runs at 3600MHZ.
I also applied a small undervolt of -100mv to -150mv I think its small anyhow, I forgot what I put down exactly..
All of this I hope means that my CPU isn't really being punished tooo much. the memory controller on the CPU also should have an easier time with DDR4 3600 compared to DDR5 speeds.
I haven't had issues my self, yet. But I will keep this in mind and see if I find anything that isn't out of the ordinary.
It likely doesnt matter too much, but my RTX 3080 10G is also undervolted with a locked speed.
Max voltage is set to 0.850mv
Max coreclock is set to 1830 MHZ.
(Tho.. it did create 2 more steps against my wishes, because.. Nvidia I guess. so at times it can go ip to 1840 or 1860.)
reply
Tenggaming01
This is one of the weirdest things to ever happen to me am having a 13400f which is really unstable the cpu is picky when it comes to graphics card anything from rx6700-6900xt the cpu doesn’t seem to recognize it but the pc boots whereby my ez debug gpu light would be on indicating the graphics card can’t communicate with the processor if those AMD cards are inserted I initially thought my board was the problem..so I tossed a 12400f into the motherboard board and everything work fine with any card so I placed the 13400f on a friends MSI b760m and it’s still picky with graphics card.only nvidia card I have tried with this processor(13400f) and it worked was an rtx 3070
The processor is a tray processor luckily I have used it for like a year and my reseller gave me 3years warranty so I might RMA it
What do you think could be the issue In my wildest imagination I haven’t seen a picky processor before haha
reply
This is one of the weirdest things to ever happen to me am having a 13400f which is really unstable the cpu is picky when it comes to graphics card anything from rx6700-6900xt the cpu doesn’t seem to recognize it but the pc boots whereby my ez debug gpu light would be on indicating the graphics card can’t communicate with the processor if those AMD cards are inserted I initially thought my board was the problem..so I tossed a 12400f into the motherboard board and everything work fine with any card so I placed the 13400f on a friends MSI b760m and it’s still picky with graphics card.only nvidia card I have tried with this processor(13400f) and it worked was an rtx 3070
The processor is a tray processor luckily I have used it for like a year and my reseller gave me 3years warranty so I might RMA it
What do you think could be the issue In my wildest imagination I haven’t seen a picky processor before haha
reply
FFRyanBL
I got a 13900k really close to launch. It was about 3-4 months before I started having to turn down turbo boost speeds to keep it from closing random programs. The problem was so new that almost no one had this problem for like 7 months after mine started, there was only 1 reddit thread I found. Never had any errors codes or crash info, it would just boot up and after 20 minutes start closing programs randomly every 1 second - 5 minutes. Kept getting worse and was turning down turbo .1ghz or so each month to keep it working. Down to 4ghz before I gave up. Support was like try replacing EVERY OTHER COMPONENT before we'll consider RMA. So I pretty much bought and built an entire second computer and the problem was still there. Support was like oh not our problem there's no errors.
I gave up and built ANOTHER system and switched to AMD
reply
I got a 13900k really close to launch. It was about 3-4 months before I started having to turn down turbo boost speeds to keep it from closing random programs. The problem was so new that almost no one had this problem for like 7 months after mine started, there was only 1 reddit thread I found. Never had any errors codes or crash info, it would just boot up and after 20 minutes start closing programs randomly every 1 second - 5 minutes. Kept getting worse and was turning down turbo .1ghz or so each month to keep it working. Down to 4ghz before I gave up. Support was like try replacing EVERY OTHER COMPONENT before we'll consider RMA. So I pretty much bought and built an entire second computer and the problem was still there. Support was like oh not our problem there's no errors.
