Imagine 2nd hand market flood with faulty CPUs. Like socket 1700 will be a least popular platform for aware customers and at the same time a doom trap for unaware ones...
Maybe this is the moment AMD can finally establish a much larger market share, not by selling more new CPUs, but by Intels second hand market collapsing due to these faulty CPUs.
there goes my resale value lol. I have kept my 13900k at low TDP I think 95watts? Good temps no crashes. I never tried to OC and dump a lot of voltage on the CPU anyway...on my first few runs I noticed how hot the CPU gets and I began turning off the high bios voltage
@@scanlly Nobody knew about this issue in 2022. Intel found out in 2023 and didn't say anything, and just this year the gaming community and a few server companies spoke up about instability.
Something tells me " could be" is an understatement... When I worked for hp at one of they're dev' fabs we had contamination problems and we developed a plasma cleaning process to remove the contamination but had to scrap a lot of wafers ... Once contaminated the contaminants can diffuse into many insulating or conducting structures... there's no real going back ... Any contaminated CPUs will only get worse no matter how much intel adjusts the voltage/ power specs.
@@Baldmaxx For the time being, you can use a program called Process Lasso, to find the core that is bad by using Nvidia's (chinese site)driver selecting webpage. Use the drop-down menus(click around, select random cards etc.), it will cause the web browser to crash. Isolate the power core(select the web browser process in process lasso, set core sets through affinity) that is causing the web browser to crash. Stop your games from using the bad core using process lasso. This is only a band-aide of course, but I've had success with this method and making Star Citizen crash free. Star Citizen is one of the most demanding games CPU wise. Another way --- In Bios- Load optimized defaults, lock your all P core multiplier to 55 or 56, increase load line calibration to level 7 (Asus bios), set a core voltage + offset to target 1.4 volts (be careful, use increments of .05 -save/reboot-check voltage with CPUZ) or lower. You may be able to get away with less, depending on the degradation of the cores. Use Intel XTU AVX2 stress test (5 minutes) and test for failure. Adjust your voltage only lower than 1.4 or lower your core multiplier to achieve stability.
@@grtitann7425 I knew AMD was better, but at the time AM5 and DDR5 pricing was through the roof in Canada. Saved over 1k staying with Intel and the DDR4 I had. After RMA - I'm selling it all and picking up a 7950x3d. Screw Intel.
@@aliancemdthere are multiple things that are implied and one just needs to know (never admit fault so it is much harder to sue and no recalls are required, always try to make it look like user error for maximum profits. To mention a few). But I'd say it is pretty solid theory. Just keep the mouth shut and react minimally when public pressure is too high.
Apple did replace all of the Nvidia defectives, that's why they're not on speaking terms. They also did respond to antennagate, and the i6 being a bit _too_ futuristic (is that PTSD why they don't do foldables? lol) The only one I'm not too happy off is the superthin keyboards which make the Core m Macbook and the 16-19 MBP no-go devices, but I guess they just _really_ wanted to stick to that design because slimness tunnel-vision, and I guess it must've been mitigated enough because it seems to be a forgotten issue.
@@scanlly the scale of this issue was not know back in then, they published the information as they learn about it. Tech UA-camr don't just talk about it now to make a buck, they talk about a known issue so who plan to buy a new PC or a new CPU know about the existence of this issue. Also, would you really be ready to say that Tech youtuber such as Gamer Nexus that literally offer to buy defective CPU from viewers to send to an failure analyst lab at their own cost to investigate on this issue and see if really it was a via condensation issue, does this to get rich? Those kind of test cost lots of money and the lab literally told him that with such a complex CPU, there was like 50% chance of not being able to see the issue. No Tech Channel ever claimed to know everything, they share information as they learn about it. Like Der8auer said, this issue is an issue that is developing and get worst with the time, they could not know that there was an issue 2 or 3 years ago like you said, especially not at this scale. We just learn from Level1Tech about 2 week ago how wide spread this issue was, based on Video Games Telemetry and reports from server company and PC company, before that, all they knew was that some people was experiencing some issues and as always Intel play the card of "Small amount of people" in their public statement.
@@scanlly the rage bait merchants like gamers nexus & hardware unboxed only knew about the issue because some server owners leaked information to a completely different youtuber first. they had no idea, but now they're making thousands off clickbait nonsense and attacking intel for not coming out without knowing what had gone wrong and how to fix it. the dumb thing is the voltage issue only impacts people too lazy to know what their technology is doing, and didn't set manual voltages & clocks.
@@bigpoppa1234 Too lazy to know, you mean like people that are just using a computer with that processor that are not into bios-es and hardware tech... 99% of people basically. Ofcourse they don't know and that shouldn't be the reason not making any noise about it, it's an issue that needs to be addressed, Intel is just a company.
Disagree on one opinion: No matter what caused the permanent damage to CPUs (via oxidation alone or “faulty” microcode resulting in too high voltages or a combination) this should absolutely warrant a recall of every unit already in circulation.
It's interesting to contrast this with AMD's actions just today. They have apparently found a QC issue of some kind with Zen 5 and are literally recalling every single unit shipped for testing and possible replacement. Yes, it's much easier to do in the run-up to launch but that doesn't eliminate the contrast.
@CoffehKittehGrey I think board readiness is a red herring. The whole point of AMD's socket longevity approach is that the new processors need to work on existing boards.
intel: what?! you want us to go out of our way to help customers make their cpus last longer instead of buying a new one in a few years? having it die just out of warranty is what we call "the sweet spot" 😀
wow bro, you just made a major discovery…what’s next on your list? the earth being round? or that it revolves around the sun? we got a genius here people😂
Huh? Intel chips last for decades. They've never been concerned about you buying another chip in a few years. The only people I know that buy a new chip every few years is AMD users.
@@scanllybruh all the big tech YTers have been constantly recommending gamers & builders to use AMD. Roman is an electrical engineer, he’s reporting on the news like the rest of them calm down , AMD has enough marketing & aren’t going bankrupt. Your issue should be with the companies like HP, Dell, Acer, Asus, IBM, Lenovo etc. who will continue to choose Intel over AMD for their laptops & PCs. The “PC Gamer builders” make up less than 10% of the market , you’re quite literally preaching to proxy AMD marketers & their followers.
Intel wants people to forgot so they can deny future RMAs or be like, of you updated so must be something else. This is why recalls exist. There's no trust left.
@@DingleBerryschnapps It took them over a year to acknowledge that there was a problem, then they lied about it, backtracked, and have been silent ever since. They knowingly sold defective CPUs and hoped that people wouldn't RMA inside the warranty window. It blew up in their face and now nobody trusts Intel
You can't trust their next CPU Gen either. If the problems surface over a year later. No one can test the CPUs reliably. And you can't even be sure it won't affect their GPU branch either. Because no one knows the innards of their fabs and their production line in general. Thats the general problem of cross contamination.
@@DingleBerryschnapps No, they're not acknowledging everything, it's been over a month since you left that message and WE STILL DO NOT KNOW THE TIME FRAME FROM THE EARLIEST KNOWN DEFECTIVE UNIT TO THE NON DEFECTIVE. Which means we're still in the dark, people that bought 13th gen and 14th gen still do not have a clue if their CPU is affected by the oxidisation problems.
Easiest way to reach the majority of users with Intel affected CPUs is to contact Microsoft and display a message in widows if the system has a CPU with possible issues.
Absolutely. But I am sure it ain't gonna happen. Think about the millions of schools or work computers of people not aware of nothing of the tech word looking at a message that you may or may not need a CPU RMA..... Then think how it would hurt Intel for the next time these government companies need to buy something. They would make sure it is not Intel
Excellent piece, really appreciated the technical explanation. Perhaps an interesting observation from my end: I’ve built quite a few passively cooled configurations, and one thing that stood out was that 12th/13th gen have much elevated power consumption past 85C. So with the workload remaining equal, something like a 8700 or 10900 would show for instance 50W power draw and that would increase to 55W when closing in on 100C. In other words: the CPU would become less efficient at higher temperatures, which is logical, because it pushes higher voltage to remain stable. Now for the 12th/13th I observed a marked difference, where past 85C the power draw increased much more dramatic, getting towards 70W at the 100C point. This apparently aggressive ramping of voltage is something ‘new’ for 12th gen and it looks to me Intel took a gamble here, where things quickly go south if there are even minor inconsistencies in the production process. The microcode update is therefore more of a mitigation, but like you said: the average user will take the blow for this, because it’s likely they will be stuck with an effectively broken part (and PC, ultimately).
Warranty in France is 20 years for hidden defects before sale, and 2 years from the discovery of the said defect. The Europe Warranty rules allow each members to implement hidden defects warranties. I don't know how widespread they are in other countries.
This is complicated. What ultimately provides the cpu with voltage is motherboard, and most high end boards on the market tend to add more to the vid given by MC for higher boosts. This strategy is especially obvious when XMP is active.
Oxidation would almost certainly have been caught in accelerated wear testing. The problem with the excess voltage is that it causes wear of the transistors themselves through a process called time dependent dielectric breakdown, and this is independent of current and less dependent on temperature than electromigration, it is therefore more likely to pass in the conditions of the accelerated wear tests. Why exactly they didn't properly monitor the voltages during operation to be within the intended range on the other hand is quite perplexing.
@@GreenJalapenjo The way Intel is describing it makes it seem like a 'limited hangout' where they disclose a real problem that they had and solved while trying to distract from the much larger problem (that their cpus are running unsafe voltages that degrade them rapidly).
I don't think it's plausible that they didn't know the peak voltage reached! More likely they retroactively had to conclude that the semiconductors can't sustain as much voltage as they thought they would.
913:05 it's easy: they will not. The average customer will be clueless unless the store they bought it from clues them in. And especially beyond the (in many places quite short) normal warranty period, the stores have a stake in not telling them anything. They would have to deal with returning, arranging exchange, etc. Intel is not going to foot the whole bill for that, especially not if it's 2-3 years down the line. Oh, and not 5-6 years. I've just moved off a Haswell 4790K that was fine, but received zero OS support sadly. It even had a modest OC all these years. Rock stable.
My i7 4770k was liquid metal in between the IHS and the die with a 4.4 GHz overclock for several years with absolutely zero issues. Those days have long passed
@@scanlly many of us smaller companies have been shouting off the rooftops for years about the instability related to these CPUs. We were told that we were just hating on Intel and AMD Fanboys when we sell both products alongside each other and actually see the failure rate and abysmal RMA process
I had a 4690k for years at 4ghz. I have gone to a 14600k. I have only had crashes from over clocking. Few bios up dates have helped. I have it stable at 5.5ghz + .030 volts on ETU.
if it were just a slow decay from corrosion, they could "just" replace customers CPU with new ones. But if it's a fundamental flaw, you can't. You could give out 12th gen as a temp "solution" but if you, for example, change motherboard and chipsets every generation....
Didn't they say that the oxidation were only on 13th gen and has been fixed? And since they have moved unto 14th gen, I would say that if you are affected, then you should be able to go to their site, as Steve mentioned, to look up if you have an affected CPU, and then you should be able to RMA it directly to Intel, and they should be providing you with an equal or better version of the product, meaning if they can't give you a 13th gen, then you should receive a 14th gen version of your CPU, which I believe could be "cheaper" for them to just say that 13th gens have a "no questions return policy" and to replace with 13th or 14th gen versions, if someone RMAs a CPU to a retail store or so. But they should never be able to give you a 12th gen CPU as a replacement, as that would be a downgrade, and not be in compliance with a product of equal value.
@@TigonIIISupposedly, yes that's what they're claiming but 14th chips are rapidly having issues too. For the rest they're saying it's just voltage issues. Somehow there's a VID algorithm issue or something that took them this long to address. This would also cause accelerated degradation. Either way it doesn't really matter except that in theory these chips that aren't already damaged enough to fall out of spec could be fine after a microcode update....
