nvidia-smi cheat sheet
# Basic
nvidia-smi # Summary
nvidia-smi -L # List GPUs + UUIDs
nvidia-smi -q # Full query
nvidia-smi -q -d MEMORY,TEMPERATURE,POWER # Filtered
# CSV monitoring
nvidia-smi --query-gpu=name,utilization.gpu,memory.used,temperature.gpu \
--format=csv,noheader -l 1
# Topology
nvidia-smi topo -m # GPU topology
nvidia-smi nvlink -s # NVLink status
nvidia-smi nvlink -e # NVLink errors
# MIG
nvidia-smi -i 0 -mig 1 # Enable MIG on GPU 0
nvidia-smi mig -lgi # List GPU instances
nvidia-smi mig -lci # List compute instances
sudo nvidia-smi mig -cgi 19,19,19,19,19,19,19 # 7× 1g.5gb
sudo nvidia-smi mig -cci # Create compute instances
sudo nvidia-smi mig -dci && nvidia-smi mig -dgi # Delete all
# Device modification
sudo nvidia-smi -pm 1 # Persistence mode on
sudo nvidia-smi -i 0 -pl 250 # Power limit 250 W
sudo nvidia-smi --gpu-reset # Reset all GPUs