My thoughts on the Intel 13900k and 14900k crashes
Вставка
- Опубліковано 12 вер 2024
- Clip from the weekly Tech Thursdays show on 7/18 ua-cam.com/users/li...
Check out the livestream every Thursdays.
Workstation HEDT PC:
CPU: AMD Ryzen™ Threadripper™ 7960X 24-Core, 48-Thread - amzn.to/3N2UDOL | newegg.io/ncfa...
Cooler: Noctua NH-U14S TR5-SP6 - amzn.to/3TMubgD
RAM: G.SKILL Zeta R5 NEO Series 128gb 6400mhz - amzn.to/410rN7G | newegg.io/ncc5...
Motherboard: GIGABYTE TRX50 AERO D sTR5 - amzn.to/41Go9A8 | newegg.io/nc87...
Case: LIAN LI LANCOOL III - amzn.to/3S0GQLL
GPU1: GeForce RTX 4090 Founders Edition - amzn.to/3RzR1We
GPU2: Radeon RX 7900 XTX - amzn.to/3vji4xw
Streaming PC:
CPU: AMD Ryzen 9 7950X3D - amzn.to/45nHlmJ | newegg.io/ncd8...
Cooler: Noctua NH-D15 - amzn.to/3MXEjzh | newegg.io/nc47...
Motherboard: Gigabyte X670E AORUS Master - amzn.to/45nqqku | newegg.io/nc83...
RAM: G.Skill Trident Z5 96gb 6400mhz 2x48gb - amzn.to/45rpifu | newegg.io/nc45...
GPU: Sapphire Nitro Radeon RX 7900 XTX -amzn.to/45ptWut | newegg.io/nc34...
PSU: MSI MEG Ai1300P PCIE5 PSU - amzn.to/43r6Wdx | newegg.io/nc59...
Case: Fractal Torrent RGB - amzn.to/45rlGex | newegg.io/nc14...
Webcam: Logitech C920 - amzn.to/3viZEgg
#13900k #14900k #intel
Almost, you are correct that issue is right there in the middle, the ring is getting fried, but the root cause of it is the voltage, the P core, E cores and ring get their voltage from the same rail , so when a P core boosts to 6ghz and requests 1.55v all three (P/E cores/ ring) are getting those 1.55v , the cores can take it but the ring can't so it starts degrading until eventually is completely fried, that is why CPUs show different symptoms (errors on P cores, or e cores, or memory) since everything goes thru the ring, most of the CPUs are gonna be affected but the severity will vary since chips with better silicone quality will take more time to degrade since they don't get so high voltages, there should be some golden CPUs that won't be affected at all if is so good that it never requests more than 1.45v but most 13900k/14900k do use 1.45v+, for those is only matter of time , as for the microcode patch they are just gonna lower the voltage request limit so the CPUs that haven't fully degrades can last till the warranty expires. But the ones already degraded they are trash, only option for those is RMA
was about to say that as well, but he did say quite early that the issue was voltage and therefore current related. Increased current will make the electromigration occur.
even with baseline profile suboptimal chips request 1.5v. to chase profit margins (selling i7 quality as i9) there's a TON of suboptimal chips out there. twp i9 sku's too. you have to wonder will those chips requesting 1.5 and higher boost near advertised speed after the update? if they do will it be stable with too little voltage?
@@JoeL-xk6bo if they are degraded they will not be stable at the advertised boost frequencies. U need to overvolt them even more to get that, until they die.
The clockspeed will be reduced, by how much that will time and intel tell.
@ivantsipr Exactly, there are a lot of good videos out there that explain it well
Yup that is the main issue. To high voltage on the Vcore. The ringbus is fine for the high power delivery. It's not fine with the high voltage it gets. Buildzoid showed today that on a Gigabyte motherboard on default it can reach 1.6volt on single core bench in R23. That is the main culprit!
I’m confident by now Intel understands what the problem is. But they don’t want to admit it for liability and reputation reasons.
They knew many months ago what the issue was.
The oxidation issue is not fable, it will come to clear light when Intel has to spew it all out in their class action lawsuit.
The ring shares voltage with the P & E cores,, it doesn't come from VCCSA. So if you're pumping 1.5v for your P cores to boost to 6ghz you're also feeding that much voltage to your ring and e-cores.
Get to the point, and don’t repeat what’s NOT the problem over and over.
I used 14900KF for a month with daily crashes (one to two a day on Win11 with huper-v running web dev machine). Once I disabled the turbo boost entirely, it never crashed. That seems better than to turn off all e-cores?