I use to work at Intel, and most of the people I knew who chose to leave did so cause we could tell there was no real plan. For example in Litho we had to explain to our manager what our job was, what we could and couldn't do, and even what the alarms meant on a daily basis for months. It would get so bad that people would yell at him "I CANNOT DO THAT. IT WILL BREAK THE MACHINE." Beyond that they constantly lay off experienced engineers to hire new ones fresh out of college so they could get tax breaks from the state government. When you get rid of all the experienced people this is the kinda thing that happens. We all knew it was coming.
It sounds as if they have followed Boeing's example and replaced engineers with business economists (fresh from university, of course, because those are cheaper).
This would be disaster for stock owners so sorry but nope. It is always better to piss off customers than reduce next quarter profits. That is the logic why customers are always screwed no matter what they do...
@@timop6340 well they are already loosing market share to amd, if the problem becomes worse and more apparent turning their back on customers would be a death sentence to their stock price
The ending bit of what you said is exactly what I was thinking. My cpu is already damaged from the microcode issue regardless of if it is stable or unstable currently right at this moment in time. I have had this 13900KS since it was released and it runs on average 19-20 hours every day since. I don't see how just telling ppl to update their microcode would make this okay at all. They sold me a product and then caused damage to it, that alone should result in an RMA. When I paid over $700 at the time for this CPU, I was not aware intel was going to cause permanent damage to it for this entire 1.6 years that I have had it. Warranty only lasts 3 years, if mine isn't unstable now, it's probably not gonna start showing any signs of being unstable within the remaining 1.4 years of my warranty period. Depending on how they handle this, is gonna be the decision maker on whether I trust Intel in the future or not. The other commenter that wrote extending the warranty period to 7 years for affected CPU's, is a good happy medium imo.
But it’s not just about approaching customers about a microcode update. I assume all >65w 13th and 14th gen CPUs are damaged to some extent, if they were being used at all.
ive read some i5-12450HX LOQs died though i am sure it is a low power consumption CPU with 8(4+4) cores and max 4.4GHz for the P cores, is it possible to be affected with the oxidization issue ? i mean the 12th gen is not new but the LOQ varient is 2024 model.
Usually when a company refreshes a product, that product becomes more refined, and a an all around better Intel 13th & 14th refreshes are doing an irreparable damage to their brand
Regarding the oxidization issue. Intel said in a subreddit post "addressed with manufacturing improvements and screens". Does that mean none of the affected CPUs were released to the channel? If not why not provide a date code etc so that users can verify if they are using any of these affected CPUs? The lack of detailed information regarding this issue is frustrating as hell. I assume that the affected CPUs were all of the 13th/14th gen CPUs produced during some time period of 2023. The microcode fix is to be released to motherboard makers in mid-August. It is then up to the motherboard makers to release BIOS updates for their affected boards. Will all manufacturers release BIOS updates for all of their affected boards? I am not holding my breath on that. There is also no mention of whether this issue affects ALL 13th and 14th gen CPUs or just the higher-end ones.
@@jeffscomp ive seen i5-12th gen HX in LOQs 2024 model dies so i am a bit confused right now cuz it is not a high end and it is considered low-powered CPU. what is ur opinion about it ?
Same. Wanted a good gaming CPU for the 1st time in my 39 year life. Went with the 7800x3d and was considering 13700k or 13900k but tbh I just didn't particularly like the sound of E cores for my gaming system.
@GregoryShtevensh yeah 78x3d for gaming is 100% a no brainer. I'll be picking up a 9950x3d when it releases so I can get the best of both worlds. I really hope AM5 is a banger of platform just like AM4
@@InfernoTrees yeah I mean even for basic work stuff. It's been more than enough. I've never felt like it was poor in any particular area. 16 threads has always been enough for my needs that way. It has gotten pretty hot a few times though. Nothing to really worry about as I have it set to throttle at 81c. It's hit 80c and 79c a few times. Never throttled yet though. I'm in Australia and running a 240mm rad. I think a 320 is the way to go with these chips. But my case couldn't fit one
I don’t understand why people go Intel, Ryzen 7000 is more efficient, more performs better in games (X3D) and has comparable productivity performance. While also having a future proof platform supported for many more years.
I honestly think that it does have to be a recall on the CPUs. Without making extraordinary efforts to replace every CPU that has had a significant decrease in overall life span from this, you basically have a large number of people who will get a lot less time with the computer than they normally would, and the resale value of the systems will be disproportionately lower than previous and subsequent generations. Unfortunately I am not sure if there is a way to detect how much degradation any CPU has until it gets to the point where it affects stability, and if there isn't, then Intel would have to refund or replace every single affected CPU to fix these problems and I am not sure that they can replace enough of them to satisfy a recall. In any case they absolutely will be trying to avoid a recall as that would be extremely bad for Intel if they had to do that, so the only way that will likely happen is if a government forces them into doing a recall, which at least in the US is very rare unless it is an actual safety issue.
I agree on the necessity for a recall on the CPUs. But you are aware that this problem is not limited to the CPUs on PCs with motherboards? They're now finding the same issues on laptop CPUs. Swapping a CPU on a PC motherboard I can do myself. There's no way I can see people swapping CPUs on laptops.
@@andrewbauer1322 first off, Intel is claiming that is a different issue, I am not sure if I believe it and not sure which would be better. Second, that type of thing would go through the laptop manufacturers and they would probably do the CPU replacement.
@@jeffreyparker9396 Agreed. Also, I missed another difficulty regarding CPU replacement in a laptop..."CPU replacement in a BGA socket requires proper equipment and experience. Then there is the issue of BIOS support. Most OEMs BIOSes don't support many CPUs ... usually only the models they shipped with that series. So even if the new CPU is socket and chipset compatible the BIOS may not have the microcode to support it. So while an upgrade maybe possible its usually not feasible." (from a Moderator on Tom's Hardware)
@@andrewbauer1322 while I would agree that OEM BIOSes might not get additional CPU support added in later updates, I haven't really heard of many people ever having issues upgrading CPUs in those systems other than cooling or VRM limits.
My first 14900K with my asus strix z790 itx was definitely affected as the temps would regularly spike to 110c on 2-3 cores and voltage spiked to 1.65v+ this was using Asus AinOC feature though so im pretty sure that was a definite contributing Factor. 2nd 14900K + Dark Hero z790 zero issues gaming or encoding it has the Intel V1 Direct Die Contact Frame from TG and its working Beautifully temps only hit 70c recently because ambient temp went up since its 35c outside. This cpu+Contact frame from TG have keep this cpu at 45c while gaming voltage has stayed at a steady 1.34v -1.35v from what i can see no spikes near the 1.71v of old setup. 3rd 14900KF no issues other then stock ilm wich will need to be changed out considering another direct die since it gives the die 200% more surface area to cool from. Awaiting word if the Rumored 10-12P core CPUs will be affected or if this generation is another 14nm++++++++ like ole days.
Hey Roman, My 13900k only crashes when idle... Thanks for this video, yours were the first one to point out issues not only in high usage, but idle voltage scenarios. I have that exactly. I was debugging for long, but its hard to, because I had crashes only when my pc was idle. Random software, usually Chrome and Discord would just crash. I mitigated this with running youtube videos in a tab, and the issue was gone, always. Never had any problems playing games. Since then to avoid complete instability, on newer and now newest bios by Asus (z790e gaming wifi), and yea that solved the complete instability but sometimes still happening.
As you experienced, 12th gen is affected, just not as badly as 13th and 14th gens. Seems like the entire P and E core architecture choice was a mistake.
@@kaspervendler1726 I've started seeing comments under other videos over the past few days where even i7-12700K owners are rallying around disabling E cores. They bought into the P + E core architecture, and now throw away performance instead of managing it with undervolting, no more than 1GHz delta between all cores, and Process Lasso or CoreCtrl or the like. These CPU's require careful configuration but they were marketed toward the lowest common denominator.
I agree they need to be proactive in reaching out to customers, but how do you reach the unengaged group you mention? I don’t see an easy way for Intel to communicate with customers who just set and forget their computer.
@@der8auer-en Why are you TECH people not understanding that a lot people buy tech online. To buy tech online you need to make an account on said website for about 95% of the websites. If you still bought the said hardware without an account, 100% of the websites will ask you for your e-mail address. Do you see where I am going? 99.999% of hardware bought online can be traced back to an e-mail of the customer. Intel should contact all the stores that sell their CPUs and ask the stores to send e-mail to customers that bought said CPU. They sold millions of these faulty CPUs, it's time they invest time and money fixing their dumb, lazy, major accident!
Any time i buy a higher-price item above 100 currency-units at a computer store, they note down my address, they put it on the invoice and they store a copy of the invoice in their system. So by all reason the vast majority of people can be reached by snail mail, they will need to allocate like 3 dollars per customer to pay the stores for the outreach work. The store isn't allowed to share my address to Intel or anybody else, and it's not allowed to mail me advertisement, but it is allowed to contact me with important matters of this kind, especially if you officially designate it a recall. Recall doesn't mean every single item needs to be sent in, they can absolutely issue a recall with a DIY service option. Most CPUs are in prebuilts. It will be up to the builders whether those want to contact customers for proactive service or just handle the returns as they happen. Second hand market gets toxic, but that's to the benefit of Intel. They could alleviate it with retroactively extended warranty, without proof of purchase, say 7 years from manufacturing date, but they aren't going to do that. I think in a few months, everyone will receive microcode update anyway. BIOS isn't the only option, Windows also receives microcode patches via Windows Update, the NT bootloader is capable of injecting them.
@@der8auer-en Like someone else mentioned above, Intel should reach out to Win and Unix system dev-teams, asking them to implement a notification pop-up regarding 13/14 CPU issues. The most easy way.
@@der8auer-en Both, Windows and Linux are capable of updating microcode on the fly. Microcode updates are not persistent so they have to be applied on each bootup. That's why it's better to do the update with MB firmware, but OS update is also good solution. Ask MS to release an update and add required blob to Linux kernel and you have most CPUs updated.
Microcode updates can (and are) provided via windows update. Linux distributions also load updated microcode early in the boot process. As long as Intel talks with Microsoft and pushes that update to all supported windows versions, most users should have a patched system. The real issue is how they are going to deal with the early degradation of the CPUs. Extended warranty? Voluntary RMA? How about high performance laptop parts? They also have very high boost clocks and voltages for single core loads. How is the degradation there at high temperatures? There's still a lot of work to be done to understand how widespread this issue is. I don't believe that it affects just some higher TDP CPUs because other CPUs also have the same boosting behavior, with a little less voltage the lower you go on the performance tier scale
Yes windows forced microcode likely in october. Allegedly, laptop is not affected. By all reason, laptop is tuned a lot more towards power economy than top end performance, so it's plausible. But given Intel's recent conduct, i'll truly believe it when i don't see it 3 years down the line.
@@inkredebilchina9699 depends where you live at. Here I can only apply for an RMA. If they see that the part is damaged they can replace the item with the same or equivalent or give part or all of the money back
I did an RMA for my 13900K and I have to say that it was a pleasant experience, I contacted them august 21st, they replied with some questions about my PC parts and settings, I answered and they approved my RMA request. it took 4 days from initial contact to receiving my new CPU (they sent me a 14900K) The 14900K was shipped the same day they received my 13900K
Intel can distribute microcode updates through Windows Update, without touching the BIOS/EFI. When the system boots, the UEFI loads its microcode version onto the CPU, but when the OS boots, it can load an alternate version. Theoretically, if the issue really is just down to the CPU requesting too much voltage, then in a few weeks, most systems affected will silently get their microcode patched in the background and halt the damage. I can't help but imagine that a sizable number of users will end up with rapidly-aged CPUs with a couple years worth of voltage margin left. They'll hang onto them for a while, probably planning to upgrade when 16th gen or Zen 6 come out, only to have them die, and they'll be none the wiser. Even if Intel can fix this, we're just beginning to see the extent of the fallout.
I can't see any way that any OS can overwrite the very microcode it is already running. The UEFI BIOS can definitely do so, hence the need for this to be incorporated in a BIOS update.