I gave up and built ANOTHER system and switched to AMD
reply
ShadySKWASHA
Great video, glad I kept my 12900k instead of getting the 13900KS I was planning to buy to keep my EVGA Z690 Classified build going longer... Haven't had any issues, undervolted mine to keep my gaming room a bit cooler cause as you see in my PFP, I got a feline CPU heater on top of my 420mm Arctic Liquid Freezer 2 AIO... Might still risk it and return the 13900KS till I find a stable one. For the record I have AMD system too, my laptop and Steam Deck is AMD, my other movie and downstairs desktop is a Ryzen 5800X... Ran a 5800X3D before and had some issues there too, it's tech, it happens, just gotta learn to enjoy the process, troubleshooting is part of being a tech enthusiast... However, I might switch to AM5 for the efficiency at some point down the line lmfao, this power bill is a bit rough!
reply
Great video, glad I kept my 12900k instead of getting the 13900KS I was planning to buy to keep my EVGA Z690 Classified build going longer... Haven't had any issues, undervolted mine to keep my gaming room a bit cooler cause as you see in my PFP, I got a feline CPU heater on top of my 420mm Arctic Liquid Freezer 2 AIO... Might still risk it and return the 13900KS till I find a stable one. For the record I have AMD system too, my laptop and Steam Deck is AMD, my other movie and downstairs desktop is a Ryzen 5800X... Ran a 5800X3D before and had some issues there too, it's tech, it happens, just gotta learn to enjoy the process, troubleshooting is part of being a tech enthusiast... However, I might switch to AM5 for the efficiency at some point down the line lmfao, this power bill is a bit rough!
reply
Fender178
Brian from TechYesCity said that there is an issue with the IO on the 13900K and 14900K respectively and he had this issue where the machine seemed sluggish or he would choke his entire system when he was doing a lot of actions per minute which is a Real Time Strategy game term during his video editing. He was told by someone at last year's Computext that they moved the IO hub on the 13 and 14th gen Intel CPUs and it was crudely done which makes me believe that leading to the problems that Wendell was seeing. And I also think that the Out of GPU memory error is tied to that IO problem because I am willing to bet the GPU is getting confused during a specific IO error when it is using GPU memory and when that specific IO error occurs the GPU thinks it ran out of memory.
reply
Brian from TechYesCity said that there is an issue with the IO on the 13900K and 14900K respectively and he had this issue where the machine seemed sluggish or he would choke his entire system when he was doing a lot of actions per minute which is a Real Time Strategy game term during his video editing. He was told by someone at last year's Computext that they moved the IO hub on the 13 and 14th gen Intel CPUs and it was crudely done which makes me believe that leading to the problems that Wendell was seeing. And I also think that the Out of GPU memory error is tied to that IO problem because I am willing to bet the GPU is getting confused during a specific IO error when it is using GPU memory and when that specific IO error occurs the GPU thinks it ran out of memory.
reply
Turzel
I had this start happening like a month or two after I got my 13th gen processor (13900k) in like 2022 or something. I thought it was something with my GPU or ram and then once I narrowed it down to the processor I thought it was just a faulty product I got. After intel replaced it via warranty and the same thing started happening I thought I was crazy because I couldnt find anything online. I was only tipped off it was a widespread voltage issue after some random Techlinked video mentioning it last year. Luckily I've undervolted it and turned the turbo-boost modes on my MB off and it's been working fine. Still, it's very disappointing that this is happening and if there isnt a quick fix Intel puts out for it I hope they are able to reimburse/replace my current one.
reply
I had this start happening like a month or two after I got my 13th gen processor (13900k) in like 2022 or something. I thought it was something with my GPU or ram and then once I narrowed it down to the processor I thought it was just a faulty product I got. After intel replaced it via warranty and the same thing started happening I thought I was crazy because I couldnt find anything online. I was only tipped off it was a widespread voltage issue after some random Techlinked video mentioning it last year. Luckily I've undervolted it and turned the turbo-boost modes on my MB off and it's been working fine. Still, it's very disappointing that this is happening and if there isnt a quick fix Intel puts out for it I hope they are able to reimburse/replace my current one.
reply
DangerGnom
My guess: it is a problem from the ILM. The bending of the CPU, combined with the 24/7 load, deforms the cpu to a point, that along the bend all your dedicated functions of the cpu are affected. If you have a scematic/picture of the cpu, and you know which part of the cpu does what (i.e. memory controller - your DDR54200 issue), and you draw a line along the bend. you can see which other functional parts could be affected and corellate that to the different issues you discovered. Maybe that explains some of your discoveries. My reasoning is: until 12th gen, they used a different loading mechanism on Intel. And maybe the Z690 crowd out here that uses an 13th/14th gen cpu, has most likely replaced the default ILM with a frame. so they don't have this issue.