How to turn it off? Can you teach it?
@@syedaminjazmisysulaiman6286 You have to enter your BIOS and generally you are searching a single option "Intel Turbo Boost" which had to be turned off/disabled. Search how to enter your BIOS specifically for your motherboard model.
Yesterday I updated the MSI Z-790 PRO WIFI BIOS to see if that will prevent the crashed leaving the original settings. It seems stable for now.
I only ran at stock profile when I used my 13900kf when it was on an b660 itx board, but even then after observing the high voltages I lowered the voltages to the cores quite fast.
none of the users with custom settings have issues with their systems because they have not allowed the cpu to run at 1.5v+. usually the cpus will do very high frequencies at 1.35 or if u have a really bad one at 1.45 if u want them to boost to 6ghz+.
high voltages on its own is not bad, it is only bad when it is during load when the current(ampere) increases.
I have two i5 - 13600kf's. One is faulty (no POST) and the other one is brand new. I used the brand new CPU to disable all the e-cores since the old one was completely dead. I then put the faulty one back. Guess what? The PC booted. Your assumptions are probably correct.
Yeah, this is just wrong. I'm with @ivansipr - I've seen a PC in an idle state suddenly BSOD. Because some background task started up a single core that boosted into the stratosphere.
And I've now seen multiple videos of just a single core, or sometimes two, getting a boost to 5.8 or even 6.0GHz. And to reach that freq. the core gets fed 1.55 or even 1.6v. Which will degrade the core, the cache, the ring, even the IO.
But when ppl Synch all of their cores and stop that stupid single core boosting behavior, and limit the power and voltage, their CPU has no issues.
Why is there crazy high single core boost? So they get a top score on Cinebench from some reviewers. What a useless reason to kill your CPU.
Lock the boosting to all cores, lock down your power draw and your CPU should last forever. More or less.
Yep :)
The issue seems to be the fact that they doubled the ammount of E-Cores for the Raptor Lake Family of CPU's, and the interconnect (THE RING) isn't ready to handle all the extra E-cores, with each increase in clockspeed you increase the ring aswell, for example 6ghz core speed comes usually with at least 5,7 ghz of Uncore (RING) clocks, and it requires significantly more Vcore to be stable.
My theory is that the RING can actually handle all the Ecores, but can't do it while having 5,7-6,0 ghz P-Core clocks (and with that high Vcore). The short answer here is that Intel didn't validated enough the Raptor Lake Architechture, 10 nm node isn't competitive with Zen 4 as is and they made changes in a hurry to remain relevant.
Intel will reduce the VID voltage request in the next BIOS update. This essentially means they're prolonging the life of your CPU just enough to surpass the warranty period, helping them avoid issuing refunds.
Exactly. But how did they miss this in the new 14th gen after 13th gen issue??
Curious to see how that will impact performance though
@@clarenceoveur9497
Buildzoid from "Actual Hardcore Overclocking" makes a new video testing performance drop when you limit voltage.
@@clarenceoveur9497
Buildzoid from "Actual Hardcore Overclocking" recently made a new video testing performance drop when you limit the voltage.
cRaptor lake 💩
Here's the thing that bothers me about this whole Intel CPU fiasco..
The UA-cam hardware testers are now claiming that the problem is bigger and may include 'all' 13 and 14 gen Intel cpu's but where's the tests to demonstrate the claims? How easy would it be to test the claims if its all 13 and 14 gen cpu's? These UA-cam hardware testers are comparing and testing everything except this, the easiest test out there. Just grab a handful of 13 and 14 gem Intel cpu's and test them, we know that installing NVIDIA drivers will trigger the issue so why are they not testing?
UA-cam gamer sites accept motherboards defaults to unlimited power and are sponsored by liquid cooler vendors to push product as a cure. Setting reasonable power profiles for long/short duration and limiting multipliers usually are not discussed. "Why buy and I9 and run with reduced power when you can buy an i7 and drive it to its max.." I have found very little information as to how processors scale. Pushing an extra 100 watts to get the last 15% seems foolish to me.
Intel did admit it's everything
Processor degradation may be related not only to the level of power supply voltage, but also to the operation of the gate that controls this level.
The gate, which allows power to the processor and controls the level of this voltage, may experience degradation when operating at high frequencies. This may be related to the fact that at high frequencies, the gate may experience a greater load, which in turn can lead to its wear and degradation.
In addition, gate degradation can lead to changes in the dynamics of the switch on/off, which can significantly affect the operation of the processor.