@@ColinDyckes I wasn't speculating; I was making a statement of fact. See: Intel Microcode Update Guidance "Early OS Microcode Update The operating system (OS) should check if it has a more recent microcode update (higher update revision) than the version applied by the BIOS. If it does, the OS should load that microcode update shortly after BIOS hands off control to the OS. This microcode update load point is called the early OS microcode update. This can be done by an early boot software layer such as a Unified Extensible Firmware Interface (UEFI) driver, a bootloader, the operating system, or any other software layer provided by the operating system (like early startup scripts). The early OS microcode update should be done on each core as early as possible in the OS boot sequence, before any CPUID feature flags or other enumeration values are cached by the OS software. It is also required to be loaded before any user-space applications or virtual machines are launched. This is necessary so that the update is loaded before the affected microcode is ever used, enabling any relevant mitigations for potential vulnerabilities as early as possible. If the BIOS does not load the most recent microcode update, Intel recommends loading that update during the early OS microcode update."
@@Zigglegarf That just seems like a recipe for disaster. I can imagine that system getting hacked by nefarious agents who will write code that can then be injected directly into the CPU without the user being aware. This also counts towards Microsoft itself injecting code without the knowledge or consent of the user as an "update".
@@ColinDyckes Not necessarily, it'd need a second place to stash the update and a mechanism to switch to it on the fly, but quite doable. Just not trivial to implement, but then even a simple modern cpu is far, far more complex.
@@thephantomchannel5368 Microcode is a bit different than most software and frankly if you can inject malicious microcode during os boot you probably don't need to.
The biggest issue I see is CPUs failing after the warranty has expired through no fault of the user. Another issue is, say Intel increases the warranty to 6 years to cover this, will Intel be able to RMA these CPUs with something that will go in the same socket? Presumably this generation will no longer be manufactured or have any remaining stock. That could be a serious issue.
Should I RMA now or wait till August Microcode Update? I am currently on latest bios 1402 with my Dark Hero motherboard but my 14900K is still unstable.
Running 13600K all core 5.7 P 4.5 E 5.0 Cache 1.38v with no issues. I would say RMA now because those cores which were boosting to 6 ghz at 1.5v on your cpu are now permanently damaged. You should not go over 1.4v on any cores. Just lock everything on the same frequency at around 1.4v or below when you get a new one or it will happen again.
@@club4ghz Thanks for your input man. Since I am paranoid now. Could you give me the exact bios settings I should use on a Dark Hero 1402 bios once I get a replacement chip? I should lock all p cores to 5.7 ghz you say. Should I use Auto svid behavior with auto cpu load line? I hear Intel failsafe svid is dangerous.
@@cemsengul16 It's hard to say what exact settings you need to use without studying your board and CPU first. I'm on ASRock board and after tweaking everything on mine I have 0.92v Idle 1.38v gaming. 1.26v stress tests (downclocking itself) 90c TJ max 200W PL1 PL2 locked. (Default is 180W on 13600K) C-States - Disabled Intel Speed Shift - Disabled = (This gets my Idle voltage down to 0.92v) Voltage - Auto, FIVR - Adaptive +205 Offsets P-E-Cache. It's so much offset added because my CPU stock voltage at 5.1 ghz is only 1.2v. I'm on LLC4 to have voltage drop together with clocks under stress. You can experiment with load line to get voltage drop in stress tests, but you need to be careful with that as setting wrong load line can overshoot voltage and damage your CPU under stress tests if it's set too high. LLC may be completely different on your board than on mine. Always monitor your voltage when tweaking your CPU and running tests.
I don't know if this effects 12th gen, but anecdotally, I have a 12th gen system (laptop i7 12700h) and it has to be one of the most unstable systems I have ever used, frequent blue screens, random crashes, constant popping crackling audio (very widespread problem with 12th gen mobile) I think i'm going to throw this thing at a wall and buy an AMD based laptop.
I know this issue affects 13th gen and 14th gen Intel chips, particularly the 13900/14900 parts, but since Raptor Lake (and desktop Meteor Lake) is really just an evolution of Alder Lake, is there a ramification for 12th gen also (yes, Alder Lake runs at lower power levels and frequencies)? My 12900 system appears to be working fine, but perhaps it is also suffering accelerated aging and I wouldn't know about it? Thankfully I also have a 7950X system.
If I'm not having problems with my 13900K should i upgrade the Bios when this new microcode comes out? Im undervolting my CPU to -0.75v, my PL1/PL2 are set to 253W and ICCMAX is 400A. The PC was built like 2 months ago and till now no problems.
I have a 13900HX and have had zero stability issues or signs of failure, is this something i should still worry about, or did i perhaps get a sample from an unaffected plant?
It's because it affects such a small proportion of their CPUs. According to The Register article this morning, Desktop CPUs are about 20% of the x86 market, of which about 10% are K skus, of which only a certain proportion are affected. So you're looking at less than 2% of all their processors. Even if they issue full refunds on all those K skus, the financial impact will be a blip. It will only start being a problem if it affects their laptop CPUs or if the OEM suppliers are affected (Dell, Lenovo, US Government etc)
(from an engineering perspective of course it's a disaster, these should have been caught in Release Testing or V&V - their biggest disaster since FDIV. But financially, it's just a blip, people forget how tiny the enthusiast market is in the grand scheme of things)
@@peterwstacey all different versions of 13th and 14th gen are affected, even T variants, not just the K ones, just to a lesser degree as they're not pushed as hard. Because it's not just the K cpus OEMs will be affected if a noticeable amount fail before the usual 3 year cycle in businesses. It just showed up on busy servers first as they're pushed hard and run 24/7.
@@Bareego 13th and 14th gen laptop CPU's are also affected (to a lesser degree just like T series) because both laptop and desktop share the same architecture and since server market was the one who was first to report issues i don't think this is just a blip considering companies like citadel put pressure onto intel to fix this or else and SI's are not a small thing because many people choose a prebuilt over a custom PC which is main source of intel's income
@@Bareego Let's see. Intel are only acknowledging it on the K skus for now. If they acknowledge it on the rest of the product stack, as the internet rumours say, then expect the analysts to start to make noises 👍 Edit - also note that those OEMs and huge contractors like the US Government do a X-year lifecycle anyway, so if the non-K CPUs are affected but their median lifespan drops from say 20 years to 10, they won't care. So it depends
Why is nobody talking about gaming laptops? The cpu is soldered on the motherboard, if there is a degraded cpu it is a 2500+€ damage. How will intel deal with laptop owners?
I think so. But I just locked the voltage in the BIOS and now I have 1.4v peak in idle and 1.2v under load at 200w max whereas before it was going up to 1.5v at idle. I'm waiting for microcode and some watercooling shit for slightly increase watt and see what's goin' on. it's on a 13900kf
How do you lock the VCore? I built a PC like 2 months ago with a 13900K CPU, im undervolting my CPU to -0.75v, my PL1/PL2 are set to 253W and ICCMAX is 400A. Till now not a problem but im curious too, if i have this settings should i update Bios with the new microcode when it comes out?
@@MrSethGr hey, its an Asus Z790 Prime A WIFI. I have now the Actual VRM Core Voltage set to Offset Mode and i undervolted to -0.75 because that the max i can go on this CPU. Instead of Offset Mode i should set it to Manual and then input manually a voltage under CPU Core Voltage Override? My pc is not old, its now about two months since I built it and runs fine but i dont want my CPU to get bad. I do have some good Cinebench R23 scores with the CPU undervolted, i can even get 40K if nothing much runs in the background!
Regardless of the problems now, i enjoyed this good reporting/education thanks! What i like about all this is I think we've all be thinking for a while that intel and AMD have started shaving the margins a bit closer. i'd like to get back to a proper engineering buffer. I dont' need every last drop of performance. That should be opt in. I want stability. Not xeon stability (4 cores and 6x the cost) but something closer to what we had 25 years ago with consumer where the boars and cpus came with sane defaults and had most of the performance and they generally never failed for 10 years opr more.
It just says 13th and 14th Gen Intel everywhere online... But the stories started with 13900k and 14900k... Are all 13th and 14th Gen CPUs affected? Roman is also mention "all CPUs above 65 Watt" which confuses me even more...
It seems to affect i9s first due to their higher power intake however it seems to be affecting low tier CPUs just at a slower rate. With reports of i7s and some i5 apparently showing signs of degregation
As I understand, Windows and Linux can "update" microcode before fully starting the OS. The update is not permanent. The update is loaded before the OS startup. So as long you do not stay in the BIOS (which will not have the microcode update), you should be fine.
Thanks for the balanced and perhaps less strident report than some other channels recently. The fairest way forward I think would be for Intel to offer an extended warranty on the parts potentially impacted by accelerated ageing.
With all respect to the work you are doing, the microcode update will not solve the problem. First of all, it just postpones the CPU degradation to the date after warranty will expire - the same damage controlling tactic that nV did to the laptop GPUs ~10 years ago. They don't want to earn their responsibility and lose too much profit. Second, if the microcode update is slowing down these CPUs, then it is not the product that you wanted to buy, it is not the advertised one and the benchmarked one.
I need a recall and full refund…. Class action lawsuits are USELESS. All media need to press Intel and give them bad PR. I want my board refunded too…. Spent too much on a z790 apex encore just have 3 CPUs swapped core i9 14900k
Yeah but ASUS boards are tricked out garbage as well with weird bugs, instability and most probably in typical ASUS ROG fashion, the some parts of the CPU are silently getting hammered with high voltages to maintain stability.
I am with you bought a 14900k and a Z690 Maximus Hero EVA edition, but I also what half my HDD refunded as I have 4 western digital black 6TB in raid 5, amd doesn't support raid 5 so I will go from 18TB in storage to 12TB
keep dreaming, Intel would only be willing to refund the cpu or replace it but the board that's a different product that has nothing to due with intel they would not refund a board
I’ve go a question. When a cpu is being made aren’t there a whole bunch of “machines” (fabs?) that are making all the dies all the cores in a factory? Is the problem that Intel is facing all of those fabs or just one that was messing up when it was making a cpu? I don’t know all the terminology or processes of cpu production, just trying to learn.
most likely using the extra time to bin more and getting the sku's a little higher, as this will be a delay Intel cross the broad, remember this problem could be in server die's also and if so the enterprise customers we be put ahead of DIY market. Now the fastest way for Intel to get usable replacement cpu out there, is to get TSMC to make intel "12gen+" CPU Intel of cause will call them 15 gen Re-Fresh-Lake CPU on TSMC 10nm or 7nm and Intel will also need to make a MB bios/firmware update and later on they will say that 1700 socket was 4 gen socket
I would imagine that intel figured that the CPU's would last just long enough to get them out of a warranty claim. That has obviously majorly backfired on them.
Intel can definitely address the challenge of informing Windows users of having a potentially defective 13th or 14th gen CPU via Windows Update. Intel can work with Microsoft's Windows Update team to detect systems with such CPUs and an update can be applied where a pop-up appears on the desktop screen after every reboot/logon displaying a warning message about the installed CPU with information about the need to install the latest BIOS for their motherboard and how to get in touch with Intel to RMA their CPU. The pop-up should only be disabled once the user clicks to select a check-box stating that they acknowledge the issue. Microsoft have 100% done this before with regards to cetain Toshiba laptops that shipped with faulty OEM Dynabook AC power adapters, so there is no excuse that it's "too hard" or "can't be done". Great video by the way!
Whole issue should be reviewed from original timeline: 1. Intel 13th Gen release; 2. Intel spotted problem with Via Oxidation ("addressed back in 2023" as they said), but hide it; 3. Problem with limits appear and Intel blame motherboard vendors for it; 4. Intel fixed problem with limits by new BIOS and profiles; 5. Problem with CPU appear again, which isn't voltage (Via oxidation which Intel hide). So, half released CPUs might be fixed with microcode since they don't have via oxidation problem, but first batches can't. They should be RMA. Intel just need to share manufactured dates\serial range of this CPUs.
Intel has never said that oxidation was problem for all chips that were made in that timeline. They said there was a problem in 2023 and we do not know how many chips with the problem made it out of fabs and we don't even know how long that problem lasted or how many fabs were affected. What you wrote is pure speculation.
