German_John

Sorry for taking so long - I fell ill over the weekend. I first tried a BIOS update, but that didn't fix anything. Still the same odd RAM behaviour and crashes while installing Windows 11. (Note: I actualy re-investigated the RAM issues a bit. Working configs were DIMMs in slots 1 or 1 and 2; DIMMs in only slots 3 and 4 would not POST; DIMMs in 1, 2, and 3 wouldn't POST; DIMMs in all four slots POST, but only the DIMMs in slots 1 and 2 were recognised) So I installed the new CPU and Windows 11 installed without a hitch. Since I had already tried re-seating the old CPU previously, I think I can rule out issues with mounting pressure, though a part of me wants to try the old chip again and see if it works... Maybe on a weekend someday. I think my memory controller was on its way out. I still have absolutely no idea why the Ubuntu install worked flawlessly, but I suppose it could have just been a lucky break. After all, the installed Windows 10 also worked okay-ish, and the Windows 10 installer got farther than Windows 11. So odd behaviour all around. I'll mark this solved for now; if it ends up crashing up again, I'll make a new thread.

Glad I'm not the only one who's a bit stumped. I haven't tried letting it sit in the BIOS actually. The RAM issues today made me suspect a hardware fault after all (memory controller maybe). I ordered a new CPU - if it doesn't work, I can just return it, and if it does work, I'd have needed a new CPU anyway. I guess it's possible that Windows Update is to blame, or it just kinda happened at the same time (or the CPU started dying and that caused a failed update) - hard to tell, especially since I wasn't there.

Okay, I've tried a few more things and I think my CPU is on its way out - though that does not explain why Ubuntu worked, even stress tests. New things I've tried: New USB stick - the same one from which I installed Ubuntu yesterday. No change. Single stick of RAM - no change Single known good stick of RAM - no change However, during the single stick of RAM test(s) I noticed: Slot 1 (A1 I believe): same results as previously Slot 2 (B1 I believe): first, did not post, then posted after a few attempts. Obviously, no success Slot 4 (B2 I believe): same behaviour as Slot 2 -- both tested with different sticks of RAM I then re-seated my CPU and tried again. Same results: Slot 4 does not work, Slot 1 works as before. In the BIOS, I then saw the following behaviour: If I install RAM in Slots 1 and 2 (both the same channel), they are detected as expected If I add one in Slot 3, all 3 are detected fine If I add the last one in slot 4, both Slot 3 and Slot 4 are not detected. I didn't check that in the BIOS yesterday, but I definitely remember Ubuntu detecting all 32GB of RAM, since I had to manually configure the stress test and monitored it. So... this probably sounds like a CPU failure? Or could it be the motherboard? And I'm still extremely confused as to why Ubuntu worked fine yesterday, though I suppose CPU failures can manifest in strange ways. I'd expect it to crash under a stress test, but who knows.

I'll try another USB drive, though I don't think that's the issue. Windows 10 install got to the point where I could remove the drive - and it's brand new, literally took it from its packaging today. I'll try though, simple enough. I used the official tool. Maybe rufus would be better? Official Microsoft seems to not exactly be a guarantee of quality...

I've now tried further troubleshooting steps: Creating the installation medium on a different computer (that has no issues) Installing without my RTX2070 instaled, using the integrated graphics Creating a Windows 11 medium instead of a Windows 10 one None of them worked. Windows 11 actually dies earlier than Windows 10. It crashed after selecting the partition twice (both times at 16%), but also before that. Windows 10 only crashed when during setup, so after I used the USB drive for installation. I honestly have no idea what's going on. I guess I'll have to come back to this tomorrow.

Hm, interesting. That might be it. Though it wouldn't explain why the crashes began in the first place. I don't think I'm installing optional preview updates, but I suppose it's possible I checked a box somewhere at some point. Would you agre that it's probably not hardware related? Currently stress testing in Ubuntu (CPU, GPU, and RAM) and it's been stable for the past 20 minutes. I'm also creating another Windows install medium on a different computer, maybe that was the issue.

