# GPU status
nvidia-smi
# Look for NVIDIA Xid errors from this boot
sudo journalctl -k -b | grep -Ei 'NVRM|Xid|nvidia'
# Previous boot — particularly important
sudo journalctl -k -b -1 | grep -Ei 'NVRM|Xid|nvidia|AER|PCIe|pcieport'
# PCIe errors generally
sudo journalctl -k | grep -Ei 'AER|PCIe Bus Error|pcieport'
# GPU PCIe information
nvidia-smi -q | grep -A20 -i 'PCI'
# Hardware location/link status
lspci -vv -s "$(lspci | awk '/NVIDIA/{print $1; exit}')" | grep -Ei 'LnkCap|LnkSta|AER'