I purchased a 12 13, 14 gen Thermal Grizzly contact frame and received some plastic 3d printed thing. Its says Thermal Grizzly on it.?? Is this really Thermal Grizzly product? How does it not deform under heat?
Im sure intel did this to make the intel fans buy CPU more often so they always have the latest and best performance. Like they gave them 2% performance each year over 15 years. Intel is such a kind company :)
Perhaps degradation with time is purposeful. As chips become more powerful manufacturers will be incorporating planned obsolescence. Todays i7 or i5 can function a very long time in a simple task home machine. Most desktops do not play AAA games. I still use a 4790k from 10 years ago. It is perfectly fine for normal non-gaming use and will remain so for another good several years. This presents a big profitability problem for CPU manufacturers. I see Intel so hesitant introducing a six core i3. It would be a 100 dolar chip with potential to function in a simple task computer for 15 years or more. The industry is maturing and planned obsolescence for CPU's is the likely course.
Had to RMA my 14900kf a few weeks back. Bought in in January and by April it was super unstable and pretty much unusable. Ended up figuring out if I disabled P-Core 4 that made things run much more stable. Got CRC errors when unzipping NVIDIA drivers, but when that core was off, worked fine. RMA process was fairly painless and quick. New CPU seems to be running much better and I already have the microcode BIOS update in place, so hopefully it'll last more than a few months. Intel better get this figured out.
I've got 3 friends who recently bought pre-builts with intel CPUs (13900k, 14700k & 14900k) following advice a from a mutual friend who ''knows tech'' - one of them I've managed to give a warning too and they've already dropped the voltage as recommended by others, but the other 2 just don't care ''it turns on fine, I don't understand what the issue is, I'll worry about it when it goes wrong' and don't understand about microcode or BIOS settings - going to be fascinating watching their machines over next few months. When the first person contacted the company that built the PC they said they'd tested everything, weren't aware of any Intel problems and they'd need to speak to Intel directly if the CPU had stability issues. Interesting way to treat consumers when this issue is out in the public and has been for a bit.....
I have a 14900K and I'm kinda freaked out. If this is a manufacturing flaw, yeah. This could be a massive problem for Intel. It means they sold a defective product. This could be a very expensive mess for them. I'm tempted to just slap my 12900K back in my rig and RMA my 14900K out of an abundance of caution. I haven't had serious issues other than game crashes when the power limits issues popped up.
If you havent had issues, dont do a damn thing. Again its not ALL cpus, its a specific set of CPUS more then likely from a few of the fabs. If you start having issues yes, you will have to RMA it. But again, thats IF you have issues.
@@64BBernard it sounded like he was having issues. Game crashes when power limit issues pop up. Personally, I wouldn’t take my chances. If I can get a replacement, I would. It’s a lottery of potential future issues otherwise.
I really think it needs to be a recall at this point. I'm also concerned about BGA parts - Intel says laptops aren't affected, but that would also a much more expensive problem and I'm not sure we should really believe them.
I have a 2022 13900k Intel cpu. I got approval for warranty replacement yesterday but I had to ask if it'll still be covered since it is delidded(largely because of stability issues despite already having a custom water cooling system). No answer from intel yet. So now my CPU may have issues with oxidation and definitely has stability issues since i can't overclock without issues anymore(could sustain 5.8ghz all core and 6.1ghz 2 cores daily) now I've had to drop ram to 6000mt/s and stock cpu clocks. Can't run 6400mt/s or even a 100MHz OC on any p or e core. Still have some stability issues that aren't explained by anything else. I knew delidding had risks but i didn't know intel would do this either. Guess im waiting for 12p bartlett lake or switching to AMD (if they are issue free in 9000). At least it is pretty stable for work, just cant use my water-cooled 4090 for games 30% of the time.
@@WrexBF Yea it's well known. I'm actually pushing them on it since delidding hasn't affected anything and it's only their own failures causing this. Normally you don't have to worry because CPUs typically just don't fail. I got a weird response and waiting on another now.
@@gagarin777 I saw that, so kind of hoping reviewers don't get golden samples since they supposedly are dumping the first batch. They probably knew it'd cause a huge backlash as this Intel issue is ongoing.
Yes, in this case user's knowledge does matter too. If you always checking your cpu voltage or familiar with OC/BIOS tuning troubleshooting stuff since day 1, you should already lower voltage if too high or something weird happens. The cpu who get aging, damaged etc may mostly happens to those people who completely never checked their cpu voltage etc.
If you're on an lga 1700 system with 13th or 13refresh cpu that's acting odd it's probably the cheapest fix with good performance. I've heard (NOT VERIFIED) that 12900 k's were having issues a smidge above normal levels, but that could just be the node or arch or both.
Thanks for the detailed breakdown. I have a question. If I have a 14900k that is unstable and affected but it was lapped and polished in a last ditch effort because I was having very big temp diff between cores. Can I rma that ? I never took the lid off. I still have original box and receipt from microcenter. It’s 4 or 5 months old at most.
You can try, but they'll possibly deny the warranty because it's lapped... This whole situation sucks for 13/14th gen owners, especially those going to more extreme ends on their systems. I have a delidded 7950x running bare die on a WB with LM, I know that any warranty claim would have been void at the time I delidded it, but with a manufacturing defect, they should still honor it. So my advice would be to try, but don't have high hopes they'll approve the RMA. If you have screenshots, etc from the instability and high temps before lapping, then that may help to convince them, wish you the best of luck.
Rather than electrical conductivity problem the main issue with oxidation is the mechanical stress exerted inside the silicon packaging due to bulging of the affected via.
Does the latest news about potential leaks from Intel indicating that the ring bus is getting cooked by the voltages further aggravate the issue. Will the microde correct for this?
Or maybe Intel just realized, that building CPUs that last for 10+ years is not the best business strategy anymore. They just overshot a bit so these CPUs now already fail withing the warranty period.
Uhh, does this mean windows do not distribute CPU microcode updates? Because in linux I have microcode updates as part of regular system updates, so I would expect new microcode to be automagically distributed and updated without need to explicitly ask users to do something on their own.
Nice, interesting and technical discussion about vias and the entire concept. Also, it will be interesting to see second hand marketing prices after this situation, if Intel doesn't recall or RMA their 13/14th gen CPUs.
The more that comes out about it, and the more that ppl talk about it, the more I think this is not only a major hit to intel in general, but also to their efforts to get 3rd parties to use their foundries...
Thank you for doing a well researched, balanced and non-biased video on this subject! Don’t make reference to GN’s videos. Steve’s running a tabloid UA-cam channel. Does not compare well to what you are doing. Thank you once again!
I literally just ordered a 13700k and platform a few days ago...I haven't recieved it yet, but i'm wondering if a use a static voltage if it will mitigate the microcode voltage issue or does that not matter? I could probably return it, but I really wanted an Intel system...
Imagine 2nd hand market flood with faulty CPUs.
Like socket 1700 will be a least popular platform for aware customers and at the same time a doom trap for unaware ones...
Maybe this is the moment AMD can finally establish a much larger market share, not by selling more new CPUs, but by Intels second hand market collapsing due to these faulty CPUs.
there goes my resale value lol. I have kept my 13900k at low TDP I think 95watts? Good temps no crashes. I never tried to OC and dump a lot of voltage on the CPU anyway...on my first few runs I noticed how hot the CPU gets and I began turning off the high bios voltage
@@Fergunator just sold it, before its value drop
@@scanlly Nobody knew about this issue in 2022. Intel found out in 2023 and didn't say anything, and just this year the gaming community and a few server companies spoke up about instability.
yeah recently made a new build. First time using AMD and went with AM5 for future-proofing. So glad I tried something new!
Something tells me " could be" is an understatement... When I worked for hp at one of they're dev' fabs we had contamination problems and we developed a plasma cleaning process to remove the contamination but had to scrap a lot of wafers ... Once contaminated the contaminants can diffuse into many insulating or conducting structures... there's no real going back ... Any contaminated CPUs will only get worse no matter how much intel adjusts the voltage/ power specs.
Great!
Looks like my $500 i9-13900K is just an expensive paperweight. I have a 30 percent failure rate (freezing, crashing to desktop) while gaming.
And that is the problem isn't it, once it goes bad, you can slow it down but not cure it.
@@MasticinaAkicta It's the osteoporosis of integrated circuits.
@@Baldmaxx Do you buy it just for games or more than that?
@@Baldmaxx For the time being, you can use a program called Process Lasso, to find the core that is bad by using Nvidia's (chinese site)driver selecting webpage. Use the drop-down menus(click around, select random cards etc.), it will cause the web browser to crash. Isolate the power core(select the web browser process in process lasso, set core sets through affinity) that is causing the web browser to crash. Stop your games from using the bad core using process lasso. This is only a band-aide of course, but I've had success with this method and making Star Citizen crash free. Star Citizen is one of the most demanding games CPU wise.
Another way --- In Bios- Load optimized defaults, lock your all P core multiplier to 55 or 56, increase load line calibration to level 7 (Asus bios), set a core voltage + offset to target 1.4 volts (be careful, use increments of .05 -save/reboot-check voltage with CPUZ) or lower. You may be able to get away with less, depending on the degradation of the cores. Use Intel XTU AVX2 stress test (5 minutes) and test for failure. Adjust your voltage only lower than 1.4 or lower your core multiplier to achieve stability.
My 13900k nightmare continues. 🤬
Me too!
$500+ down the drain....
Your fault.
You knew Ryzen was better but insisted on giving your money to a convicted monopolistic anticonsumer company.
Just stop buying intel. Been fleeced too many times by them to buy any more of their products.
Lol Ryzen has been better but you held on to your pride.
@@grtitann7425 I knew AMD was better, but at the time AM5 and DDR5 pricing was through the roof in Canada. Saved over 1k staying with Intel and the DDR4 I had. After RMA - I'm selling it all and picking up a 7950x3d. Screw Intel.
my theory: Intel knew about the oxidation problem but hoped it would effect only out of warranty CPUs
Intel just following Apple Footstep, heck the number even similar. 13 14 got so much problem like screen and overheat on iphone
It’s not a “theory”, it’s a brain fart. The best it can be is a “hypothesis” but definitely not a “theory”
@@aliancemdthere are multiple things that are implied and one just needs to know (never admit fault so it is much harder to sue and no recalls are required, always try to make it look like user error for maximum profits. To mention a few). But I'd say it is pretty solid theory. Just keep the mouth shut and react minimally when public pressure is too high.
Apple did replace all of the Nvidia defectives, that's why they're not on speaking terms. They also did respond to antennagate, and the i6 being a bit _too_ futuristic (is that PTSD why they don't do foldables? lol)
The only one I'm not too happy off is the superthin keyboards which make the Core m Macbook and the 16-19 MBP no-go devices, but I guess they just _really_ wanted to stick to that design because slimness tunnel-vision, and I guess it must've been mitigated enough because it seems to be a forgotten issue.
@@Fay7666 what about the failing display cable that was too short? only few models were considered "defective" while others didn't.
Intel's handling of the situation is becoming as big a story as the CPU issues themselves. Great work as always guys.
@@scanlly the scale of this issue was not know back in then, they published the information as they learn about it. Tech UA-camr don't just talk about it now to make a buck, they talk about a known issue so who plan to buy a new PC or a new CPU know about the existence of this issue. Also, would you really be ready to say that Tech youtuber such as Gamer Nexus that literally offer to buy defective CPU from viewers to send to an failure analyst lab at their own cost to investigate on this issue and see if really it was a via condensation issue, does this to get rich? Those kind of test cost lots of money and the lab literally told him that with such a complex CPU, there was like 50% chance of not being able to see the issue.
No Tech Channel ever claimed to know everything, they share information as they learn about it. Like Der8auer said, this issue is an issue that is developing and get worst with the time, they could not know that there was an issue 2 or 3 years ago like you said, especially not at this scale. We just learn from Level1Tech about 2 week ago how wide spread this issue was, based on Video Games Telemetry and reports from server company and PC company, before that, all they knew was that some people was experiencing some issues and as always Intel play the card of "Small amount of people" in their public statement.
Intel will release the microcode after Ryzen 9000 release after all the reviews are done.