Thus, it is possible that processor degradation may be related to the operation of the gate that controls the voltage level, especially when operating at high frequencies.
3-month old 13700K here. No symptoms of failure--yet--but on heavy CPU load (benchmarking) HWINFO 64 always shows ring temperatures on red (thermal throttle), even if the individual core readings don't indicate throttle. Could it be related to what's explained here?
What kind of cooling solution are you using? Do you use your system in an AC environment?
@@sharthakghosh970 NZXT Kraken 280 AIO. No air conditioning, temperatures are quite cool in my location at this time of the year.
i guess intel cpus are just that hot either that or you forgot to remove the sticker on your aio (and multiple more probable factors but i'm too lazy)
@@j_shelby_damnwird is that temperature rise you are observing now or you are observing because this whole issue panned out recently? I think best would be to apply the microcode update as soon as possible and not put load on that cpu for some time.
@@sharthakghosh970 I'm keeping close tabs on all readings since this issue started making headlines. I've installed the most recent BIOS with the microcode, yes. Although after watching this and learning about the oxidation issue I highly doubt that's gonna do much good. You're right, I'd better ease it with the stress tests and benchmarking for now.
Wish I knew more about the Arizona plant that had the oxidation issue
It was the HVAC system that had rust
@@AJ-iu6nw "Oxidation" refers to the process where the copper layers on the CPU begin to rust during the deposition process. This rusting occurs as layers are being added.
@@xponen but it was the hvac though right
This all will come to light when the class action lawsuit will force them to make company papers public.
And this is why they reduce clock speeds on Arrow Lake. This was discovered by all users on the 12900K. Think about every problem the cpu had from heat, stutter, crashing, memory, etc. Sure intel and microsoft released patches to work around and mitigate...but what solved EVERY PROBLEM? Turning e cores off. That's why you can also get perfect stability by underclocking cores, or underclocking memory. There is too much shit going on in these chips especially to be pushed so hard in frequency.
Clarification needed. Is it only the 13900k and 14900k or every 13th and 14th gen including laptop cpus have this issue.
every raptor lake cpu with base power 65w or higher could be affected
*not all 13th and 14th gen are raptor lake, some of them is a refresh from 12th gen
*since it's specifically mentioning power, it means that doesn't matter if it's laptop or desktop, if it draws more than 65w then it's at risk of instability
@kuekejuu5057 got it thanks for the clarification.
"no reason to do a recall" hot damn, you are so sure about?
A class action lawsuit can make stuff go side way FAST!
The problem with this theory is that the 13th/14th gen processors are degrading even in servers where the heat not exceed around 80C according to Wendell. Or maybe we don't get proper heat measurements from there. Maybe it's not only the temperature but voltage/temperature/contact frame/bad motherboard/bad microcode these all together causing the problem.
Note:
I have an 13900K with custom water-loop and contact frame and with 3x420mm radiator more than a year. ZERO issues from my side.
easiest way to describe it is that the ring is been asked to supply too much current to every core and cache and its burning the ring out technicaly your power input is burning out .
lets say each ecore can draw 80 watts no problem and each p core can handle 150 watts, now 16x 80 watts then 8x 150 watts thats 2400watts that the cores would love to eat away at but the ring cannot supply it , the ring is supplying up to 580watts thats the highest power draw i saw on my 13900k cpu which is burning the ring power supply out and shoving voltage through other parts that are pushing them too far over specs. ever core in the past can handle upto 150watts pl1 was 125 and pl2 was 150 and its never been a problem till we added the large amount of ecores , 12th gen only had 8 ecores which means it would take a max of around 450 watts thats what my 13900k does if i disable 8 of the ecores its max draw goes down to about 470watts . thats why the 13th and 4th generation are the ones having the problems the rtx 4090 had failures and they have now limited the power the cards can draw down to a respectable 500watts so the silicon is able to take a max long term of 450w to 500 watts for the die size they use if we want more watts we have to increase die size so the wires are thicker and can carry more current for the trasnistors i would have made the cpu use different rings for each set of cores and cache as with the current design what ever needs the highest voltage wins which means all the other parts of the processor on the same ring loose efficiency if they dont need as much voltage . if my ring voltage was not attached to my core voltage i could run that cooler and probably run it at 51 all day at the moment cause of heat i run it at 48
So what the sollution? My com keep crashing while playing stable fps
I RMAd my 13900k after I found the last group of e cores was causing mine to fail. I ran per core stress test on OCCT and found failures in that group. In order to stabilize my cpu, I had to downclock that group to 3.9 ghz
Thanks for sharing your thoughts. I built my first PC a little over a month ago and installed a 14900K. So far I’ve been okay and not encountered the known issues (thankfully). I’ve updated the BIOS whenever an update has been released and will be awaiting the Intel micro code update.
