Computerus

[Mongrel]

So, I'm looking at new PSUs potentially (Black Friday sales make that both easier and more difficult), but I had a small wrinkle in the meantime.

At one point when we were in the middle of the thermal paste application we had a bad application (or what we thought was a bad application), such that any game loaded would be incredibly low FPS and lag severely in all respects, including input. Like, open main menu and then immediately close game without actually bothering to load an instance or save, levels of bad. This has now reoccurred and is continuous, including through restarts and complete shutdowns with an extended cooling period.

Because I still don't know if it really is the PSU, and here we have an actual ongoing symptom of failure (instead of random outages), what do I even look at to try and figure out what the bottleneck is? Of course now that there's an actual ongoing failure, hopefully I can just shortcut this by swapping PSUs briefly to test.

Forgot to mention I had switched to HWinfo64, which is apparently the most unversally-recommended software for this. It has an incredible amount of information, but of course in most cases I lack the knowledge to know what much of it means or what to look for. I also for the life of me can't figure out where its stress testing utilities are (which it's supposed to have?).

In any case, if the problem is the power supply, I wish I knew what hell we did during heat sink replacement that corrected a massive graphical processing fault (and this could be a CPU or GPU-based issue!), albeit temporarily.

[Mongrel]

Well, I swapped PSUs with Starr and... my computer didn't even come on, not even the case lights which come on as long as your PSU is switched on, even if the computer itself is totally shut down.

I did have to switch some cables because her PSU has different sockets, but again, we have the exact same computer build other than the case and the PSU (and I have an additional HDD). It's not like we have a big mess of cables either, literally just Mobo, CPU, GPU, and for her 2x 6-prong PCI-E for her case stuff, and for me 1x SATA cable for both my HDD and case stuff, with the weird flat-5 prong SATA connection (my PSU has those sockets for accessories, hers doesn't).

I thought, well, maybe it's my Mobo and it just gave out now from all this handling? But no, we replaced both our original PSUs and each computer runs again.

I DON'T UNDERSTAND.



EDIT: Graphics issue turned out to be the drivers just went... missing? (Windows display error 31, no drivers) This was directly after one of the hardlock crashes (first time since I upped the fan speeds). Reinstalling fresh drivers is always good from time to time, but uh, mine were literally working before the failure since I was right in the middle of a game and then they just, I dunno, got up and fucking walked away? Windows is a funny beast, I guess. And bad power can fuck your shit up... if that does turn out to be the issue. Though now I have no idea what will happen if I buy a new PSU! Fun!

So I still don't understand, but it's probably not the GPU failing? Maybe?

[Mongrel]

New PSU obtained and installed. Everything seems to be working fine for now, but it will take days or even weeks to be sure. Maz & Upthorn, I'd like to stress that I'm not going to blame either of you if it doesn't, because that was as valid a suggestion as anything (and having a spare PSU is a sensible idea anyway).

I can only assume the issue of Starr's PSU not working in my comp is because of some proprietary cable finicky-ness. We didn't pull out her PSU cabling because her case's power bay is an enclosed area which is a nightmare to work in.

(I've noticed that newer cases of identical design add a generous access cutout to the side of the power bay - better late than never, ya dimbulbs.)

[Mongrel]

Welp! That wasn't it!

I don't even know what to do now. Rerun tests I guess.

[Mongrel]

Okay, possibly an angle here. Was looking up the C: check in the event viewer and realized I should have been checking here much earlier. Errors, including critical hardware failures have in fact been logged all along.

A fatal hardware error has occurred.

Reported by component: Processor Core
Error Source: Machine Check Exception
Error Type: Cache Hierarchy Error
Processor APIC ID: 5

The details view of this entry contains further information.

The Good news is that there's a number of threads covering this situation on Reddit, the AMD forums, the Windows forums, and a number of other places. The Bad news is that it actually isn't a consistent issue and different users have solved the issue with

- Various driver/firmware reinstalls
- Fiddling with various BIOS settings
- Replacing GPU
- Replacing CPU
- Replacing RAM
- Replacing Mobo

So, uh

[Mazian]

I did have to switch some cables because her PSU has different sockets,

You always want to swap all cables when exchanging modular PSUs. On the PSU side, even if (some of) the connectors may match between two models, the pinouts are not standardized and may be entirely different.