@@scanlly the rage bait merchants like gamers nexus & hardware unboxed only knew about the issue because some server owners leaked information to a completely different youtuber first. they had no idea, but now they're making thousands off clickbait nonsense and attacking intel for not coming out without knowing what had gone wrong and how to fix it. the dumb thing is the voltage issue only impacts people too lazy to know what their technology is doing, and didn't set manual voltages & clocks.
@@bigpoppa1234 Too lazy to know, you mean like people that are just using a computer with that processor that are not into bios-es and hardware tech... 99% of people basically. Ofcourse they don't know and that shouldn't be the reason not making any noise about it, it's an issue that needs to be addressed, Intel is just a company.
i misread "handling" as "mishandling" and it changed literally none of the meaning of the sentence...
Intel's three-year limited warranty should be immediately changed to 7 years.
This…. Or give us a refund. Remember guys Intel had dominance for years. Don’t give it to them, they HAVE THE FUNDS TO TAKE care of their customers
First, it shouldn't even be a limited warranty but changed to a full warranty!
What warranty? Its only on paper.
Actually offering replacements is a different story.
@@Multimeter1intel would need to refund or compensate for the motherboard purchase too
@DarkStarProGamer you clearly have no idea what your talking about. 👎👎😒🤦
Disagree on one opinion: No matter what caused the permanent damage to CPUs (via oxidation alone or “faulty” microcode resulting in too high voltages or a combination) this should absolutely warrant a recall of every unit already in circulation.
It's interesting to contrast this with AMD's actions just today. They have apparently found a QC issue of some kind with Zen 5 and are literally recalling every single unit shipped for testing and possible replacement. Yes, it's much easier to do in the run-up to launch but that doesn't eliminate the contrast.
@@Lemurion287 Do you really think that Intel wouldn't do the same if given the chance? A recall at this point is not feasible.
False. They should only extend warranty for the small batch that was impacted.
@@the12gaugeshotty its not over yet.
@CoffehKittehGrey I think board readiness is a red herring. The whole point of AMD's socket longevity approach is that the new processors need to work on existing boards.
intel: what?! you want us to go out of our way to help customers make their cpus last longer instead of buying a new one in a few years? having it die just out of warranty is what we call "the sweet spot" 😀
they're just helping make the choice to upgrade easier. Thaanks Intel!
Laptop motherboard strategy
wow bro, you just made a major discovery…what’s next on your list? the earth being round? or that it revolves around the sun? we got a genius here people😂
@@medovk tha actual fuck are you on about bud?? Op comment had nothing to with discovering or even figuring anything out lmao
Huh?
Intel chips last for decades. They've never been concerned about you buying another chip in a few years.
The only people I know that buy a new chip every few years is AMD users.
"could be" is rather optimistic.
No problems here. Oh I’m an enthusiast.
Yh its not like the server cpus have 100% failure rate lol
@@Typhon888enthusiastic about contributing to user benchmark?
@@Typhon888 Don't overdose on that cope.
@@scanllybruh all the big tech YTers have been constantly recommending gamers & builders to use AMD. Roman is an electrical engineer, he’s reporting on the news like the rest of them calm down , AMD has enough marketing & aren’t going bankrupt. Your issue should be with the companies like HP, Dell, Acer, Asus, IBM, Lenovo etc. who will continue to choose Intel over AMD for their laptops & PCs. The “PC Gamer builders” make up less than 10% of the market , you’re quite literally preaching to proxy AMD marketers & their followers.
Intel wants people to forgot so they can deny future RMAs or be like, of you updated so must be something else. This is why recalls exist. There's no trust left.
yup, all aboard the slippery slope
Forget what?
They're literally acknowledging everything.
😂🤷
@@DingleBerryschnapps It took them over a year to acknowledge that there was a problem, then they lied about it, backtracked, and have been silent ever since. They knowingly sold defective CPUs and hoped that people wouldn't RMA inside the warranty window. It blew up in their face and now nobody trusts Intel
You can't trust their next CPU Gen either. If the problems surface over a year later. No one can test the CPUs reliably. And you can't even be sure it won't affect their GPU branch either. Because no one knows the innards of their fabs and their production line in general. Thats the general problem of cross contamination.
@@DingleBerryschnapps No, they're not acknowledging everything, it's been over a month since you left that message and WE STILL DO NOT KNOW THE TIME FRAME FROM THE EARLIEST KNOWN DEFECTIVE UNIT TO THE NON DEFECTIVE. Which means we're still in the dark, people that bought 13th gen and 14th gen still do not have a clue if their CPU is affected by the oxidisation problems.
Easiest way to reach the majority of users with Intel affected CPUs is to contact Microsoft and display a message in widows if the system has a CPU with possible issues.
Absolutely. But I am sure it ain't gonna happen. Think about the millions of schools or work computers of people not aware of nothing of the tech word looking at a message that you may or may not need a CPU RMA..... Then think how it would hurt Intel for the next time these government companies need to buy something. They would make sure it is not Intel
The problem is that it isn't MS fault, problem or even any business.
They should really issue a recall.
And the message should be displayed on a blue screen, of course
@@Tautolonaut "temporary licensors" .
So ask microsucks to spy on users and spam them
Excellent piece, really appreciated the technical explanation. Perhaps an interesting observation from my end: I’ve built quite a few passively cooled configurations, and one thing that stood out was that 12th/13th gen have much elevated power consumption past 85C. So with the workload remaining equal, something like a 8700 or 10900 would show for instance 50W power draw and that would increase to 55W when closing in on 100C. In other words: the CPU would become less efficient at higher temperatures, which is logical, because it pushes higher voltage to remain stable. Now for the 12th/13th I observed a marked difference, where past 85C the power draw increased much more dramatic, getting towards 70W at the 100C point.
This apparently aggressive ramping of voltage is something ‘new’ for 12th gen and it looks to me Intel took a gamble here, where things quickly go south if there are even minor inconsistencies in the production process. The microcode update is therefore more of a mitigation, but like you said: the average user will take the blow for this, because it’s likely they will be stuck with an effectively broken part (and PC, ultimately).
Warranty in France is 20 years for hidden defects before sale, and 2 years from the discovery of the said defect. The Europe Warranty rules allow each members to implement hidden defects warranties. I don't know how widespread they are in other countries.
This is complicated. What ultimately provides the cpu with voltage is motherboard, and most high end boards on the market tend to add more to the vid given by MC for higher boosts. This strategy is especially obvious when XMP is active.
the issue is intel discovered the defects after their initial sale. so a large range of those would be subject to the 2 year warranty
@@dead-claudia What? If you bought the CPU and Intel told that there's a defect after the damage is already done that's still a hidden defect.
@@dead-claudia intel's own warranty is already 3 years minimum
Oxidation would almost certainly have been caught in accelerated wear testing. The problem with the excess voltage is that it causes wear of the transistors themselves through a process called time dependent dielectric breakdown, and this is independent of current and less dependent on temperature than electromigration, it is therefore more likely to pass in the conditions of the accelerated wear tests. Why exactly they didn't properly monitor the voltages during operation to be within the intended range on the other hand is quite perplexing.
13th gen was only launched in late October 2022. You're saying that in 2 months you returned 8?
@@scanllyyou returned 8? So you're telling me after the 3rd one you didn't ask for your money back?
What are you doing dude
Wait but the oxidisation issue is literally confirmed by Intel in a Reddit comment tho
@@GreenJalapenjo The way Intel is describing it makes it seem like a 'limited hangout' where they disclose a real problem that they had and solved while trying to distract from the much larger problem (that their cpus are running unsafe voltages that degrade them rapidly).
I don't think it's plausible that they didn't know the peak voltage reached! More likely they retroactively had to conclude that the semiconductors can't sustain as much voltage as they thought they would.
913:05 it's easy: they will not. The average customer will be clueless unless the store they bought it from clues them in. And especially beyond the (in many places quite short) normal warranty period, the stores have a stake in not telling them anything. They would have to deal with returning, arranging exchange, etc. Intel is not going to foot the whole bill for that, especially not if it's 2-3 years down the line. Oh, and not 5-6 years. I've just moved off a Haswell 4790K that was fine, but received zero OS support sadly. It even had a modest OC all these years. Rock stable.
My i7 4770k was liquid metal in between the IHS and the die with a 4.4 GHz overclock for several years with absolutely zero issues. Those days have long passed
@@scanlly many of us smaller companies have been shouting off the rooftops for years about the instability related to these CPUs. We were told that we were just hating on Intel and AMD Fanboys when we sell both products alongside each other and actually see the failure rate and abysmal RMA process
The average customer doesn`t have high end cpu`s they have looser cpu`s.
The people who have those high end cpu`s are mostly very well informed.
I had a 4690k for years at 4ghz. I have gone to a 14600k. I have only had crashes from over clocking. Few bios up dates have helped. I have it stable at 5.5ghz + .030 volts on ETU.
@@scanllyyou have no idea what's going on and just want to cry about youtuber don't you?
if it were just a slow decay from corrosion, they could "just" replace customers CPU with new ones. But if it's a fundamental flaw, you can't. You could give out 12th gen as a temp "solution" but if you, for example, change motherboard and chipsets every generation....
God i LOLLED so hard on your last sentence!
Intel shooting themselves in both feet by themselves.
Didn't they say that the oxidation were only on 13th gen and has been fixed? And since they have moved unto 14th gen, I would say that if you are affected, then you should be able to go to their site, as Steve mentioned, to look up if you have an affected CPU, and then you should be able to RMA it directly to Intel, and they should be providing you with an equal or better version of the product, meaning if they can't give you a 13th gen, then you should receive a 14th gen version of your CPU, which I believe could be "cheaper" for them to just say that 13th gens have a "no questions return policy" and to replace with 13th or 14th gen versions, if someone RMAs a CPU to a retail store or so. But they should never be able to give you a 12th gen CPU as a replacement, as that would be a downgrade, and not be in compliance with a product of equal value.
@@TigonIIISupposedly, yes that's what they're claiming but 14th chips are rapidly having issues too. For the rest they're saying it's just voltage issues. Somehow there's a VID algorithm issue or something that took them this long to address. This would also cause accelerated degradation. Either way it doesn't really matter except that in theory these chips that aren't already damaged enough to fall out of spec could be fine after a microcode update....
Bartlett Lake is existing exactly for that.
@@TigonIII Well if the oxidation issue is only for 13th gen, what causes the 14th gen issues?
Intel: 🔥🔥🔥This is fine. 🔥🔥🔥
But they haven't said it's fine. 🤷
@@DingleBerryschnapps bots have trouble with sarcasm eh?
I use to work at Intel, and most of the people I knew who chose to leave did so cause we could tell there was no real plan. For example in Litho we had to explain to our manager what our job was, what we could and couldn't do, and even what the alarms meant on a daily basis for months. It would get so bad that people would yell at him "I CANNOT DO THAT. IT WILL BREAK THE MACHINE." Beyond that they constantly lay off experienced engineers to hire new ones fresh out of college so they could get tax breaks from the state government. When you get rid of all the experienced people this is the kinda thing that happens. We all knew it was coming.
It sounds as if they have followed Boeing's example and replaced engineers with business economists (fresh from university, of course, because those are cheaper).
They need to extend the warranty period of all 13th and 14th gen cpu. this is the only way to do right by their customers
This would be disaster for stock owners so sorry but nope. It is always better to piss off customers than reduce next quarter profits. That is the logic why customers are always screwed no matter what they do...
@@timop6340 well they are already loosing market share to amd, if the problem becomes worse and more apparent turning their back on customers would be a death sentence to their stock price
der8auer: Watch GN's videos about it.
LTT: I have never heard of this man in my life
"he could have phoned me"
that s what i was thinking 😂 LTT refuses to talk about oxidation...