I only use the PC to game and don’t run the benchmark tests and will stay away from those even more so given the current issues.
Hoping I’ll be okay … let’s see.
Any way keep up the good work.
Will Arrow Lake have similar problems to Meteor Lake? ARL has 2 more P-cores than MTL. Given Intel’s ring bus problems, I’m not optimistic.
Would you recommend turning off the e-cores on a 12th gen "12900k" ? or do you think the limited number of e-cores on the 12900k is safe?
12900k runs fine as is no worries atleast that is what I have been told its not having these issues.
How do you know if you have a bad 13900k mine is 1 year and a half old. Fairly stable a couple of crashes within that time nothing major. Temps are good but I do get random spikes to 87 whilst gaming occasionally. But sits at 65 to 75c normally.
Are you essentially saying that the network switch is overheating?
Like, if you imagine the cores as a Beowulf cluster, then the nodes themselves are fine, but the switch linking them is bad. Yes?
every part can fail.
Thank you for explaining this. Now I understand the basics of why the i9s are the ones effected.
I made the right decision going with AMD this gen.
But AMD desktop and mobile CPUs' fail rate isn't any better, even when not considering the explosion issue on Ryzen 7000. Lots of 4800H and 3600 are failing if you paid attention.
@@Alan_Skywalker guess I'm lucky then that my 7800X3D is still fine.
I’m pretty sure they knew about these issues from the get go, therefore they will launch at a certain point a new 13th / 14th gen processors with these problems fixed.
Hopefully. My 13th gen i7 just got a 1 year
Bartlett Lake-S is their fix. 12 p-cores 0 e-cores on one of the chips coming 2025.
The problem solution is EZ
Buy 7800x3d, upgrade possibilities later and never trust Intel again with their every year cpu launch
the problem is related to the idea of buy something from intel and expect a product in return, you should just say just buy it, it just works, do not look at benchmarks, give us money now so we dont sink to fab only
intel fanboys(the like 5 that are left) are imagining a future where nothing of this happened, and amd goes bankrupt, so intel can keep testing glue to glue together their future cpus, the fumes of that glue caused what happened i bet
the race to the bottom, but only intel is in it
if there wasnt this many gamers and companies affected, it would be fun, but is in reality, just sad.
don't think so, vccsa is locked on mobile and non K skus and they also crash apparently
I had one of my cores go bad with i9 13900KF (had to limit it to 5Ghz).
two trains use same rail, thats why its crashing. good job Intel
no one is speaking about specific cpu which caused this errors because i am using intel i9 13900k i have faced a lot of bsod before but not every day i have made temperature control limit on bios now it is not creating any issues , i agree what your telling but controlled over clocking issue but exteneded supported clocking limit will not cause any such issues , this issue is coming from few cpu not on all cpu's
From proven stability and reliability. to RNG and gacha to see if you are lucky or not, affecting also 65w and locked CPU with B boards.... this its just a dissaster
The main culprit was on die power controller.
I said this from the very start , the e-cores was the dumbest idea ever maybe for a laptop but on a desktop it’s just dumb
Typical UA-camr, bulks vieos with repetition and nonsense. You told the audience 4 times in a row that the cores are not the problem. It is completely unworthy of my time to continue watching your artificially inflated video at that point. If you have no content, reduce the length of your video, make succinct points, stop wasting everyones time.
He’s incapable
Yeah, that kinda rattled me too
Nice theory but completely off the mark.I have just returned two 14900KS's with issues. The first one i bought at launch. I ran it at 56 P cores, 44 Ecores, 46 ring. It was underclocked by 0.06 and fully stable under every benchmark that matters. I played games for around 4 months straight on it then one day BSOD. I have RMA'd it and got my new one. My new one already needs 0.02 on x57 on the V/F curve to be stable. You're claiming its an e-core issue. Its not, its a degradation problem with the p-cores. I would say the oxidation issue intel mentioned is more wide spread than they are making out. Running a KS at the settings i run it at will not cause any issues, yet iv had issues. Im a gamer, game ALOT. A few people i know have also has similar issues with there 14900k's and one of them games without his e-cores on the whole time. Claims it gives him more FPS (i told him it doesnt). Yet he had to RMA his also. Sorry but what you said just isnt true.
