{"slug": "nvidia-smi-cheat-sheet", "title": "nvidia-smi cheat sheet", "summary": "The **nvidia-smi** (NVIDIA System Management Interface) is a command-line tool for monitoring, managing, and diagnosing NVIDIA GPU devices, providing data on performance, temperature, utilization, power, and memory. It supports real-time monitoring with options like `-l` for periodic updates, `dmon` for device monitoring, and `--query-gpu` for exporting detailed metrics in CSV format for scripting and automation. The tool also allows administrative control, including setting power limits, locking clock speeds, enabling persistence mode, and terminating GPU processes.", "body_md": "nvidia-smi\n(NVIDIA System Management Interface) is a command-line tool that provides monitoring, management, and diagnostic information for NVIDIA GPU devices.\nIt communicates directly with the NVIDIA driver and GPU, and can:\n- Monitor GPU performance, temperature, and utilization\n- Manage power, clock speeds, and ECC\n- Control persistence mode and compute modes\n- Query detailed metrics for automation and monitoring\nnvidia-smi\nShows a summary table with:\n- GPU index, name, and UUID\n- Driver & CUDA versions\n- GPU & memory utilization\n- Power consumption and temperature\n- Active processes using the GPU\nnvidia-smi -l 5\nnvidia-smi -lms 500\nnvidia-smi --filename=/var/log/gpu.log -l 5\nnvidia-smi dmon\nExample:\n# gpu pwr temp sm mem enc dec mclk pclk\n0 85 64 23 5 0 0 405 1110\nnvidia-smi --query-gpu=index,name,uuid,temperature.gpu,utilization.gpu,memory.used,memory.total --format=csv\nOutput:\nindex, name, uuid, temperature.gpu, utilization.gpu [%], memory.used [MiB], memory.total [MiB]\n0, NVIDIA RTX A6000, GPU-02afcc1a-…, 58, 72 %, 13456 MiB, 49152 MiB\nnvidia-smi --query-gpu=memory.used,memory.total --format=csv\nnvidia-smi --query-gpu=temperature.gpu,power.draw --format=csv,noheader,nounits\nnvidia-smi --query-gpu=name --format=csv,noheader\nnvidia-smi pmon -c 1\nnvidia-smi pmon\nExample output:\n# gpu pid type sm mem enc dec command\n0 3024 C 23 5 0 0 python3\nTerminate a process:\nsudo kill -9 <pid>\nsudo nvidia-smi -pm 1\nsudo nvidia-smi -pm 0\nsudo nvidia-smi -pl 250\nsudo nvidia-smi --lock-gpu-clocks=900,1500\nsudo nvidia-smi --reset-gpu-clocks\nsudo nvidia-smi --lock-memory-clocks=405,1215\nsudo nvidia-smi -i 0 --gpu-reset\nnvidia-smi -q -d ECC\nsudo nvidia-smi -e 1\nExample:\nsudo nvidia-smi -c 3\nnvidia-smi -q -d SUPPORTED_CLOCKS\nnvidia-smi -q -d PERFORMANCE\nnvidia-smi -q -d PCI\nnvidia-smi -q -d FAN\nnvidia-smi --query-gpu=name,utilization.gpu --format=csv,noheader\nnvidia-smi --query-gpu=temperature.gpu --format=csv,nounits\nnvidia-smi --query-gpu=index,uuid,temperature.gpu,power.draw --format=csv -l 10 --filename=gpu_stats.csv\n- Use\n--query-gpu\nwith--format=csv,noheader,nounits\nin scripts. - Use GPU UUIDs for consistent identification.\n- Combine with\nwatch\n:watch -n 2 nvidia-smi\n#!/bin/bash\nLOGFILE=\"/var/log/nvidia_smi_monitor.csv\"\necho \"timestamp,gpu_index,uuid,utilization.gpu,memory.used,memory.total,temperature.gpu,power.draw\" > $LOGFILE\nwhile true; do\nnvidia-smi --query-gpu=timestamp,index,uuid,utilization.gpu,memory.used,memory.total,temperature.gpu,power.draw --format=csv,noheader >> $LOGFILE\nsleep 5\ndone", "url": "https://wpnews.pro/news/nvidia-smi-cheat-sheet", "canonical_source": "https://gist.github.com/omerfsen/8ecb620675525ac724a92bdf5a31a4b3", "published_at": "2025-11-06 11:58:31+00:00", "updated_at": "2026-05-22 20:07:34.850382+00:00", "lang": "en", "topics": ["developer-tools", "hardware"], "entities": ["NVIDIA", "nvidia-smi", "CUDA", "RTX A6000"], "alternates": {"html": "https://wpnews.pro/news/nvidia-smi-cheat-sheet", "markdown": "https://wpnews.pro/news/nvidia-smi-cheat-sheet.md", "text": "https://wpnews.pro/news/nvidia-smi-cheat-sheet.txt", "jsonld": "https://wpnews.pro/news/nvidia-smi-cheat-sheet.jsonld"}}