"It might last through our small warranty so its fine guys" - Intel
The ending bit of what you said is exactly what I was thinking. My cpu is already damaged from the microcode issue regardless of if it is stable or unstable currently right at this moment in time. I have had this 13900KS since it was released and it runs on average 19-20 hours every day since. I don't see how just telling ppl to update their microcode would make this okay at all. They sold me a product and then caused damage to it, that alone should result in an RMA. When I paid over $700 at the time for this CPU, I was not aware intel was going to cause permanent damage to it for this entire 1.6 years that I have had it. Warranty only lasts 3 years, if mine isn't unstable now, it's probably not gonna start showing any signs of being unstable within the remaining 1.4 years of my warranty period. Depending on how they handle this, is gonna be the decision maker on whether I trust Intel in the future or not. The other commenter that wrote extending the warranty period to 7 years for affected CPU's, is a good happy medium imo.
But it’s not just about approaching customers about a microcode update. I assume all >65w 13th and 14th gen CPUs are damaged to some extent, if they were being used at all.
PL1 and PL2 watts should also be considered.
Fortunately my 13th gen rig has an i5 13500.
You assume wrong
@@j.p.h.8126 Your PL2 is 154W.
ive read some i5-12450HX LOQs died though i am sure it is a low power consumption CPU with 8(4+4) cores and max 4.4GHz for the P cores, is it possible to be affected with the oxidization issue ? i mean the 12th gen is not new but the LOQ varient is 2024 model.
So glad I switched to the 7800x3D when i did. Not that anything was wrong with my 13700K, but avoiding potential issues is nice
the 7800x3D is the greatest cpu of todays - with only 8 cores it beats the entire competition in gaming ❤
Now 14nm Intel chips starts to look like the “ol’ reliable” option.
Well the ones with enough +'s at the end at least.
12th gen is already fine. 10 and 11th gen are useless.
Usually when a company refreshes a product, that product becomes more refined, and a an all around better
Intel 13th & 14th refreshes are doing an irreparable damage to their brand
Regarding the oxidization issue. Intel said in a subreddit post "addressed with manufacturing improvements and screens". Does that mean none of the affected CPUs were released to the channel? If not why not provide a date code etc so that users can verify if they are using any of these affected CPUs? The lack of detailed information regarding this issue is frustrating as hell. I assume that the affected CPUs were all of the 13th/14th gen CPUs produced during some time period of 2023.
The microcode fix is to be released to motherboard makers in mid-August. It is then up to the motherboard makers to release BIOS updates for their affected boards. Will all manufacturers release BIOS updates for all of their affected boards? I am not holding my breath on that. There is also no mention of whether this issue affects ALL 13th and 14th gen CPUs or just the higher-end ones.
It’s mostly the higher end 125 watt ones
@@jeffscomp ive seen i5-12th gen HX in LOQs 2024 model dies so i am a bit confused right now cuz it is not a high end and it is considered low-powered CPU.
what is ur opinion about it ?
So glad I dodged a bullet with the 13900 last year - ended-up with a 7950x for no other reason than sheer dumb luck :/
Good call 😂
Same. Wanted a good gaming CPU for the 1st time in my 39 year life. Went with the 7800x3d and was considering 13700k or 13900k but tbh I just didn't particularly like the sound of E cores for my gaming system.
@GregoryShtevensh yeah 78x3d for gaming is 100% a no brainer. I'll be picking up a 9950x3d when it releases so I can get the best of both worlds. I really hope AM5 is a banger of platform just like AM4
@@InfernoTrees yeah I mean even for basic work stuff. It's been more than enough. I've never felt like it was poor in any particular area. 16 threads has always been enough for my needs that way.
It has gotten pretty hot a few times though. Nothing to really worry about as I have it set to throttle at 81c. It's hit 80c and 79c a few times. Never throttled yet though. I'm in Australia and running a 240mm rad. I think a 320 is the way to go with these chips. But my case couldn't fit one
@GregoryShtevensh yeah I thibk people underestimate how 16threads on zen 4 is still gonna be good for that stuff lmao. Fellow Australian here too btw!
I don’t understand why people go Intel, Ryzen 7000 is more efficient, more performs better in games (X3D) and has comparable productivity performance. While also having a future proof platform supported for many more years.
It's not even close to being comparable in productivity.
I honestly think that it does have to be a recall on the CPUs. Without making extraordinary efforts to replace every CPU that has had a significant decrease in overall life span from this, you basically have a large number of people who will get a lot less time with the computer than they normally would, and the resale value of the systems will be disproportionately lower than previous and subsequent generations. Unfortunately I am not sure if there is a way to detect how much degradation any CPU has until it gets to the point where it affects stability, and if there isn't, then Intel would have to refund or replace every single affected CPU to fix these problems and I am not sure that they can replace enough of them to satisfy a recall. In any case they absolutely will be trying to avoid a recall as that would be extremely bad for Intel if they had to do that, so the only way that will likely happen is if a government forces them into doing a recall, which at least in the US is very rare unless it is an actual safety issue.
If they can sell that many on a new node
They definetely manufacture as much when it already matured and has better yields
I agree on the necessity for a recall on the CPUs. But you are aware that this problem is not limited to the CPUs on PCs with motherboards? They're now finding the same issues on laptop CPUs. Swapping a CPU on a PC motherboard I can do myself. There's no way I can see people swapping CPUs on laptops.
@@andrewbauer1322 first off, Intel is claiming that is a different issue, I am not sure if I believe it and not sure which would be better. Second, that type of thing would go through the laptop manufacturers and they would probably do the CPU replacement.
@@jeffreyparker9396 Agreed. Also, I missed another difficulty regarding CPU replacement in a laptop..."CPU replacement in a BGA socket requires proper equipment and experience. Then there is the issue of BIOS support. Most OEMs BIOSes don't support many CPUs ... usually only the models they shipped with that series. So even if the new CPU is socket and chipset compatible the BIOS may not have the microcode to support it. So while an upgrade maybe possible its usually not feasible." (from a Moderator on Tom's Hardware)
@@andrewbauer1322 while I would agree that OEM BIOSes might not get additional CPU support added in later updates, I haven't really heard of many people ever having issues upgrading CPUs in those systems other than cooling or VRM limits.
My first 14900K with my asus strix z790 itx was definitely affected as the temps would regularly spike to 110c on 2-3 cores and voltage spiked to 1.65v+ this was using Asus AinOC feature though so im pretty sure that was a definite contributing Factor.
2nd 14900K + Dark Hero z790 zero issues gaming or encoding it has the Intel V1 Direct Die Contact Frame from TG and its working Beautifully temps only hit 70c recently because ambient temp went up since its 35c outside. This cpu+Contact frame from TG have keep this cpu at 45c while gaming voltage has stayed at a steady 1.34v -1.35v from what i can see no spikes near the 1.71v of old setup.
3rd 14900KF no issues other then stock ilm wich will need to be changed out considering another direct die since it gives the die 200% more surface area to cool from.
Awaiting word if the Rumored 10-12P core CPUs will be affected or if this generation is another 14nm++++++++ like ole days.
Hey Roman,
My 13900k only crashes when idle...
Thanks for this video, yours were the first one to point out issues not only in high usage, but idle voltage scenarios. I have that exactly. I was debugging for long, but its hard to, because I had crashes only when my pc was idle.
Random software, usually Chrome and Discord would just crash. I mitigated this with running youtube videos in a tab, and the issue was gone, always.
Never had any problems playing games.
Since then to avoid complete instability, on newer and now newest bios by Asus (z790e gaming wifi), and yea that solved the complete instability but sometimes still happening.
12th gen isnt impacted at all by this? I switched from a 12900k to a 5800x3d few years ago after fighting stability.
As you experienced, 12th gen is affected, just not as badly as 13th and 14th gens. Seems like the entire P and E core architecture choice was a mistake.
@@savagej4y241WAY to much powerdraw was a mistake, and the power draw insanity started with 12900k
@@kaspervendler1726 I've started seeing comments under other videos over the past few days where even i7-12700K owners are rallying around disabling E cores. They bought into the P + E core architecture, and now throw away performance instead of managing it with undervolting, no more than 1GHz delta between all cores, and Process Lasso or CoreCtrl or the like. These CPU's require careful configuration but they were marketed toward the lowest common denominator.
I agree they need to be proactive in reaching out to customers, but how do you reach the unengaged group you mention? I don’t see an easy way for Intel to communicate with customers who just set and forget their computer.
Yea thats pretty much the problem. I also don't have a good answer to that
@@der8auer-en Why are you TECH people not understanding that a lot people buy tech online.
To buy tech online you need to make an account on said website for about 95% of the websites.
If you still bought the said hardware without an account, 100% of the websites will ask you for your e-mail address.
Do you see where I am going?
99.999% of hardware bought online can be traced back to an e-mail of the customer.
Intel should contact all the stores that sell their CPUs and ask the stores to send e-mail to customers that bought said CPU.
They sold millions of these faulty CPUs, it's time they invest time and money fixing their dumb, lazy, major accident!
Any time i buy a higher-price item above 100 currency-units at a computer store, they note down my address, they put it on the invoice and they store a copy of the invoice in their system. So by all reason the vast majority of people can be reached by snail mail, they will need to allocate like 3 dollars per customer to pay the stores for the outreach work. The store isn't allowed to share my address to Intel or anybody else, and it's not allowed to mail me advertisement, but it is allowed to contact me with important matters of this kind, especially if you officially designate it a recall. Recall doesn't mean every single item needs to be sent in, they can absolutely issue a recall with a DIY service option.
Most CPUs are in prebuilts. It will be up to the builders whether those want to contact customers for proactive service or just handle the returns as they happen.
Second hand market gets toxic, but that's to the benefit of Intel. They could alleviate it with retroactively extended warranty, without proof of purchase, say 7 years from manufacturing date, but they aren't going to do that.
I think in a few months, everyone will receive microcode update anyway. BIOS isn't the only option, Windows also receives microcode patches via Windows Update, the NT bootloader is capable of injecting them.
@@der8auer-en Like someone else mentioned above, Intel should reach out to Win and Unix system dev-teams, asking them to implement a notification pop-up regarding 13/14 CPU issues. The most easy way.
@@der8auer-en Both, Windows and Linux are capable of updating microcode on the fly. Microcode updates are not persistent so they have to be applied on each bootup. That's why it's better to do the update with MB firmware, but OS update is also good solution. Ask MS to release an update and add required blob to Linux kernel and you have most CPUs updated.
Microcode updates can (and are) provided via windows update. Linux distributions also load updated microcode early in the boot process. As long as Intel talks with Microsoft and pushes that update to all supported windows versions, most users should have a patched system.
The real issue is how they are going to deal with the early degradation of the CPUs. Extended warranty? Voluntary RMA?
How about high performance laptop parts? They also have very high boost clocks and voltages for single core loads. How is the degradation there at high temperatures?
There's still a lot of work to be done to understand how widespread this issue is. I don't believe that it affects just some higher TDP CPUs because other CPUs also have the same boosting behavior, with a little less voltage the lower you go on the performance tier scale
the solution is quite simple. return it while you can and go other way.
@@inkredebilchina9699 You can't return a year old part wtf.
Yes windows forced microcode likely in october.
Allegedly, laptop is not affected. By all reason, laptop is tuned a lot more towards power economy than top end performance, so it's plausible.
But given Intel's recent conduct, i'll truly believe it when i don't see it 3 years down the line.
@@SianaGearz why not? if it is underperforming or have other issues, you certinally can ask for your money back.
@@inkredebilchina9699 depends where you live at. Here I can only apply for an RMA. If they see that the part is damaged they can replace the item with the same or equivalent or give part or all of the money back
Well now i know what to do instead of sleeping
I did an RMA for my 13900K and I have to say that it was a pleasant experience, I contacted them august 21st, they replied with some questions about my PC parts and settings, I answered and they approved my RMA request. it took 4 days from initial contact to receiving my new CPU (they sent me a 14900K)
The 14900K was shipped the same day they received my 13900K
Intel can distribute microcode updates through Windows Update, without touching the BIOS/EFI. When the system boots, the UEFI loads its microcode version onto the CPU, but when the OS boots, it can load an alternate version. Theoretically, if the issue really is just down to the CPU requesting too much voltage, then in a few weeks, most systems affected will silently get their microcode patched in the background and halt the damage.