Nah dude. The compute fabric might be a slight issue compared to the real issue. I know the have not updated it from the 12th gen I9. However the real issue is the single and dual core boost algorithm! They don't die from power usage but form way to high VID and Vcore! The 14900K can boost up te 6ghz and the 14900KS even to 6.2. The Vcore can exceed 1.6volt wich is causing the 2 cores that are the best to get over current every single time it is boosted.
Just use HWINFO 64 or an other tool to see the Vcore. It is in the single and light load for example in Cinebench R23 that those cores get way to high voltage. I9 cores are the best binnes, however the VID table that can be read from the bios and is different for each CPU due to bin quality. It can vary between 1.398 and 1.503 volt and those were just a couple of hundred samples.
The issue is that anything over 1.4 is going to cause degradation over time. Anything over 1.5v even way faster. There are boards like Gigabyte and Asus that are know to exceed 1.6volt Vcore during single and light load.
Those server boards and those cpu's were dying only because of the high single core use for those games, not for sustained all core load as the all core load is way lower.
Actually hardcore overclocking(buildzoid) has a new video out today, but Framechasers is saying this already for months.
I have a 14700KF and from day 1 undervolted and also i just recently changed the additional turbo mode cpu voltage in undervolt so it does not get that high voltage anymore with boosting. Result i gained speed and lost heat and power usage.
I have heard it before tha compute fabric would be the issue but i don't think it is at all!
Honestly Intel should hire you immediately. Your knowledge seems to be so superior than 99% of UA-camrs who are making videos not knowing anything about this issue.
Actual Hardcore Overclocking was actually the 1st UA-camr to have a hypothesis about the ring bus getting fried, so I would give him props too for that assumption.
Well he's wrong, so maybe Intel will hire him.
The problem is the high voltage required for high p-core boost being applied to the ring and everything attached.
13th and 14th gen request too much voltage.
I9 have highest boost so highest voltage, highest failure rate.
Intel is hiring spindoctors now, not people who can diagnose and repair
@@arielis78buildzoid is daddy
@@incandescentwithragedoes it also effect the i5 13600k or am i good
But it’s not only thr icore 9. There people having issues with the icore 7 also. 13700k and 14700k
Lol, this video.
Isn't it high voltage that you can't even see in HW monitor. Think you need a oscilloscope to see the voltahe spikes
Higher uncore clocks the higher voltage goes the ringbus , take consider their are folks push ddr5 to 10000mt not just 8000-9000mt.. memory voltage goes to 1.6v . Where do you think will happen next? It's same problem with AMD exploding CPU due to 2:1 ratio for expo memory thus board vendors fixed it by adding 3:2:1 ratio for xmp higher memory clocks thus pushing fclk 2100 while soc voltage stay around 1.2 to 1.34v but never goes 1.4v that will cause damage to fclk due higher memory speeds gonna 8000mt . They can support it but not same performance as Intel does it with 2:1 ratio clocks while almost did but fabric clock is too weak on AMD CPUs to see any benefits to use 8000mt dram sticks.
My Asus ROG Strix z790 E Gaming with i9 13900K is stable with the 1801 Bios.. I Had to roll back after updating to 2402, it became unstable^^
Did you use stock settings or Intel's default for the new bios? I'm on the new bios same motherboard and CPU as you and haven't had any issues luckily.
I have a degraded 13700k your theory on disabling Ecores to fix the issues does not work the problem is there on mine. I disabled all Ecores and my system still hard crashes. this also with XMP off. The going theory is the ring bus is burning out.
Do you have the frequency and voltage locked or are you letting it do the stock Intel boosting behavior? I'd imagine you could tune it manually and get it stable. There's tons of resources/forums online that can walk you through it step by step. Even if you can RMA it, which you should if you can, I would still tune it manually when you get a new one. That's the point of K-series chips after all, but not only that, its the insane voltage that is used when boosting 1-2 cores that is ruining CPUs.
@@The_Chad_ Yes i locked the Pcores to 5ghz disabled the Ecores and manually set voltages and turned off XMP and it still hard crashes. At this point I have tried just about everything i can find. Im just going to rebuild my PC with the new AMD stuff when it comes out. I am a small business owner and i use my pc for work daily so RMA isnt going to cut it because i will have to buy something to replace it while its being RMA. Plus I have no confidence that a replacement Intel chip will be any better. At the end of the day Intel pushed these chips way to hard and over stated there performance.
What were your settings before it burned out? I have a 13700K that has been running 24/7 for 18 months.
