13700K Degraded in Minutes?

[Graptik]

I upgraded to a 13700KF about a week ago and have been tweaking my overclock of it ever since. I got it stable at 5.5GHz all P-Cores with E-Cores disabled. I also overclocked my RAM from 3600MHz CL16 to 3900MHz CL17.
This configuration was rock solid stable, passing hours of Cinebench, Y-Cruncher, TestMem5, Realbench.
Everything was going fine until I decided to see if I could get my RAM stable at CL16 with 1.45V. I managed to get this stable but then the next day, I started to get blue screens on Windows.
It has gotten to the point now where I can only get the system stable at 5.1GHz all core, even opening applications at anything above this will cause the system to crash. I have clean installed Windows twice but the issue persists.
The weird thing is that the RAM does not seem to be affected at all, I'm still using my 3900MHz CL17 overclock and even the CL16 1.45V overclock is still stable, it is only the core ratio that is affected.
I'm wondering if using 1.45V on the RAM has degraded the CPU, even though it was only using that voltage for a day. That is the only explanation I can think of as to why there has been such a massive drop in the capabilities of the CPU.
This is extremely frustrating and any help is appreciated.

Specs:
CPU: i7 13700KF
RAM: Kingston Fury Renegade 32GB 3600MHz CL16
GPU: EVGA 2080ti FTW 3 Ultra
PSU: Corsair 750 RM750X
Cooler: Noctua NH-D15S

 

[MartinPrince]

What were your SA and VDDQ voltages?

 

[Graptik]

What were your SA and VDDQ voltages?

VDDQ was left on auto which is 1.2 and SA was 1.13. My motherboard actually gives a lot of voltage (1.4V!!) to System Agent on Auto for some reason, but I put it at 1.13 and it was stable.

 

[MartinPrince]

And what were your Vcore voltages for the overclock and what motherboard?

 

[Robert896r1]

Manually set all your ram timings to avoid instability from retraining.

Pull the power. Leave it off for a min. Plug power back in, clear cmos and then load your profile.

Reminder that when you tune ram, it requires more vcore as well.

 

[Cyber-Mav]

Doubt it was stable to begin with since you didnt stress test it properly

 

[Graptik]

And what were your Vcore voltages for the overclock and what motherboard?

CPU Vcore was at 1.28. Motherboard is MSI Z790 Pro-A. I've found out that it's not even stable at stock now. Resetting everything to optimized defaults in the bios, it won't even run 5 minutes of cinebench without giving an error. Something has definitely gone wrong, just not sure whether it's the motherboard or the CPU.

 

[Graptik]

Doubt it was stable to begin with since you didnt stress test it properly

I absolutely did test it properly. As I stated in the OP, it ran hours of Y-Cruncher, Prime, Cinebench, TestMem5 and Realbench, now it won't run 5 minutes of Cinebench at stock.

 

[Cyber-Mav]

Didn't see core cycler on that list.

 

[=XDC=FluphyBunny]

Didn't see core cycler on that list.

Give over.

OP try reseating your CPU.
Guessing here but time to start at the beginning if defaults aren’t working.

 

[cjgardens]

I absolutely did test it properly. As I stated in the OP, it ran hours of Y-Cruncher, Prime, Cinebench, TestMem5 and Realbench, now it won't run 5 minutes of Cinebench at stock.

Needs to be 72 hours prime95 no errors to be stable

 

[McGraw]

Needs to be 72 hours prime95 no errors to be stable

I've never run a stress test for more than a few hours. If it's stable in normal use it's stable

 

[cjgardens]

I've never run a stress test for more than a few hours. If it's stable in normal use it's stable

I do about 10 mins then play some games

 

[OC2000]

If wasn't 100% stable to begin with, you may have corrupted the BIOS. Reflash the Bios and start again as mentioned

 

[MrFunex]

Needs to be 72 hours prime95 no errors to be stable

Flat out for 3 days? A bit overkill?

Surely after 30-60 minutes, regardless of how beefy your cooling choice is, everything’s running in a constant state?

 

[cjgardens]

Flat out for 3 days? A bit overkill?

Surely after 30-60 minutes, regardless of how beefy your cooling choice is, everything’s running in a constant state?

See post 13.

I was joking because of all the overkill testing OP does

 

[Cyber-Mav]

I've never run a stress test for more than a few hours. If it's stable in normal use it's stable

If your normal use of the computer is for anfew hours then i guess that stress duration may be sufficient

 

[Tetras]

I'm wondering if using 1.45V on the RAM has degraded the CPU, even though it was only using that voltage for a day. That is the only explanation I can think of as to why there has been such a massive drop in the capabilities of the CPU.
This is extremely frustrating and any help is appreciated.

I wouldn't think so, people have used that kind of dram voltage with 12th gen CPUs. What were you using to get 5.5 Ghz stable on the P-cores?

 

[MartinPrince]

CPU Vcore was at 1.28. Motherboard is MSI Z790 Pro-A. I've found out that it's not even stable at stock now. Resetting everything to optimized defaults in the bios, it won't even run 5 minutes of cinebench without giving an error. Something has definitely gone wrong, just not sure whether it's the motherboard or the CPU.

Your voltages weren't particularly high at all so I doubt that they are the reason that your CPU isn't working correctly now. I would RMA the CPU if you haven't done already as that is most likely to have some kind of defect.

 

[pastymuncher]

Have you tried updating the motherboard bios? These are new boards and can be buggy at first so it's worth updating the bios when a new release comes out. The latest for your board is 7E07vA4 and was only released last week. There have been 4 new bios releases after the launch bios, all of which improve compatibility of memory kits.