I can't help but imagine that a sizable number of users will end up with rapidly-aged CPUs with a couple years worth of voltage margin left. They'll hang onto them for a while, probably planning to upgrade when 16th gen or Zen 6 come out, only to have them die, and they'll be none the wiser. Even if Intel can fix this, we're just beginning to see the extent of the fallout.
I can't see any way that any OS can overwrite the very microcode it is already running. The UEFI BIOS can definitely do so, hence the need for this to be incorporated in a BIOS update.
@@ColinDyckes I wasn't speculating; I was making a statement of fact.
See: Intel Microcode Update Guidance
"Early OS Microcode Update
The operating system (OS) should check if it has a more recent microcode update (higher update revision) than the version applied by the BIOS. If it does, the OS should load that microcode update shortly after BIOS hands off control to the OS. This microcode update load point is called the early OS microcode update. This can be done by an early boot software layer such as a Unified Extensible Firmware Interface (UEFI) driver, a bootloader, the operating system, or any other software layer provided by the operating system (like early startup scripts). The early OS microcode update should be done on each core as early as possible in the OS boot sequence, before any CPUID feature flags or other enumeration values are cached by the OS software. It is also required to be loaded before any user-space applications or virtual machines are launched. This is necessary so that the update is loaded before the affected microcode is ever used, enabling any relevant mitigations for potential vulnerabilities as early as possible.
If the BIOS does not load the most recent microcode update, Intel recommends loading that update during the early OS microcode update."
@@Zigglegarf That just seems like a recipe for disaster. I can imagine that system getting hacked by nefarious agents who will write code that can then be injected directly into the CPU without the user being aware. This also counts towards Microsoft itself injecting code without the knowledge or consent of the user as an "update".
@@ColinDyckes Not necessarily, it'd need a second place to stash the update and a mechanism to switch to it on the fly, but quite doable. Just not trivial to implement, but then even a simple modern cpu is far, far more complex.
@@thephantomchannel5368 Microcode is a bit different than most software and frankly if you can inject malicious microcode during os boot you probably don't need to.
The biggest issue I see is CPUs failing after the warranty has expired through no fault of the user. Another issue is, say Intel increases the warranty to 6 years to cover this, will Intel be able to RMA these CPUs with something that will go in the same socket? Presumably this generation will no longer be manufactured or have any remaining stock. That could be a serious issue.
Should I RMA now or wait till August Microcode Update? I am currently on latest bios 1402 with my Dark Hero motherboard but my 14900K is still unstable.
Running 13600K all core 5.7 P 4.5 E 5.0 Cache 1.38v with no issues. I would say RMA now because those cores which were boosting to 6 ghz at 1.5v on your cpu are now permanently damaged. You should not go over 1.4v on any cores. Just lock everything on the same frequency at around 1.4v or below when you get a new one or it will happen again.
@@club4ghz Thanks for your input man. Since I am paranoid now. Could you give me the exact bios settings I should use on a Dark Hero 1402 bios once I get a replacement chip? I should lock all p cores to 5.7 ghz you say. Should I use Auto svid behavior with auto cpu load line? I hear Intel failsafe svid is dangerous.
@@cemsengul16 It's hard to say what exact settings you need to use without studying your board and CPU first. I'm on ASRock board and after tweaking everything on mine I have 0.92v Idle 1.38v gaming. 1.26v stress tests (downclocking itself) 90c TJ max 200W PL1 PL2 locked. (Default is 180W on 13600K)
C-States - Disabled
Intel Speed Shift - Disabled = (This gets my Idle voltage down to 0.92v)
Voltage - Auto, FIVR - Adaptive +205 Offsets P-E-Cache. It's so much offset added because my CPU stock voltage at 5.1 ghz is only 1.2v.
I'm on LLC4 to have voltage drop together with clocks under stress. You can experiment with load line to get voltage drop in stress tests, but you need to be careful with that as setting wrong load line can overshoot voltage and damage your CPU under stress tests if it's set too high. LLC may be completely different on your board than on mine. Always monitor your voltage when tweaking your CPU and running tests.
I'd try to RMA if you're already experiencing instability, at least you'd get a brand new chip that will not damage itself until fix.
I'd RMA anything when it doesn't work as advertised.
anybody know if intel is accepting sanded cpus back for rma?
I don't know if this effects 12th gen, but anecdotally, I have a 12th gen system (laptop i7 12700h) and it has to be one of the most unstable systems I have ever used, frequent blue screens, random crashes, constant popping crackling audio (very widespread problem with 12th gen mobile)
I think i'm going to throw this thing at a wall and buy an AMD based laptop.
Can you please post a guide on how to diagnose whether CPU is degraded?
Im so glad I was impatient and got the 12900k lol
@@kamilhorvat8290I have 12900k and 13900k both running fine can’t wait for 15th gen unless amd fix 3d chip with dual ccd
@@kamilhorvat8290 12th gen isn't affected, issues would have appeared by now.
@dekaker9933 I wouldn't be glad if there was no upgrade path on the platform I bought.
Iam about to upgrade my 9900 to 13900 ! do you suggest to switch or its really a big issue topic ? thx to argu
The only way to make this right is to extend the warranty to a minimum 5 years but honestly my 8700k is still running great and I bought it in 2017.
How is this relevant?
Intel CPUs have always lasted a long time...I have several sandy bridge 2000 series systems that still run like they did when they were new...
I know this issue affects 13th gen and 14th gen Intel chips, particularly the 13900/14900 parts, but since Raptor Lake (and desktop Meteor Lake) is really just an evolution of Alder Lake, is there a ramification for 12th gen also (yes, Alder Lake runs at lower power levels and frequencies)? My 12900 system appears to be working fine, but perhaps it is also suffering accelerated aging and I wouldn't know about it? Thankfully I also have a 7950X system.
Extended warranty for at least 5 years otherwise will never trust intel ever again
If I'm not having problems with my 13900K should i upgrade the Bios when this new microcode comes out? Im undervolting my CPU to -0.75v, my PL1/PL2 are set to 253W and ICCMAX is 400A. The PC was built like 2 months ago and till now no problems.
Intel it's far far far too late, play stupid games win stupid prizes
They 100% already knew this from the 1300 series launch...
All those RMA from 1300 series they surely knew years before.
I have a 13900HX and have had zero stability issues or signs of failure, is this something i should still worry about, or did i perhaps get a sample from an unaffected plant?
It's rather strange how stock analysts are silent about this issue and its potential impact on Intel financially.
It's because it affects such a small proportion of their CPUs. According to The Register article this morning, Desktop CPUs are about 20% of the x86 market, of which about 10% are K skus, of which only a certain proportion are affected. So you're looking at less than 2% of all their processors. Even if they issue full refunds on all those K skus, the financial impact will be a blip. It will only start being a problem if it affects their laptop CPUs or if the OEM suppliers are affected (Dell, Lenovo, US Government etc)
(from an engineering perspective of course it's a disaster, these should have been caught in Release Testing or V&V - their biggest disaster since FDIV. But financially, it's just a blip, people forget how tiny the enthusiast market is in the grand scheme of things)
@@peterwstacey all different versions of 13th and 14th gen are affected, even T variants, not just the K ones, just to a lesser degree as they're not pushed as hard. Because it's not just the K cpus OEMs will be affected if a noticeable amount fail before the usual 3 year cycle in businesses. It just showed up on busy servers first as they're pushed hard and run 24/7.
@@Bareego 13th and 14th gen laptop CPU's are also affected (to a lesser degree just like T series) because both laptop and desktop share the same architecture and since server market was the one who was first to report issues i don't think this is just a blip considering companies like citadel put pressure onto intel to fix this or else
and SI's are not a small thing because many people choose a prebuilt over a custom PC which is main source of intel's income
@@Bareego Let's see. Intel are only acknowledging it on the K skus for now. If they acknowledge it on the rest of the product stack, as the internet rumours say, then expect the analysts to start to make noises 👍
Edit - also note that those OEMs and huge contractors like the US Government do a X-year lifecycle anyway, so if the non-K CPUs are affected but their median lifespan drops from say 20 years to 10, they won't care. So it depends
Why is nobody talking about gaming laptops?
The cpu is soldered on the motherboard, if there is a degraded cpu it is a 2500+€ damage. How will intel deal with laptop owners?
Do you think you need the new microcode if you manually locked your vcore and clocks in the bios?
I think so. But I just locked the voltage in the BIOS and now I have 1.4v peak in idle and 1.2v under load at 200w max whereas before it was going up to 1.5v at idle. I'm waiting for microcode and some watercooling shit for slightly increase watt and see what's goin' on. it's on a 13900kf
Probably a good idea. Stuff in the firmware has to deny the request for excessive voltage, the microcode prevents the request in the first place.
How do you lock the VCore? I built a PC like 2 months ago with a 13900K CPU, im undervolting my CPU to -0.75v, my PL1/PL2 are set to 253W and ICCMAX is 400A. Till now not a problem but im curious too, if i have this settings should i update Bios with the new microcode when it comes out?
@@AexoeroV depends on your mobo. but you can choose different mode of voltage regulation. what's your mobo?
@@MrSethGr hey, its an Asus Z790 Prime A WIFI. I have now the Actual VRM Core Voltage set to Offset Mode and i undervolted to -0.75 because that the max i can go on this CPU. Instead of Offset Mode i should set it to Manual and then input manually a voltage under CPU Core Voltage Override? My pc is not old, its now about two months since I built it and runs fine but i dont want my CPU to get bad. I do have some good Cinebench R23 scores with the CPU undervolted, i can even get 40K if nothing much runs in the background!
Regardless of the problems now, i enjoyed this good reporting/education thanks! What i like about all this is I think we've all be thinking for a while that intel and AMD have started shaving the margins a bit closer. i'd like to get back to a proper engineering buffer. I dont' need every last drop of performance. That should be opt in. I want stability. Not xeon stability (4 cores and 6x the cost) but something closer to what we had 25 years ago with consumer where the boars and cpus came with sane defaults and had most of the performance and they generally never failed for 10 years opr more.
It just says 13th and 14th Gen Intel everywhere online...
But the stories started with 13900k and 14900k...
Are all 13th and 14th Gen CPUs affected?
Roman is also mention "all CPUs above 65 Watt" which confuses me even more...
The 900s notice first
Not enough info to say for now, my 13600k has been stable for a year now, through stress tests and all.
It seems to affect i9s first due to their higher power intake however it seems to be affecting low tier CPUs just at a slower rate. With reports of i7s and some i5 apparently showing signs of degregation
something ive been thinking about for a while. if you put a nitrogen cooled pc in a vacuum would it remove condensation???
I know microcode updates are usually included in a bios update, but can microcode updates be done through a windows update?
As I understand, Windows and Linux can "update" microcode before fully starting the OS. The update is not permanent. The update is loaded before the OS startup. So as long you do not stay in the BIOS (which will not have the microcode update), you should be fine.
@@benjarobin
Thank you for answering that question. That was my concern as well.
sort of yes, both windows and linux can load an updated ucode into the cpu during the boot process, just later than the bios would
Thanks for the balanced and perhaps less strident report than some other channels recently. The fairest way forward I think would be for Intel to offer an extended warranty on the parts potentially impacted by accelerated ageing.
Now imagine someone buying used Intel system in a year or two that may be damaged and fails in a year.
Intel Oxidation Lake coming soon to eBay...
With all respect to the work you are doing, the microcode update will not solve the problem.
First of all, it just postpones the CPU degradation to the date after warranty will expire - the same damage controlling tactic that nV did to the laptop GPUs ~10 years ago. They don't want to earn their responsibility and lose too much profit.
Second, if the microcode update is slowing down these CPUs, then it is not the product that you wanted to buy, it is not the advertised one and the benchmarked one.
I believe the operating systems also contain and upgrade the microcode of the CPUs during boot time.
I also heard something like that. But don't really know how it works
Are the 13th gen Intel laptop cpu's also affected?
I need a recall and full refund…. Class action lawsuits are USELESS. All media need to press Intel and give them bad PR. I want my board refunded too…. Spent too much on a z790 apex encore just have 3 CPUs swapped core i9 14900k
That ain't happening. RMA the CPU again and then wait for the microcode update.
