VISI

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

Quiz

A few quick questions based on this unit. Mark it complete when you are done.

Question 1 / 3

What is sudo nvidia-smi --gpu?