WHEA Uncorrectable Error among others

@RJARRRPCGP I can't reply to private messages yet so I have to thank you from here :)

Under normal circumstances (meaning mostly chrome tabs open and not stressing or gaming) cpu temp is around 40-45 degrees and motherboard just shy of 50ies. During the tests they did reach the low 90ies, I think the highest I saw was 92.

Unfortunately I don't have any spare parts to test with as they were used to build another pc which limits my options.

I saw the Prime95 link but it's way above my knowledge of how a computer works. I'll run the FFT tests and see what happens. I googled the errors that Prime95 found and they both seem overclocking related even though I've never messed with overclocking in any form. Wouldn't that point to a problem with the cpu rather than the memory? Although at this point I'm willing to believe a genuine curse might be responsible, I'm probably looking at a new memory and cpu right?

On the other hand, since updating the BIOS I haven't crashed yet, though without a reliable way of triggering a crash it's a waiting game till it happens... or not. :/

Not a good sign, you shouldn't even be getting Prime95 errors like that unless you were getting greedy with maximum core overclocks, usually happens not until at least 4.5 Ghz+ with your listed gen of K. And that would be when you upgrade the cooling and enable CPU load line calibration, which stabilizes the Vcore.

If you don't replace your heatsink with a better one, you're highly likely going to face throttling (auto underclocking) and possibly random shutdowns.

Prime95 is designed to stress physical components of the memory sub-system of a computer.
It's 3 tests stress RAM, CPU cache, and the memory controller - and each different test stresses them in different ways.

While it can be due to overclocking, or to an unstable CPU core - the most common problem (and also the least expensive) is problems with the RAM (especially when it happens with the Blend stress test). It's to your advantage to pursue further diagnostics (such as the Intel Processor diagnostic, or by testing the RAM sticks individually, testing them for longer times with MemTest86+, or even trying the Small/Large FFT tests).

If you don't pursue further diagnostics, you'll just end up throwing money at the problem until it goes away. The possibilities for failure are in this order:
- RAM (most likely)
- CPU
- motherboard (least likely)

If the problem is with the motherboard and you start throwing money in order of most to least likely - you'll end up buying new RAM, then a new CPU, then a new motherboard. You're lucky in that the RAM is likely under warranty - so getting it exchanged shouldn't cost you anything. But if the other items aren't under warranty, you may end up spending hundreds of Euro's just to figure it out.

Good luck!

Well small FFT test gives me an error less than 5 mins into it. Ran it 2 times and got these errors:

FATAL ERROR: Rounding was 0.5, expected less than 0.4
Hardware failure detected, consult stress.txt file.

FATAL ERROR: Rounding was 2422.086809, expected less than 0.4 (wtf)
Hardware failure detected, consult stress.txt file.

I also ran two large FFT tests, one was around 15 mins and the other 30 mins. I got no errors on either. If we assume there's a pattern here does this eliminate any possibilities?

Actually, it makes things a bit uncertain.
If you only fail the Blend test, we assume it's the RAM
If you only fail the Small FFT's, we assume it's the CPU cache
If both happen, but the Small FFT's fails faster than the Blend test - then our best guess is that it's the CPU cache
But, I'm concerned that it's not certain enough to warrant purchasing a new CPU - that's a lot of expense for something that's not very sure.

From post #5:
Run the Blend test on the individual sticks.
Run MemTest on the individual sticks
Run the Intel Processor Diagnostics

Hopefully that'll help point to the problem component!