You are not logged in.
Had no issue on Ubuntu, same underclock
Here is a coredump, using a RX 580 and amdgpu drivers
Offline
Ensure your microcode is installed and applied early
Offline
Ensure your microcode is installed and applied early
Microcode should refer to cpu, not to gpu if i am not wrong. Am i wrong?
Offline
A broken CPU microcode can have far reaching effects. Also which kernel was used on ubuntu and is used now?
Offline
A broken CPU microcode can have far reaching effects. Also which kernel was used on ubuntu and is used now?
From the link you posted, having amd-ucode installed should be enough for GRUB
Not sure about the Ubuntu kernel, months have passed, i have latest Linux but tested also LTS
One thing i remember for sure is that i was using amdgpu-pro instead of amdgpu drivers
Offline
Offline
Do things run fine if you do not tweak anything and use your card's default voltages+clocks? If the problem only shows up with your tweaks, can you experiment a bit with less aggressive settings? Can you find ones that work stable, just like when using default settings? Or do the crashes start to happen whenever you do any type of tweak, no matter how modest the tweak is?
Maybe the undervolt/overclock settings you used previously weren't 100% stable. A different Mesa and kernel version might now have caused the problem to expose itself. I saw this kind of thing happening more than once over the years with both Nvidia and AMD. I had settings that I thought ran fine but then a different driver version or a certain new game showed that my tweaks had actually been too aggressive.
Offline
Do things run fine if you do not tweak anything and use your card's default voltages+clocks? If the problem only shows up with your tweaks, can you experiment a bit with less aggressive settings? Can you find ones that work stable, just like when using default settings? Or do the crashes start to happen whenever you do any type of tweak, no matter how modest the tweak is?
Maybe the undervolt/overclock settings you used previously weren't 100% stable. A different Mesa and kernel version might now have caused the problem to expose itself. I saw this kind of thing happening more than once over the years with both Nvidia and AMD. I had settings that I thought ran fine but then a different driver version or a certain new game showed that my tweaks had actually been too aggressive.
I might, yes, but i have been running the same settings for a year on Ubuntu with no issue, same setting for a few weeks on Manjaro with no issue, and i game daily on Windows with the same settings, so i doubt that's the issue
I noticed sometimes even at default speed, but saving a profile, causes a shutdown
Offline
Found this in the dump
Core was generated by `radeon-profile'.
Program terminated with signal SIGSEGV, Segmentation fault.
#0 0x00007fe3c7ffc03e in ?? () from /usr/lib/libQt5Gui.so.5
Offline