i9-13900K/14900K Degradation Explained - Not Complicated

I know you were just rambling, and this was a few months ago. But here's one who points at 14900Ks running in server motherboards, so no overclocks. Running Minecraft server at ~58-82°C 30% of the CPUs are dead within 2 months, at stock turbo settings. They lowered the turbo so now only 5% are dead after those 2 months. Link: https://www.youtube.com/watch?v=yYfBxmBfq7k

100% accurate. You can talk about voltage, vdroop, load, LLC, IA DC etc etc. All it comes down to is what Wattage you are pulling into your chip and at what temps. Running an AVX based stress test for 24 hours at Thermal throttle pulling 350+ is just so fucking obviously dumb. It's like red lining your car all day when you were just checking that it started.

Great vid! Very educational

Another great video bro

Intel should basically have to offer a refund to anyone who wants one.

Oxide breakdown would cause catastrophic failure wouldn't it? Slight electromigration may change the AC characteristics without catastrophic failure however, though too much will cause that effect.

I have new settings which I am using for my i9-13900k and i9-14900k. I don’t like the uncertainty of undervolting so prefer working with power and clock limits. With these settings I got a couple of FPS BETTER on CP2077 and Forza Horizon Benchmark, went from 99th to 98th percentile on PC Mark Extended, passed the Time Spy Stress Test, Time Spy was within a whisker of the average for my hardware, Cinebench R23 finished between 37-38000, CPU-Z, Intel Diag tool and Prime95 was stable and < 90 degrees. MCE off, PL1 and PL2 limit to 225, limit P-core boost to 5.3 GHz and E-core boost to 4.0GHz, and use balanced power profile in Windows (although I do disable core parking to keep system highly responsive). Oh and just XMP on the RAM. I didn’t change LLC or voltage offset values. With these settings, you should have no stability issues, be able to run on an air cooler like the NH-D15 (what I use on my 3 i9 13th/14th gen systems) and will barely notice any performance changes in gaming or productivity.

I remember correctly the intel (when using xtu) sets the maxmuim for 13900k\s and 14900k\s as 320w under 70c before down bin 375a (400a for KS), 13700k is 280w under 70c before down bin 350a. I just use the stock 253 watt 307a ( I have shitty 360 aio). good video !

we are back to prescott/smithfield time after 20 years lol to begin with, I don't know why ppl choose bigger process (intel 7 vs TSMC 5/6nm)

Set pl1/pl2 to 180/300W for my i7 14700K, max temps are 92C.

is there a software that can check if my CPU has been degraded just like there is software for HDD and SDD/m.2 etc ?

I ran my 14900K since I bought it on January 11th with the Intel limits enforced. So 253W PL and a slight undervolt done with AC/DC LL and LLC4 in a Z790 Asus motherboard. AC LL - 0.30 DC LL - 1.02 LLC - 4 PL set to 253W I'm also running a 360 AIO and temps never went over 85-87~ degrees during all core workloads and 60-65 degress in gaming. The CPU was stable until a few days ago when I tried to play Calisto Protocol and the game crashed during the shader compilation (this being the first DX12 game I played in a while). Since then I tried a few more DX12 games and they all give WHEA and the "Out of video memory" error. For now it seems stable at LLC - 5 and DC LL - 0.8 I don't know if the undervolt caused the degradation or something else did it.

Can a RAID 0 cause a instability on those chips ? i got laptop i9 13980hx processor and honestly RAID 0 was making windows very unstable and sort of laggy and also I have unstable fps in every game on my laptop from day 1 of using it and just wondering if this could be because of chip degradation?

MY PC ENDED ITSELF A MINUS FORUM 690 THEN A MONTH LATER MY 75 INCH TV POOF GONE IT BLEW NOT THE BREAKER

The mobile processors have the similar issues despite the lower frequency and lower TDP =)

Dont buy a 14th gen i9 if you dont want your chip to degrade, the stock voltage is already too high for daily use case, most motherboards already undervolt them, if you are on a asus board, set svid behaviour to intel's fail safe and you will see the real stock voltage, you can undervolt for sure but remember these chips already crash at the motherboards' stock setting during some games' shader compilation process (there's a lot of reports about intel chips crash on some UE5 games recently and can be solved by downclock it)

I see those datacenter admins watching 100% CPU failure and wondering what they did wrong. lol Apparently it could only be everyone but Intel fault....but, you know, 100% fail in Linux.

:skull: I went back to my super old notes when I first got this pc and did the initial stock 2-3 runs of everything for score comparisons before moving on to undervolting, and saw I wrote: R23, 2 runs, 40.8k, max 91c, max 320w 1.288v. 2.5B: 1 run, 66.650s, 343w, max 100c. That was at default bios LOL. Well.. I only ran it once or twice or thrice like that before doing undervolting, and my all core voltage requirements seem to be the same-ish now a year later, so I think I got away OK... I stopped doing Y-C 2.5B 1 minute run after some point in the early months. Probably ran it 15~20 times max over lifetime including fails since it took so much power.

dear sir, if i run my 13900ks (240-250w) consistently 6 hours a day on 85-87 celc degrees, will it degrade soon ? sorry i dont understand much of video, bacause of my poor english....thank you

The only CPU that has caused BSODs in Windows 11...