reply
My guess: it is a problem from the ILM. The bending of the CPU, combined with the 24/7 load, deforms the cpu to a point, that along the bend all your dedicated functions of the cpu are affected. If you have a scematic/picture of the cpu, and you know which part of the cpu does what (i.e. memory controller - your DDR54200 issue), and you draw a line along the bend. you can see which other functional parts could be affected and corellate that to the different issues you discovered. Maybe that explains some of your discoveries. My reasoning is: until 12th gen, they used a different loading mechanism on Intel. And maybe the Z690 crowd out here that uses an 13th/14th gen cpu, has most likely replaced the default ILM with a frame. so they don't have this issue.
reply
Mordrede100
I am absolutely not sure, but maybe someone here can help me. I have a friend who just got a z mobo, 14700k, ddr 5 6000mhz ram. Until I told him to turn off the mobo xmp profile to use the ram on default 4200mhz, experienced constant crashes in unreal engine games. Does this issue have any connection to this 0 crashes since, so it was 100% the xmp and whatever it changes in the whole system.
(everything was updated software-wise, except for the BIOS, I have read in that cpu gen and previous ones that bios updates usually fix it or if you turn xmp off, BUT in the video case no xmp was used, and not z chipset in the server side. However, if it is a power delivery/control issue fixed with bios updates previously, could it be done now )
reply
I am absolutely not sure, but maybe someone here can help me. I have a friend who just got a z mobo, 14700k, ddr 5 6000mhz ram. Until I told him to turn off the mobo xmp profile to use the ram on default 4200mhz, experienced constant crashes in unreal engine games. Does this issue have any connection to this 0 crashes since, so it was 100% the xmp and whatever it changes in the whole system.
(everything was updated software-wise, except for the BIOS, I have read in that cpu gen and previous ones that bios updates usually fix it or if you turn xmp off, BUT in the video case no xmp was used, and not z chipset in the server side. However, if it is a power delivery/control issue fixed with bios updates previously, could it be done now )
reply
celeriumlerium8266
I've had my 13900k since a couple weeks after release (Australia) and have yet to experience any crashing issues. Until this story started I was running it on auto mode on my Z790 Aorus Master, after I had manually enabled C-states and power limits, but my power limits have a max amperage of 350a instead of intels recommended 307a, with the 253/253 pl1 and pl2. DDR5 5600 CL34 and 360mm AIO.
Couple of my friends shared their steam library with me to play games they can't play over discord with screen share, everything I played worked fine of my system with no crashing, even with the board set back to defaults xmp/rebar.
reply
I've had my 13900k since a couple weeks after release (Australia) and have yet to experience any crashing issues. Until this story started I was running it on auto mode on my Z790 Aorus Master, after I had manually enabled C-states and power limits, but my power limits have a max amperage of 350a instead of intels recommended 307a, with the 253/253 pl1 and pl2. DDR5 5600 CL34 and 360mm AIO.
Couple of my friends shared their steam library with me to play games they can't play over discord with screen share, everything I played worked fine of my system with no crashing, even with the board set back to defaults xmp/rebar.
reply
kemaldemirel1714
My 7950x has been really unstable, I had to not only remove expo from my ram but reduce it to 4800, Just to get crashes to be like once week and 4200 to get it to be once a month.
It's my first Ryzen cpu, so I was really unhappy. But looking at this video it seems I really dodged a bullet.
(memory was originally specced only for 5600 so it's not that big of decrease in performance. Also before anyone says anectodal, ryzen is known to have issues with 2x32gb ram sticks it's not just me. They weren't known when I built my pc though. But yeah it's an uncommon and not recommended setup now)
reply
My 7950x has been really unstable, I had to not only remove expo from my ram but reduce it to 4800, Just to get crashes to be like once week and 4200 to get it to be once a month.
It's my first Ryzen cpu, so I was really unhappy. But looking at this video it seems I really dodged a bullet.
