Hmm this is kinda interesting.
Because i was just reading trough this topic.
Maybe not related but still interesting somebody with a different board facing a similar symptom,
after updating the bios of the Gigabyte board.
Not sure if this might have anything to do with the recent ageesa updates maybe? 
Not sure but @John066 Ā“s issue looks pretty similar,
but only different brand of boards.
Maybe wendell might have an idea?
Because i couldnĀ“t really find any answer yet.
The interesting piece of information that i could gather from the other topic,
is that the backup bios of the Gigabyte board seems to work fine?
So there might be something related to the new Ageesa patches maybe?
Because the backup bios is likely an older version?
I think that this issue might lay a little bit deeper then just re-installing windowsā¦
I doubt that. The last time this happened was last yearā¦ They updated the bios again, last year. It kinda fixed itself the last time aroundā¦
Just to keep the information flowingā¦
Yesterday my PC just died while watching YouTube videos at one pointā¦
Today I plugged in an installation USB drive with W10. I ran repair, the drive told me it canāt repair my system.
Hmm thats kinda oddā¦
iĀ“m kinda curious what the core voltage was at the time it occured.
But of course that is now impossible to knowā¦
A sudden switch off could of course lead to a currupted,
OS install.
IĀ“m also start doubting the state of your psu kinda,
since you had issues like this before.
Itās possible itās his PSU but to be definitive logging the rail voltages and the load state of the system will give a better indication. If the voltage dips enough the system will pull current to compensate which sometimes is masked by OCP and thinking itās a cpu or oc related issue
I thought the same damn thingā¦ Thatās why I bought Seasonic Focus 650W Gold PSU the last timeā¦ It didnāt help.
okay then its not the psu. 
IĀ“m assuming you also already tried most other possible things,
like setting a fixed voltage and a medium llc level etc.
Slightly upping the soc voltage to like 1.1V ish to start.
I ā¦ Wellā¦ Umā¦ What?
Well which cpu do you have?
A 6 core Ryzen cpu should work fine on a fixed vcore of like 1.325V ish to 1.34V.
Not sure how your particular motherboard handles its cpu LLC profiles.
But often selecting level 3 is on the safe side with Asrock.
Vsoc something like 1.1V ish to start. (not higher then 1.2V)
But i forgot which particular cpu you have?
1700Xā¦
I think those are the files numbersā¦
A 1700X might need a little bit more Vcore.
But itās worth a try, otherwise something around the 1.36V ish,
then LLC level 2.
I donāt know what that is and where to look for itā¦
Load line calibration.
Iām kinda currious what your cpu would pull at stock stress testing.,
with everything on just auto setting.
I i think it should sit around the 1.42V ish mark normally for a 1700X
@PhaseLockedLoop what do you think for a 1700X on a Asrock board?
But that is something to try out, when Vcore isnāt adequate enough,
the system should basically just freeze / crash at stress testing.
Mmm if heās already at about 1.38 volts he shouldnāt push clocks past that. Itās just not going to do so in a stable Manor. That said he can try level 4. I like level 3 because itās flat and does not veer from my voltage settings. This is helpful because you know how far to push it.
Bumping the response to 500 khz is pretty sufficient to avoid stutters and micro droops.
Hereās what he can do. He can manually OC by starting at stock and using the offset. Undervolt as much as possible. When stable undervolt at stock is found. Bump it up 2 ticks. Then start there moving the clock and voltages. This will let him find the minimum. There is no set and forget voltage. Every VRM and CPU combo differs in performance
Yeah i know, dialing in manually numbers can be a hit or miss.
I donāt really know how his particular boards vrm is going to behave.
Maybe starting with offset might be a bit of a safer way to go.
However manually dialing in numbers could also be safely when stick around 1.36V / 1.38V max.
Then LLC 2 or 3, in a worst case scenario the system would freeze during stress testing.
It shouldnāt blow up theoretically 
But you will loose single core boost this way.
But yeah it obvious that he does seem to have some sort of an instability issue.
Asrock didnāt found anything wrong with his particular board,
but yeah idk what they did actually test, and what kind of cpu they used etc.
So yeah taking their word is still not really worth much.
The only issue i have with AMD right now, is that novice people,
should not have to deal with this kind of crap.
It technically should just work fine out of the box.
Itās a bit disappointing to see that we are 3 generations of cpuās and chipsets,
further it still does seem to be a bit flaky.
See the thing is a ton of people donāt understand LLC
Rule of thumb is
Levels:
1: voltage will spike
2: voltage will be slightly above
3: minimal Change in voltage
4: voltage will be slightly below
5: voltage will droop
Anything in between each level is a thing to on some boards with 8 levels of calibration.
Hereās the thing novices should not be concerned with overclocking in general. If you donāt OC and leave it at stock and donāt play with it. Even if temps are high itās still going to be fast and they are unlikely to notice issues day to day.
I also donāt fault AMD. They set the specs. Board manufacturers need to meet them. Itās on the mobo folk
The real funny part is, it fixed itself last time around. Now itās happening againā¦
Question: is there a possibility the system have something like a cold bug and whenever itās really cold outside the system have issues booting until it heats up a bit? Last time it was during winter the entire thing and this time it happened when the temperature neared 0Cā¦
I know, it sounds stupid, but is this an option?
Yup 
I personally never have experienced something like that.
Because where i life itās generally not getting that crazy cold, -10 is kinda exceptional.
And iām assuming that that your ambient temp during winters sits about 21C?
It would sound a bit unlikely to me, unless you a serious moisture problem in your said room?