You know, you can always opt out of class actions, and take your claims directly to your small claims court if you really feel that raw about it.
Yeah but ASUS boards are tricked out garbage as well with weird bugs, instability and most probably in typical ASUS ROG fashion, the some parts of the CPU are silently getting hammered with high voltages to maintain stability.
I am with you bought a 14900k and a Z690 Maximus Hero EVA edition, but I also what half my HDD refunded as I have 4 western digital black 6TB in raid 5, amd doesn't support raid 5 so I will go from 18TB in storage to 12TB
keep dreaming, Intel would only be willing to refund the cpu or replace it but the board that's a different product that has nothing to due with intel they would not refund a board
I’ve go a question. When a cpu is being made aren’t there a whole bunch of “machines” (fabs?) that are making all the dies all the cores in a factory? Is the problem that Intel is facing all of those fabs or just one that was messing up when it was making a cpu? I don’t know all the terminology or processes of cpu production, just trying to learn.
And now AMD is delaying Ryzen 9000 launch for "quality issues" as well. WHAT IS GOING ON.
Maybe the are checking the 9000 chips dont do this.
Delaying by two weeks. Sounds like a microcode patch. What you don't fix in two weeks is a manufacturing error.
who has bigger CB result!
most likely using the extra time to bin more and getting the sku's a little higher, as this will be a delay Intel cross the broad, remember this problem could be in server die's also and if so the enterprise customers we be put ahead of DIY market.
Now the fastest way for Intel to get usable replacement cpu out there, is to get TSMC to make intel "12gen+" CPU
Intel of cause will call them 15 gen Re-Fresh-Lake CPU on TSMC 10nm or 7nm and Intel will also need to make a MB bios/firmware update and later on they will say that 1700 socket was 4 gen socket
Waiting for Intel's nerf patch mid August
I would imagine that intel figured that the CPU's would last just long enough to get them out of a warranty claim. That has obviously majorly backfired on them.
time to sell our intel pc and get a hen instead🤭
Intel can definitely address the challenge of informing Windows users of having a potentially defective 13th or 14th gen CPU via Windows Update. Intel can work with Microsoft's Windows Update team to detect systems with such CPUs and an update can be applied where a pop-up appears on the desktop screen after every reboot/logon displaying a warning message about the installed CPU with information about the need to install the latest BIOS for their motherboard and how to get in touch with Intel to RMA their CPU. The pop-up should only be disabled once the user clicks to select a check-box stating that they acknowledge the issue. Microsoft have 100% done this before with regards to cetain Toshiba laptops that shipped with faulty OEM Dynabook AC power adapters, so there is no excuse that it's "too hard" or "can't be done". Great video by the way!
Whole issue should be reviewed from original timeline:
1. Intel 13th Gen release;
2. Intel spotted problem with Via Oxidation ("addressed back in 2023" as they said), but hide it;
3. Problem with limits appear and Intel blame motherboard vendors for it;
4. Intel fixed problem with limits by new BIOS and profiles;
5. Problem with CPU appear again, which isn't voltage (Via oxidation which Intel hide).
So, half released CPUs might be fixed with microcode since they don't have via oxidation problem, but first batches can't. They should be RMA. Intel just need to share manufactured dates\serial range of this CPUs.
Intel has never said that oxidation was problem for all chips that were made in that timeline. They said there was a problem in 2023 and we do not know how many chips with the problem made it out of fabs and we don't even know how long that problem lasted or how many fabs were affected. What you wrote is pure speculation.
Yep they should probably offer to replace all oxidation affected cpu's.
Is there a difference between the 12th and 13/14th gen contact plates? The 12th gens are out of stock everywhere! Want one for my sons 12600k
I purchased a 12 13, 14 gen Thermal Grizzly contact frame and received some plastic 3d printed thing.
Its says Thermal Grizzly on it.??
Is this really Thermal Grizzly product?
How does it not deform under heat?
Every other media: disaster! its end of intel
Chad Roman : its fascinating what happened on atom level xD
😂😂
Im sure intel did this to make the intel fans buy CPU more often so they always have the latest and best performance. Like they gave them 2% performance each year over 15 years. Intel is such a kind company :)
The gigachad move for AMD now would be to run a big Advertising PSA campaign about the faults of Intel processors and how you should RMA them
nah there has been enough juvenile behaviour from Intel and AMD the chad move is to just sell your product because everyone knows it is better atm
Perhaps degradation with time is purposeful. As chips become more powerful manufacturers will be incorporating planned obsolescence. Todays i7 or i5 can function a very long time in a simple task home machine. Most desktops do not play AAA games. I still use a 4790k from 10 years ago. It is perfectly fine for normal non-gaming use and will remain so for another good several years. This presents a big profitability problem for CPU manufacturers. I see Intel so hesitant introducing a six core i3. It would be a 100 dolar chip with potential to function in a simple task computer for 15 years or more. The industry is maturing and planned obsolescence for CPU's is the likely course.
Not all of them must be affected. If you have manual OC and not using these Intel OC features like eTVB, your CPU is likely to be fine.
Only if you limit voltage and power draw manually.
Had to RMA my 14900kf a few weeks back. Bought in in January and by April it was super unstable and pretty much unusable. Ended up figuring out if I disabled P-Core 4 that made things run much more stable. Got CRC errors when unzipping NVIDIA drivers, but when that core was off, worked fine. RMA process was fairly painless and quick. New CPU seems to be running much better and I already have the microcode BIOS update in place, so hopefully it'll last more than a few months. Intel better get this figured out.
11:20
So if it's a laptop CPU, you're screwed?
I've got 3 friends who recently bought pre-builts with intel CPUs (13900k, 14700k & 14900k) following advice a from a mutual friend who ''knows tech'' - one of them I've managed to give a warning too and they've already dropped the voltage as recommended by others, but the other 2 just don't care ''it turns on fine, I don't understand what the issue is, I'll worry about it when it goes wrong' and don't understand about microcode or BIOS settings - going to be fascinating watching their machines over next few months.
When the first person contacted the company that built the PC they said they'd tested everything, weren't aware of any Intel problems and they'd need to speak to Intel directly if the CPU had stability issues. Interesting way to treat consumers when this issue is out in the public and has been for a bit.....
I have a 14900K and I'm kinda freaked out. If this is a manufacturing flaw, yeah. This could be a massive problem for Intel. It means they sold a defective product. This could be a very expensive mess for them. I'm tempted to just slap my 12900K back in my rig and RMA my 14900K out of an abundance of caution. I haven't had serious issues other than game crashes when the power limits issues popped up.
Yikes. Yeah, better get the RMA process started before the massive flood of RMAs begin. :(
Do it.
If you havent had issues, dont do a damn thing. Again its not ALL cpus, its a specific set of CPUS more then likely from a few of the fabs. If you start having issues yes, you will have to RMA it. But again, thats IF you have issues.
Why do an RMA if you're not having an issue?
@@64BBernard it sounded like he was having issues. Game crashes when power limit issues pop up. Personally, I wouldn’t take my chances. If I can get a replacement, I would. It’s a lottery of potential future issues otherwise.
I really think it needs to be a recall at this point. I'm also concerned about BGA parts - Intel says laptops aren't affected, but that would also a much more expensive problem and I'm not sure we should really believe them.
I have a 2022 13900k Intel cpu. I got approval for warranty replacement yesterday but I had to ask if it'll still be covered since it is delidded(largely because of stability issues despite already having a custom water cooling system). No answer from intel yet.
So now my CPU may have issues with oxidation and definitely has stability issues since i can't overclock without issues anymore(could sustain 5.8ghz all core and 6.1ghz 2 cores daily) now I've had to drop ram to 6000mt/s and stock cpu clocks. Can't run 6400mt/s or even a 100MHz OC on any p or e core.
Still have some stability issues that aren't explained by anything else. I knew delidding had risks but i didn't know intel would do this either. Guess im waiting for 12p bartlett lake or switching to AMD (if they are issue free in 9000). At least it is pretty stable for work, just cant use my water-cooled 4090 for games 30% of the time.
AMD just delayed 9000 series launch to 15th Aug because of unspecified quality issues.
Intel will not accept RMAs for delidded CPUs.
It's common knowledge that delidding a CPU voids the warranty. You could try gluing the IHS back on the CPU and then hope that they won't notice.
@@WrexBF Yea it's well known. I'm actually pushing them on it since delidding hasn't affected anything and it's only their own failures causing this. Normally you don't have to worry because CPUs typically just don't fail. I got a weird response and waiting on another now.
@@gagarin777 I saw that, so kind of hoping reviewers don't get golden samples since they supposedly are dumping the first batch. They probably knew it'd cause a huge backlash as this Intel issue is ongoing.
Yes, in this case user's knowledge does matter too. If you always checking your cpu voltage or familiar with OC/BIOS tuning troubleshooting stuff since day 1, you should already lower voltage if too high or something weird happens. The cpu who get aging, damaged etc may mostly happens to those people who completely never checked their cpu voltage etc.
i9- 12900k is now a good buy for those of us who have rigs we can't be bothered to rebuild!
If you're on an lga 1700 system with 13th or 13refresh cpu that's acting odd it's probably the cheapest fix with good performance. I've heard (NOT VERIFIED) that 12900 k's were having issues a smidge above normal levels, but that could just be the node or arch or both.
Thanks for the detailed breakdown. I have a question. If I have a 14900k that is unstable and affected but it was lapped and polished in a last ditch effort because I was having very big temp diff between cores. Can I rma that ? I never took the lid off. I still have original box and receipt from microcenter. It’s 4 or 5 months old at most.
Unlikely but it really depends on how Intel wants to play this.
You can try, but they'll possibly deny the warranty because it's lapped... This whole situation sucks for 13/14th gen owners, especially those going to more extreme ends on their systems. I have a delidded 7950x running bare die on a WB with LM, I know that any warranty claim would have been void at the time I delidded it, but with a manufacturing defect, they should still honor it. So my advice would be to try, but don't have high hopes they'll approve the RMA. If you have screenshots, etc from the instability and high temps before lapping, then that may help to convince them, wish you the best of luck.
intel is so f..up
Rather than electrical conductivity problem the main issue with oxidation is the mechanical stress exerted inside the silicon packaging due to bulging of the affected via.
Does the latest news about potential leaks from Intel indicating that the ring bus is getting cooked by the voltages further aggravate the issue. Will the microde correct for this?
Is this for desktop cpus only or for laptops as well.
Or maybe Intel just realized, that building CPUs that last for 10+ years is not the best business strategy anymore. They just overshot a bit so these CPUs now already fail withing the warranty period.
Uhh, does this mean windows do not distribute CPU microcode updates? Because in linux I have microcode updates as part of regular system updates, so I would expect new microcode to be automagically distributed and updated without need to explicitly ask users to do something on their own.
Waiting for TechYesCity respond to the major unveiling of the issue.
As he had an other theory going for months now...
hello is it possible to know if the Intel Xeon W-3400 & W-2400 processors are reliable or not there are no tests of its cpus
Nice, interesting and technical discussion about vias and the entire concept. Also, it will be interesting to see second hand marketing prices after this situation, if Intel doesn't recall or RMA their 13/14th gen CPUs.
The more that comes out about it, and the more that ppl talk about it, the more I think this is not only a major hit to intel in general, but also to their efforts to get 3rd parties to use their foundries...
Windows & Linux usually have updates that apply an updated microcode even if bios are not updated etc
Thank you for doing a well researched, balanced and non-biased video on this subject! Don’t make reference to GN’s videos. Steve’s running a tabloid UA-cam channel. Does not compare well to what you are doing. Thank you once again!
I literally just ordered a 13700k and platform a few days ago...I haven't recieved it yet, but i'm wondering if a use a static voltage if it will mitigate the microcode voltage issue or does that not matter? I could probably return it, but I really wanted an Intel system...
just run a contact frame and make sure your cooler is properly configured right like tighten up all the screws.
Do you have a 14900K as a light source for this video? Everything seems like it's glowing a bit too much beyond the normal colors 🤔