(memory was originally specced only for 5600 so it's not that big of decrease in performance. Also before anyone says anectodal, ryzen is known to have issues with 2x32gb ram sticks it's not just me. They weren't known when I built my pc though. But yeah it's an uncommon and not recommended setup now)
reply
barnysgaming
I have had interesting customer PC in a few weeks ago. Customer with a 13900k having stability issues in game all of a sudden couldn't boot to windows. I tested RAM, storage, GPU and mobo all fine. The system with her 13900k wouldn't even boot into my windows installer USB with or without XMP and other bios settings but it all ran fine with a 14100f cpu.
Customer ended up having a 7800x3d and new motherboard and cooler as she didn't want to deal with this issue again by getting a 14900k. I used her RAM, Storage, GPU and PSU with the new mobo and cpu with no issues.
reply
I have had interesting customer PC in a few weeks ago. Customer with a 13900k having stability issues in game all of a sudden couldn't boot to windows. I tested RAM, storage, GPU and mobo all fine. The system with her 13900k wouldn't even boot into my windows installer USB with or without XMP and other bios settings but it all ran fine with a 14100f cpu.
Customer ended up having a 7800x3d and new motherboard and cooler as she didn't want to deal with this issue again by getting a 14900k. I used her RAM, Storage, GPU and PSU with the new mobo and cpu with no issues.
reply
msheard5905
I had the same issue with my 14900K system where it you start crashing after a week of normal gaming use without any overclocks besides the default ones provided by my motherboard. Had to identify if through troubleshooting to confirm that it was CPU issue. Had the CPU replaced, and same issue started again after a week. My solution, swapped out the platform to a 7800X3D system. That was done in February this year and it has been running perfect/stable since. How the tables have turned...Intel used to be my go to for stability, not any more.
reply
I had the same issue with my 14900K system where it you start crashing after a week of normal gaming use without any overclocks besides the default ones provided by my motherboard. Had to identify if through troubleshooting to confirm that it was CPU issue. Had the CPU replaced, and same issue started again after a week. My solution, swapped out the platform to a 7800X3D system. That was done in February this year and it has been running perfect/stable since. How the tables have turned...Intel used to be my go to for stability, not any more.
reply
Tallinu
I had this very issue with my 13900k. I struggled for 3 months to play any Unreal Engine games. New system with video errors related to shader cache etc. I acquired two new video cards, different memory and replaced the SSD. I can't tell you how many times I wiped windows, steam, drivers the games themselves... I finally found out if I turned off anything in the bios that boosted memory and processor speeds. My system would be stable. So far it's holding, not degrading and getting worse. Certainly not running at the specs I paid for.
reply
I had this very issue with my 13900k. I struggled for 3 months to play any Unreal Engine games. New system with video errors related to shader cache etc. I acquired two new video cards, different memory and replaced the SSD. I can't tell you how many times I wiped windows, steam, drivers the games themselves... I finally found out if I turned off anything in the bios that boosted memory and processor speeds. My system would be stable. So far it's holding, not degrading and getting worse. Certainly not running at the specs I paid for.
reply
josemachado7830
It's freaking cruel for people who buys top of the line CPUs, spend some money for extra performance and get screwed by the companies for being the most excited customers. I remember going top-dollar (for the time) to buy a 3930k on a X79 platform at the time (one of the first hexacores CPUs at the time), and the platform being extremely unstable for a long time until I had 2 different BIOS updates. Intel CPU on Intel motherboard. Never again, now I stay at least 1 step below the top-notch, everything just works out of the box better.
reply
It's freaking cruel for people who buys top of the line CPUs, spend some money for extra performance and get screwed by the companies for being the most excited customers. I remember going top-dollar (for the time) to buy a 3930k on a X79 platform at the time (one of the first hexacores CPUs at the time), and the platform being extremely unstable for a long time until I had 2 different BIOS updates. Intel CPU on Intel motherboard. Never again, now I stay at least 1 step below the top-notch, everything just works out of the box better.
reply
waifuhunter9709
My 12600k is working fine. 4.1hrtz speed.
But i cant keep it on for more than 18 hours.
After 12 hours , it start to hey slower.
After 18 hours, it impossible to work with it.
It will become soooo slow.
I checked aio, no problem. Checked temp, volt , everything fine.
But after 18 hours, it will be slower than my 20yo laptop.