I don't think that's it. The crashes usually happened when I wasn't doing anything taxing on the GPU (e.g. just sitting at the desktop). Also, I ran furmark at one point (in the original Windows 10 installation) and it ran quite happily. I'm currently (in Ubuntu) running stress tests on CPU, GPU, and RAM - all pinned at 100% usage. No crashes. Or is there some weird hardware failure on a GPU that would trigger voltage spikes randomly?

Hey all, I'm at a bit of a loss. My computer crashes seemingly randomly, no BSOD and nothing in event viewer except that a driver fails to load for a device (ROOT/SYSTEM/0004 - I can't find it in device manager) Since I've tried to reinstall Windows, event viewer is now empty and I forgot to make a backup of that. I originally had that issue about a year ago (when I first built it) as well, but that turned out to have been my own stupidity (I carried over the drive from another computer, resetting Windows fixed that). Since then, no issues. The problem seems to have started while I was away for a few days; I left my computer on so I could remote into it, but ended up not needing it. Event Viewer originally showed a failed update on the 8th of August, and frequent (but irregular) crashes since then. Unfortunately I cannot revert to the pre-udpate state since it was more than 10 days. Over the last few days, I removed all connected devices except a USB keyboard and mouse, ethernet, and a monitor over HDMI. It ran for days without issue, but all I did was create backups. Weirdly, those failed in odd ways - sometimes, it'd just be done when I came in but didn't copy over some folders (no system folders, it was missing e.g. a bunch of folders containing mp3 files). I tried to use the "reset this PC" option in Windows (also booting from safe mode) a few times, but it wouldn't work. Once it allowed me to download Windows to reset, the other times it didn't - and it failed every time. I ordered a new SSD (mostly because I wanted more space anyway, though I thought that would also rule out SSD failure), created a Windows 10 install medium and tried to reinstall Windows, but it keeps crashing. Not always at the same place. Sometimes, it'll crash before I can even select the keyboard layout; sometimes, only after I set my user. I've also tried creating it on a different computer, same results. I also tried installing without my RTX2070 installed, running off of the iGPU of my Ryzen 5700G. Also crashes every time during install. Tried a Windows 11 install medium instead of a Windows 10 one. Broke even earlier than Windows 10. I was able to install Ubuntu 24.04 without issue. It obviously looks like a hardware issue, but there's a few things that stump me: It happily worked while I was doing backups, no issues during that whatsoever Ubuntu installed without a hitch It all seemingly started after a botched update - which made me think it was some deep seated driver issue Was it maybe that I created the Windows install medium on the PC that I was trying to reinstall Windows on? Specs CPU: Ryzen 5700G Motherboard: Gigabyte B550M DS3H (rev 1.3) RAM: 4x8GB GSkill Aegis 3200MHz (currently, XMP is turned off to rule that out, so running at 2133MHz) GPU: Acer Predator RTX 2070 PSU: Seasonic G12 GM 650W Case: Fractal Pop Air Cooling: Arctic Freezer 36 Storage: 1x 1TB Crucial NVMe SSD (Windows 10) - this is the drive it started on 1x 2TB SanDisk 7100 SSD - trying to install Windows 10 to that with no success. Other troubleshooting steps I've taken (just in case they're relevant): Uninstall SteelSeries GG; a few times, I felt like it crashed when I plugged in my SteelSeries Aerox 9 mouse. I'm currently using an old Razer mouse that I still had. Update nvidia drivers - no change I deliberately did not do a BIOS update, since, if this is hardware related, I'm a bit worried about something crashing while I'm doing it.

Sorry for the very late reply - I wanted to make sure the PC is actually stable (it had a stable period of a few days before). Seems to be solid after a re-install of Windows.

Sorry, I seem to have skipped the most important bit. I upgraded CPU, motherboard, and PSU (and case and CPU cooler). I'll reinstall Windows and see what happens

See, I thought that as well, but it did also crash (apparently) in Ubuntu, which surprised me. But yeah, I guess I'll reinstall Windows tomorrow and get back. Thanks!

