[SOLVED] Frequent BSODs after reinstalling Windows, now not booting and cannot reinstall Windows again - RAM gone bad?

[twistedtoaster]

Wasn't sure if I should post this under Hardware>Memory or here, but here goes. There's a bit of background necessary, but over the past few weeks (great timing with COVID19 right?) I've been having issues with my desktop, which I am now no longer able to boot into Windows. Trying to reinstall Windows from a usb results in the installation stopping at around 10% and giving the 0x8007025D (Windows could not install the required files) error, similar to that described in this question here.

SPECS: (built September 2019)

CPU: AMD Ryzen 3600
Motherboard: Gigabyte X570 Aorus Elite
Memory: G.Skill TridentZ RGB 2x8GB DDR4 3200 MHz (14-14-14-34; F4-3200C14D-16GTZR; note that this isn't the GTZX model specifically for AMD CPUs, as availability for that was/is basically nonexistent where I am, but the GTZR hadn't given me any problems until this April)
GPU: EVGA RTX 2080 Super XC2
PSU: Corsair RM750x
Cooler: NZXT Kraken x62 AIO
SSD: 128 GB Samsung 840 PRO (from my previous desktop, used primarily as a boot drive)
HDDs: 1x 2TB WD Black and 1x 5TB WD Black drives (also from previous desktop).

First Issues:
Prior to the end of March, I hadn't had any issues with my current setup. At around this time I mistakenly double clicked a .reg backup file I had made, for when I disabled my windows key. Immediately after it ran, explorer stopped working properly, and on restarting the computer it wouldn't boot and couldn't run the repairs. I was able to make a bootable USB (after installing Windows on my OSX laptop in Parallels) and reinstall Windows after which I think things were mostly fine for about week, until started to re-enable my overclocks.

Issues reemerge:
So when going to set my RAM OC (I had been running CL15 3600 MHz; whatever was 'safe' according to the DRAM calculator for Ryzen) I <Mod Edit> up and had the profile set to 'manual' and not 'V1', so the timings were all whack and voltage high (1.4V), and like an idiot I didn't notice this while setting the timings and it wasn't until I had a couple BSODs and resetting back to SPD and recalculating the timings I realized my mistake. Setting the appropriate timings seemed to work for a couple days, but I started to have some random programs that would stop working and eventually I would BSOD. I decided to set the RAM back to SPD which didn't seem to fix anything, and I would still get a BSOD every couple of days after a couple of sleep cycles. BSODs weren't always the same, but all seemed to be memory or driver related (irql_not_less_or_equal, or apc_index_mismatch). I tried updating all drivers I could, even using Auslogic's driver updater (in hindsight, it seems from several threads on reddit that this wasn't the best idea). Jump to today, I am working on a PDF and get a MEMORY_MANAGEMENT BSOD. Briefly I was able to boot into safe mode, but now even that won't work.

Reinstalling Windows:
In a last hope that it isn't a hardware issue (even though these problems all started again after my screwup with the memory), I went to reinstall windows again. I have the same bootable USB drive I used last month, but I get the 0x8007025D error I described at the start.

Bad RAM, or software issues?
It should be noted that my earlier Windows reinstall, I kept the old files (where it creates a Windows.old folder), and hadn't gotten around to deleting the folder yet, but I don't imagine that any registry errors could have still been lingering after the reinstall. Swapping memory slots, and/or trying memory sticks one at a time doesn't seem to change anything, as I still get the troubleshooting menu on booting (or failing to boot) into Windows. The BIOS still reads and recognizes both sticks correctly, and in all four of the slots though. I was planning on running memtest today, but this happened before I setup a USB with it, and the one I wrote to a USB while in safemode didn't seem to work correctly as it would just load into a black screen with blinking cursor.

So, does it sound like I likely have a memory problem? Because at this point everything seems to be pointing to that unfortunately. I don't have any compatible RAM sticks to try to see if those work. I've never had to RMA anything, but I'm not sure if I'd even be able to RMA the memory since I don't know if this is something I did, or if the timing was coincidental. I also don't know anyone here that has memory I could borrow to troubleshoot either. If I did RMA it and got replacement ram sticks, and it still didn't work, I'd also be in the same boat.

Any help or suggestions for further troubleshooting would be much appreciated!

 

[RealBeast]

The easiest way to know is to run the free version of MEMTEST86 from a USB stick, which bypasses your OS since it runs on a small Linux build on the stick. Run it for 3 or 4 passes. No errors is good, anything else is a problem, which may be hardware or settings related.

 

[twistedtoaster]

The easiest way to know is to run the free version of MEMTEST86 from a USB stick, which bypasses your OS since it runs on a small Linux build on the stick. Run it for 3 or 4 passes. No errors is good, anything else is a problem, which may be hardware or settings related.

Is there a way to set one up from a Mac running OSX? I ran the exe and converted a USB while in safemode on my desktop, but found that it would hang on a black screen with blinking cursor when booting from it so I'm not sure if it was setup correctly, and now I can't boot back into safemode.

EDIT: I found that the readme contains instructions for doing this in terminal, so I'll attempt that and see if it works.

 

[twistedtoaster]

The easiest way to know is to run the free version of MEMTEST86 from a USB stick, which bypasses your OS since it runs on a small Linux build on the stick. Run it for 3 or 4 passes. No errors is good, anything else is a problem, which may be hardware or settings related.

So far I've run 2 passes with no errors, and will let it run the next two passes overnight (since the free version is limited to 4 passes). If it turns up with no errors in the end, or only a couple after 4 passes, is this likely to have caused the frequent BSOD'ing and corruption of the OS? Because I'm not sure what would have been the issue. All drivers were updated, and none of the BSODs I think indicated a video driver or something. I also ran all of the other usual checks like running windows memory diagnostic, sfc scanning, restricting startup programs, and checking the SMART status for my drives.

The only other thing I can think of is that after reinstalling windows the first time I updated my bios (to F12F from F4, one of the first bios's after Aorus Elite X570 board was released). In the patch notes for this and several of the earlier bios versions was "improved memory compatibility", so I'm not sure if this could be causing issues.

 

[RealBeast]

I had a lot of memory issues with using the latest bios on an Intel build with their newest CPU on a little older board, going back a few versions did the trick, so I wouldn't discount that idea.

 

[twistedtoaster]

I had a lot of memory issues with using the latest bios on an Intel build with their newest CPU on a little older board, going back a few versions did the trick, so I wouldn't discount that idea.

Hmm, in that case I'll try switching to an older BIOS later today. MEMTEST86 came back with 0 errors on all 4 passes, so I'm guessing that my memory isn't the main problem (if it is one at all). I noticed that when going into command prompt while in system recovery mode, it loads into X:\, which isn't shown on my list of volumes. Additionally, in diskpart list volume, it appears that all of my drive letters are different, with the 100 MB system recovery volume being listed as Drive C:, and my actual C: drive being drive E:, etc. Disconnecting my other HDDs didn't seem to fix this, and manually assigning the C: drive back to C: doesn't seem to work, as it is reset after trying to boot back into safe mode. My boot SSD is quite old now (8+ years), and although the SMART was listed as healthy and it is stated as healthy in diskpart, I wonder if something has corrupted on the drive?

At the moment I'm thinking of going back to an older BIOS, going into command prompt and saving the few user files I need to onto one of my storage drives, and then trying to install Windows 10 on my SSD after a full format.