I have to turn it off.
If its turn off for 6 hours its not enough.
It needs to at least turned off for 9 hours to recover but not fully.
I using it 12 hours on 12 hours off
reply
My 12600k is working fine. 4.1hrtz speed.
But i cant keep it on for more than 18 hours.
After 12 hours , it start to hey slower.
After 18 hours, it impossible to work with it.
It will become soooo slow.
I checked aio, no problem. Checked temp, volt , everything fine.
But after 18 hours, it will be slower than my 20yo laptop.
I have to turn it off.
If its turn off for 6 hours its not enough.
It needs to at least turned off for 9 hours to recover but not fully.
I using it 12 hours on 12 hours off
reply
H8ts
210 systems Seriously There are hundreds of thousands 13900k/14900k systems. If there was a 50% failure rate you would see this as return rate in every shop, which you don't. Anyway, we use 13900KF systems with 6800DDR5. These systems run 24/7 as build machines for more than a year and not had a single replacement. If there is anything like 50% failure rate, it has to be a bad batch coming from Intel. So instead of speculating, focus on serial numbers of failed CPUs and track it down to a specific factory and time made.
reply
210 systems Seriously There are hundreds of thousands 13900k/14900k systems. If there was a 50% failure rate you would see this as return rate in every shop, which you don't. Anyway, we use 13900KF systems with 6800DDR5. These systems run 24/7 as build machines for more than a year and not had a single replacement. If there is anything like 50% failure rate, it has to be a bad batch coming from Intel. So instead of speculating, focus on serial numbers of failed CPUs and track it down to a specific factory and time made.
reply
dat_21
I don't like the wording make us whole. It should be in the ballpark of fix your ASAP. These products are (potentially) defective and a borderline fraud if their fixes will lower the clock speeds and performance by a significant amount.
In fact their whole single core improvement in 13-14 gen may be a dud if it comes with a cost of becoming unstable in a few years.
This is really bad, I really though we can trust intel with stability and their validation process. I guess it's not the case anymore.
reply
I don't like the wording make us whole. It should be in the ballpark of fix your ASAP. These products are (potentially) defective and a borderline fraud if their fixes will lower the clock speeds and performance by a significant amount.
In fact their whole single core improvement in 13-14 gen may be a dud if it comes with a cost of becoming unstable in a few years.
This is really bad, I really though we can trust intel with stability and their validation process. I guess it's not the case anymore.
reply
justacomment1657
My guess is that the silicone degrades over time. It could be a process problem that just happen to occure on denser systems like the fully packed 13/14th gen CPUs.
Its quite possible that high voltage does accelerate the problem.
Datacenter systems do run on rather conservative settings but they do run without pause...so they will degrade as well maybe even faster compared to the consumer space.
Whatever this is, intel better their act together or another market will run away.
reply
My guess is that the silicone degrades over time. It could be a process problem that just happen to occure on denser systems like the fully packed 13/14th gen CPUs.
Its quite possible that high voltage does accelerate the problem.
Datacenter systems do run on rather conservative settings but they do run without pause...so they will degrade as well maybe even faster compared to the consumer space.
Whatever this is, intel better their act together or another market will run away.
reply
Sadew42
This has been a huge issue for my personal computer with a 13700K. I have to tune the multiplier down to 51x or else almost any significant CPU load will cause a crash. I also had to disable all e-cores (to be fair, I'm on Win10 and not Win11). I do air cool with a Noctua NH-D15 in a well ventilated case (3x140 and 3x120 intake, with 3x140 and 1x120 outtake), but still, I'm usually just gaming not doing full-usage 100% number crunching or anything. These CPUs should not be keyed to run at 100C
reply
This has been a huge issue for my personal computer with a 13700K. I have to tune the multiplier down to 51x or else almost any significant CPU load will cause a crash. I also had to disable all e-cores (to be fair, I'm on Win10 and not Win11). I do air cool with a Noctua NH-D15 in a well ventilated case (3x140 and 3x120 intake, with 3x140 and 1x120 outtake), but still, I'm usually just gaming not doing full-usage 100% number crunching or anything. These CPUs should not be keyed to run at 100C
reply
Add a review, comment
Other channel videos
