Intel CPU instability testing ideas?

[Akumatie]

I have a 14700K and had it since launch, and my CPU has been fine been updating my bios and just doing the 253w limit on my pl1 and 2, and I want to test my CPU see if it was affected, I did do a nvidia graphics driver install and that was fine any other ideas?

[WereCat]

Play UE5 games as for whatever reason it seems like UE5 engine is quite heavy on some instructions that usually manifest in instability on those CPUs. 

[Haswellx86]

There is no such thing as "affected". The problem was a problematically fast degradation issue.

 

All electronic components and semi-conductors degrade over time. Ten years ago, if you had overclocked your processor and tried overclocking it the same today, it will most probably not be able to hit those frequencies because of degradation over time. It's just that it is not a problem as even with the degradation it still can run within its performance spec and has not managed to actually cause a defect in the micro-architecture.

 

However the crashing issues that plagued Intel recently was a problematically fast degradation issue caused by indirect and high voltage requests. I don't think that there were only a part of the processors that were affected (or at least not by some other factors like oxidation issues which was another problem). I really just thought that these processors are just really hard workers which need a lot of juice and in turn harming themselves, but among testing and close look by enthusiasts, it really was the microcode issue requesting/allowing requests of really high voltages.

 

The only reason many people did not face any issues was because of -

 

1) Silicon lottery - Your processor might just have really good silicon quality which could withstand those voltages.

 

2) Workload - Lighter workload will preserve the life of your processor. During this problem, it was also noticed that purely single threaded workloads degraded processors faster than the all core loads because in single threaded workloads, the single core boost would be activated and boosting the frequencies and in turn, the voltages really high, where as the frequencies in all core loads will not be that high because of power limits, current limits, voltage droop, thermal limits, turbo ratio limits, and others that I am not aware of. This is a good video explaining to reduce the single core turbo multipliers, and upon that, I would still recommend undervolting and power limiting (tbh, its all the voltages and PL are fine actually. If you want performance you can raise PL).  

 

3) Settings - People with the Intel Default or Intel Default Extreme power profiles active instead of any indirect overclocking power profiles set by manufacturers are said to be more on the safer side. And as I said, undervolting and power limiting should help. Mainly, setting a core voltage limit will definitely help remove this issue. Running in balanced power profile in Windows also helps, but I would not recommend it.

 

So if you don't see any problems, your processor is running just fine.

[Haswellx86]

Watch this actually -

 

 

[Ebony Falcon]

Occt extream mixed varied for cpu stress

prime 95 small fft

geuss you can use burn test if u like jays 2 cents lol 

 

then something real world like rendering a video, install tlou and lets the shaders compile for 30 mins ect

 

[CatTNT]

For me, I used this Jayz video and used CB r15/ R23, Prime95 small ffts, OCCT, Black myth wukong benchmark tool, and Intel Burn test. In my experience, primt95 didn't help me diagnose anything, and the best tools for testing stability were OCCT, the  Wukong benchmark tool, Intel burn test, attempting to start up an mudded installation of Minecraft. 

I used Wukong at maximum settings, cinematic quality, full RT, highest DLSS settings, and if there was soft program crash or BSOD, then then my CPU settings didn't work. Using OCCT, in the CPU stability tester, I used Normal Mode, Variable load, Auto Instructions, and cycled through all cores for 2 minutes each. In the intel burn test, i did as Jay said, 100 cycles at normal load. If that all worked, I would try playing an actual game, and it seems that when I thought the CPU was stable, starting up minecraft would give me a memory access violation and BSOD a few seconds later. 

Intel 14900K, 1.39v VCORE, -2 P core turbo ratio offset, memory underclocked to DDR5-5800 CL36.