Spent months thinking my Ryzen #5900x was broken due to reboots and #mce error logs. Turns out it was my #gpu driver.

Needlessly increased the voltage curve incrementally for nothing. Spent multiple days debugging.

Good to know #mce errors can't necessarily be trusted without further investigation and analysis.

I can now reliably trigger the #mce errors on startup by running some broken #hip workload on my #5700xt

Modern processors with their increasingly smaller features really aren't build to last, they suffer from electron migration requiring increasingly larger core voltages over time. My #5900X is less then 4 years old and already crashes without offsetting the core voltage even with #PBO off.

This will require to rethink server maintenance for projects with long turnover.

Kinda wild to get a software / microcode update for a 4/5 year old CPU that reduces the minimal idle frequency from 2200 MHz down to 550.

haven't managed to measure reduction in idle power consumption but I suspect its major as these chips tended to use 35/40 watts previously.

Thanks #AMD #5900X