Hi, I’ve been having this problem with my system for like a week where it just locks up and has to be hard restarted. The screen freezes, input devices are unresponsive, bluetooth disconnects, sftp server stops etc. This can happen anytime; while I’m moving the mouse, in a game, while idling… And the worst part is that when I check the journal afterwards there’s absolutely nothing there. No errors, no coredumps, no other messages (from the freeze time)
I’ve run the aur malware checker and it’s clean. I’ve also updated the system a couple times and it feels like the frequency of freezes only increased over time.
That makes me suspicious of a hardware failure, but I’ve had similarly strange problems before and every time it was a software problem. I’ve started a memtest (clean so far) but I’m not sure what else to do on that front. What component’s failure would result in a freeze, but not a crash or any errors?
System is zen4 with nvidia graphics.
Anyone experiencing something similar? Are there new driver / system incompatibilities that I should know about? How would you debug this? Is there a way to figure out the origin of the freeze despite the empty journal?
Edit: Memtest passed 3 times. Now it’s burned into the lcd screen 🥲😭 Edit: burn-in gone after color cycling
Edit 2: I think the problem is the bios settings after all. I had a -6 curve optimizer offset, and the freezing hasn’t happened since I removed it. Silicon degrades over time, apparently.
Edit 3: Nope it froze again
Are you running low on memory? The system can appear to freeze when it runs low on memory as the kernel tries very hard to swap pages between ram and swap constantly.
Nope, plenty of memory and more than half empty too. But what you say does make sense to me, the last couple weeks I’ve also been getting momentary freezes when starting local LLMs, as if the scheduler was prioritizing transferring it to the GPU above all else.
Try reseating all your components, like ram, cpu, etc. maybe new thermal paste too.
Any way to keep dmesg running someplace you can see it? Do the text virtual consoles work? (Ctrl-alt-f3, etc). Do you have any network disks mounted over nfs?
I can run it on my second monitor, I’ll see what it says when it crashes the next time. Virtual consoles don’t work. I don’t have nfs, but I do have sftp and it stops working when it freezes.
Best of luck! If it’s a desktop you can try re-seating everything. I’ve had issues a times with both ram and GPU not being in exactly right.
Also one time I was using my motherboards auto-overclock features and had hard crashes, so you can see if those are enabled.
no clue if my scenario is anyhow close to your situation but what sounds familiar is the freeze part
while for me it was for a second or so, from time to time, it got really annoying
all started after nvidia drivers got pushed to AUR — so if your GPU is newer than mine, this is where my assumption is wrong…
checked and did everything I could, memory, drives, psu, danced naked under the full moon…
so I switched to CachyOS with guesstimation that the person maintaining the drivers is someone from the Cachy team, they even have their own repo for them… — following that, jumped from my ~3 yo Arch install mostly just having hopes
and everything is fine for now…
so, I don’t want to push you away from the distro of your choosing but, personally, I already learned for myself: until I have anything nvidia in my setup, I have more work to do and more things to check / worry about, always
The gpu does support the new drivers, but it is possible that the new drivers are what’s buggy.
from all my basic investigations back then, it might be mostly about the part of their drivers that manages power delivery and consumption / power states and so on
it might not help you at all but if you have a chance to run your setup temporarily on Cachy — I would give it a shot — there are some kernel level optimizations and tweaks done there, maybe they figured some work around for that problematic part
Try clearing bios/uefi cmos, even if you didn’t touch its settings. They are buggy meases. Update it if possible
Updated and it didn’t happen for the rest of the day. Today it happened again though
Unfortunate. One other possible workaround may be this: https://gist.github.com/dlqqq/876d74d030f80dc899fc58a244b72df0
Btw, you have amd-ucode installed, right?
i do, and I’ll reinstall just in case. I’ll also try the solutions mentioned in the link. Your google-fu is amazing ty <3
Edit: Nope, no luck
I think it the bios settings after all. I had a -6 curve optimizer offset, and the freezing hasn’t happened since I removed it. Silicon degrades over time, apparently.
Yeah I mentioned resetting everything because I had many experiences on different hardware where that helped. Glad it worked for you.
Nevermind it froze again today. Concerning…
Have you tried our lord and savior
sudo pacman -Syu?Of course
Can you spec your components cpu, men, graphic, motherboard, disk and if you have any cards insert in computer!
Aswell what do you have connections on USB pan so on…
I recognize the symptoms…
Nvm it froze again.
It’s a 7950x with x670 mobo, 64 gb ddr5 6000, 4070ti and no other cards. I have keyboard, mouse, controller and bluetooth adapters in the usb ports
I think this has been solved. It was the bios settings. I had a -6 curve optimizer offset, and the freezing hasn’t happened since I removed it.