Hey all, I've got a real head scratcher. My PC randomly crashes - hard reboots, sometimes freezes. No BSOD or anything. Specs CPU: Ryzen 5700G (recently bought used) Motherboard: Gigabyte B550M DS3H (rev 1.3; recently bought used) RAM: 4x8GB GSkill Aegis 3200MHz / 1x16GB Corsair Vengeance 3200MHz (XMP off in both cases) GPU: Acer Predator RTX 2070 PSU: Seasonic G12 GM 650W (new) Case: Fractal Pop Air (new) Cooling: Arctic Freezer 36 Storage: 1x 1TB Crucial NVMe SSD (Windows 10) 1x 960GB Kioxia SATA SSD (Ubuntu 20.04) 1x Toshiba 3TB HDD (mass storage) Issue: Random crashes, either hard reboots or freezes. No blue screen. Happened initially, went away after turning XMP off. Returned after a few days, with very high frequency (often, Windows would not be stable for more than a few minutes). Though occasionally, Windows will be stable for several hours. In Ubuntu, two crashes seemingly happened, but I wasn’t around for either one, so I can’t verify what happened, and the system was stable for many hours (as opposed to a few minutes for Windows). Troubleshooting steps I've taken (to no avail): Turning off C-States in the BIOS Swapped RAM (see spec list above) Other issues with the PC: Ground(?) noise through one of my USB audio interfaces (Presonus Audiobox USB 96) when GPU is under load; other interface (Behringer UM2-Phoria) does not have these. I don’t think it’s tripping overcurrent protection (or other PSU issues, really). Crashes do not occur under load (CPU or GPU), but seemingly random. I suspected hardware issues, but since Ubuntu was so much more stable, I’m very confused. If it were e.g. an issue with the CPU, I would expect the issue to persist more across OSs, but I might be wrong. I also do not think it’s a thermal issue. Running Prime95, the CPU does turbos up to 4.4GHz with temps going up to 60C (Packages, cores max out around 50C). GPU running Furmark gets a bit toasty (over 100C on the hotspot), but I was able to run Far Cry 6 (which is pretty hard on my 2070) for a good while without crashing. I did not do a fresh install of Windows (or Ubuntu, for that matter) after the upgrade. I would like to avoid that (for laziness reasons, lots of manual config for my work), but I did make a full backup yesterday and have a Windows 10 install stick ready (created yesterday, too). I thought I might get away with it since I stuck with AMD. Event viewer shows a few warnings and errors. Warnings mostly relate to DistributedCOM. Errors relate to TPM-WMI (I disabled Secure Boot) and NetBT. The event logging service also encounters errors for channel Nefarius-Drivers-HidHide. The most recent error before the last crash was: The Secure Boot update failed to update a Secure Boot variable with error Secure Boot is not enabled on this machine.. Though it happened several minutes before the crash, so I don't think that was it. If you have any insights, I'd be very appreciative.

As I mentioned in my edit, I am already running them at 2133MHz, it was still BSODing all the time.

Interesting, hand't heard of flex mode. I suppose I'll just torture it with driver verification for a few more hours and if that's stable, it probably won't BSOD. Thanks!

Hey everyone, after fighting BSODs for a while, I think that either a DIMM or a RAM slot is broken. No BSODs while using slots 2 and 4 (going from CPU socket), immediate BSOD while using slots 1, 2, and 4, and (so far) no BSODs using slots 2, 3, and 4 with the same sticks as before. So either way, I can't populate all slots, either because a slot is dead or a RAM stick is dead. I have a matching set of 4 8GB GSkill Aegis 3000MHz sticks. I'd much rather have 24GB than 16GB, obviously - but other posts are quite ambiguous on whether this is a good idea, generally advising against mixing and matching RAM (which I wouldn't be doing) or mentioning that it'd run in single channel. So my question would be if this would be a good idea stability wise, and whether it'd actually run in single channel. Buying a new motherboard isn't a very attractive option, as I'm running a Ryzen 1600 and wouldn't want to inveset in a dying platform - B550 doesn't support it afaik, and I might upgrade once GPU pricing becomes reasonable, so throwing money at a mobo that I might run for a year or so doesn't seem great. EDIT: Before anyone mentions it: I'm running the RAM at 2133MHz to make sure the BSODs weren't from the RAM being too fast. I know first gen Ryzen doesn't like that very much.