My desktop system is locking up, presumably due to Mathematica. At least every time it has locked up, I have been using Mathematica. My gut tells me it is a graphics issue, but I can’t swear to that. Linux shouldn’t be locking up, regardless of what a user app is doing, but the user is part of the solution in Linux.
Are there any suggestions as to what to look for, or how to get the system to log a fault that locks it up? I haven’t been doing that sort of thing for over a decade.
If you have another computer from which you can login remotely, and the lockup is only for local I/O, you should be able to access current dmesg. The journal may have clues as well:sudo journalctl -b -1 on subsequent boot, and if using Xorg, /var/log/Xorg.0.log or Xorg.0.log.old. Please paste inxi -Faz input/output here using the </> icon, to show some possible other clues.
Any problem with telling which desktop?
Operating System: openSUSE Leap 15.6
KDE Plasma Version: 5.27.11
KDE Frameworks Version: 5.115.0
Qt Version: 5.15.12
Kernel Version: 6.4.0-150600.23.42-default (64-bit)
Graphics Platform: Wayland
Processors: 32 × Intel® Core™ i9-14900KS
Memory: 94.0 GiB of RAM
Graphics Processor: Mesa Intel® Graphics
Manufacturer: ASUS
I meant locked up! I tried. I’m using the on-chip intel gpu. I don’t know if a separate video card would make things better for what I do. And, in particular if it would prevent these lockups. It seems related to specific Mathematica notebooks. I have found the larger they get, the more problems they have.
Mathematica and MatLab: Limit resource usage (ram, swap, cpu, tasks) in user land with cgroups and systemd.
Translate from german to english with a translator like deepl.com.
Check access to this looked up machine with root/admin rights on console (<Ctrl>-<Alt>-<F1>) over network (SSH) or over serial cable (terminal).
@Ljosalfr Might need to look at using taskset since you have a cpu with efficiency and performance cores… If using the intel gpu is there an options to set the GPU memory allocation in the system BIOS?
Monitor Intel GPU usage with intel_gpu_top (rpm: intel-gpu-tools).
Mathematica could overload the slow Intel GPU with OpenCL parallel processing.