@@BobMar1964 for a while I was using the intel xtu utility and the built in auto overclock wich put it upto 5.6 in 2 cores and 5.5 on the rest of the Pcores with the E cores at 4.3. I set the bios power level to the intel 253w power limit. That never did run stable so I locked all the Pcores to 5.5 and set the voltage to 1.35v. And that ran great for about 7 months then that started crashing. Then I set it to 5.3 all core at 1.32v and that didn't work. Then I tried 5ghz all core and still no good then I tried 5ghz with the all the Ecores turned off and XMP turned off and it still won't load a game with freezing and rebooting. I don't even get blue screen just locks up and shuts down.
I have been running my at 5.6 4.4 and ring at 4.8 voltage set at 1.285v and still score 40800 in cinebench using a 280 aio max 89o for 10 minutes . Stock cinebench was scoring around 39400. Looks like I found a sweet spot. 😊 till now not a single problem. Also avoid using the intel integrated network that thing make the cpu runs hot af.
I think there are two killers though? You might be avoiding one of them, but what if your cpu has a little bit of oxygen from the problematic hvac in the Arizona plant?
Limit ur wattage 253 watt or lower
@@TryHardNewsletter I might not have that oxygen issue I am based in the Nederland and I think our cpus aren’t those from Arizona. I know that long ago our cpu are mostly from places like Taiwan. Malaysia.
@@joydebnath_1485 no need to if you apply what wrote I did that and mine doesn’t go above 297watts 2-3 months straight not problem anymore. Even cinebench temp are way better i think around 87-91 . And score high fps unchanged
Good thing I just bought a i5 13600k 😎
This will age like milk...
i have a quick fix . place one E cores L2 cahse colum . on the other side of Raptor Cove .
alsaw move memory controller and display controller in the middle of Raptor Cove . it is called balance .
no software can fix this . this is design flaw ... fire all people that work on this lazy design ...
My 14500 non K should be ok then
Bro I am using i5 13500, in start it never does any crushing but now I am facing 4-5 game crush in a day. So beware my recommendation is disable ecore. 6 core 12 thread still enough
i dont understand channels like this ,,why are you speculating ,,what you saying doesnt make any sense
You forget OXIDATION!..That can affect randon parts and dont have SOLUTION!!
I don't think people realize how big of a deal this is, nobody should be buying intel products. They were already hurting for business, this isn't good this isn't good.
I just bought a 14700K actually and I like it
@@khensationalReviews the gist of what I get is that you should make sure your voltages are limited. So that they don't spike.
@@khensationalReviews I mean is ok, as long as it works :3
Intel copium is hilarious
Oh, and I forgot to mention in my other comment about what the actual problem is, the overvolt of a single core getting boosted to 5.8-6.0GHz, this is what Intel advertised. The chip HAS to go to 5.8GHz. It's not a motherboard issue. It's on the CPU box.
If the CPU DOESN'T reach 5.8GHz, that requires 1.55v or more, killing your CPU, then it's False Advertising. They'll get sued. They are SOL either way. They have to bring down your clocks, which will result in legal actions, or they fry your chip.
I don't know if they will survive this. Anything could happen, but right now they already have high exposure due to having really high debt. They are in a precarious position, to be sure...
Intel sells outdated, over-volted 10nm garbage and AMD is already usng 4nm.
Its not complicated, folks.
best way to save a 13th and 14th gen for the time is to disable 8 of the ecores so its now a 12900k
Yea because I nigaz runs 6ghz all core and there fine because they run a fix voltage with ecores off
So just completely guessing haha
Goofy AHH glasses
Game Tech Reviews : what D F are you talking about ?!?!?!?! The sole problem is the current, voltage and PWR going through too many P-cores with too high of a frequency for a hybrid 22nm and 14 nm design. The tranzistor logic is melting due to it being designed in the Sandy Bridge 2011 era and the silicium cannot stay solid to the point of the kathodes and anodes distances to be proper for the electrons to jump.
The tranzistor logic and design are simply over its limits in terms of temperatures on a nanoscopic level.
i have a shitty 7800x3 its cloacks max 4800 mhz 🤣
Just direct die that shit and put it on a custom water cooling loop and you fixed the problem by lowering the temps by 15-20 degrees. And you can even turn of the dumb e core 6ghz boost for even less heat
must be with a select set of produced cores we have had only 1 failure out of 30 at our place but will keep an eye out
1 out of 30 is still too high.. 3% failure is 10x too high for cpus