What to do with a currently stable 13700k wrt risk mitigation and potential future damage

[deadpin]

Context:

• 13700k system purchased and installed May 2023 (CPU box with id numbers gone since December…)
• PL1 173W, PL2 225W since install
• 128 gb, 4 sticks @ 4800 (running stable for my workloads since ~4 months after install)
• No noticeable decrease in stability or increase in problems etc. Routinely only reboot for security patches ~35 days or so.

The reason I chose intel to begin with was better DDR5 support. It was almost plug-in-play on day 1; it booted, trained, and ran at speed immediately. There was just the occasional hang after ~7 to 14 days of use. I eventually narrowed it down needing to bump the VDDQ/TX voltages up from 1.1v to 1.16v. I do not use XMP.

Since stability sorted itself out, my bios/micro-code is only from June 2023 as there was no need or desire to change after I updated once to try to address the initial stability issue after purchase in May.

This is a personal computer for work, not play, so I need it to be functional.

But now what? Come “the middle of August” when we get more information and micro-code, is anyone considering just kind of pretending the problem doesn’t exist to not rock the boat?

The nature of the RAM hangs is certainly concerning and I worry that moving my bios/micro-code forward so far will (re)introduce gremlins. Superstitious, maybe, but not unwarranted. And asking how likely it would be to regain RAM stability at 4800 is pointless; no one will know for sure.

Would my PL1/PL2 setting be preventing most overvolt damage maybe?

Is anyone else in a similar situation? If so, how are you approaching this mess?

[twin_savage]

@deadpin

Since stability sorted itself out, my bios/micro-code is only from June 2023

There’s a very good chance your microcode has already been updated by the OS many times already since your last BIOS update.
You should be able to poll what version of microcode you are on in command prompt/terminal.

@deadpin

Would my PL1/PL2 setting be preventing most overvolt damage maybe?

From what I’m hearing these settings won’t help much; limiting maximum core frequency seems to be the most effective way to prevent the voltage excursions.

[Charles_VT]

Setting a max multipler of like 52 seems to help, setting a dynamic undervolt couldn’t hurt

[deadpin]

@twin_savage

There’s a very good chance your microcode has already been updated by the OS many times already since your last BIOS update.

Oh boy lol. Looks like I’m running revision 113 here released May 2023. Good job Windows?? Version 125 is currently available from my MB vendor to address the first eTVB problem it looks like.

@Charles_VT

Setting a max multipler of like 52 seems to help, setting a dynamic undervolt couldn’t hurt

Thanks! This is a good escape hatch. Maybe if nothing else I should proactively do this if I don’t want to take the larger plunge…

[twin_savage]

Microsoft is the kid in the back sniffing glue.
Presumably Intel is going to force Microsoft to implement the raptor lake μcode update via Windows itself once ready, because there is no way all the normies are going to update their BIOS, nor should they due to the dangers.

for those that want to verify what their Windows system is running, run reg query HKLM\HARDWARE\DESCRIPTION\System\CentralProcessor\0 | find “e R” to check:

It seems like Windows only bothers to have μcode updates for major issues like security fixes, Linux is more generous in this regard and gives more regular updates.