The 100% CPU usage is normal and intended, we are trying to fully utilize your available hardware. This should not cause crashes by itself unless there is some configuration or hardware issue with your system. This should never cause your PC to turn off if your cooling is working correctly.
If you'd really like to use fewer cores then you can try reducing the number of worker threads at Window: Preferences: General Performance. This really shouldn't be necessary though, that's more of a work-around. If you just want to keep your system responsive then try reducing the hrc.exe priority in the task manager (or launching hrc via "start /low hrc.exe"), that's a much better way to achieve this.
The crash logs still show a 4GB limit (-Xmx4GB), you mentioned in your post that you tried to change this. It wasn't applied correctly when those crashes happened, but this wouldn't cause your PC to turn off. JVM crashes often indicate some hardware issue. Memory issues are not that uncommon with Ryzens and can often be fixed by modifying the RAM Bios settings. If your system turns off from overload then there may be an issue with your cooling. (CPU fan not working? CPU sink not fitting correctly? If the crashes happen after some time, maybe insufficiently ventilated case?)
After checking your cooling you can try to perform a stress test on your system:
https://mywindowshub.com/how-to-use-...ress-test-cpu/
The mprime95 stress test run shouldn't produce any errors, if it runs fine for ~15-30min this will rule out most issues. Again, make sure to check your fans/cooling before stress testing and monitor the temperature during the test.
Maybe try disabling PBO in bios, that should help if your system isn't sufficiently cooled. If you need additional help then please let me know the CPU type, CPU cooler, RAM type and any manual overclocking you may have done.
Last edited by plexiq; 08-17-2021 at 03:46 AM.