NVITOP
Where Power Meets Precision

NVITOP is a next-generation performance intelligence platform designed for those who demand complete visibility and
control. Built for speed, accuracy, and scale, NVITOP transforms complex system data into clear, actionable insights
helping teams operate smarter, faster, and at the top.

What Is Nvitop?

NVITOP is a powerful, real-time monitoring and performance management tool designed specifically for NVIDIA GPU environments. It provides deep visibility into GPU usage, memory consumption, processes, and system health, making it easier to track performance and detect bottlenecks efficiently.

Built for developers, data scientists, and system administrators, NVITOP offers an intuitive interface with precise metrics and live updates. It helps users optimize workloads, manage resources effectively, and maintain stable, high-performance GPU operations across local machines and server environments.

$ nvitop --monitor

Device 0: NVIDIA RTX 4090 Temp: 52°C

Fan: 35% Pwr: 125W/350W

Perf: P2 Mem: 8192/24576 MiB

Powerful Key Features

GPU Monitoring

Monitor GPU utilization, memory, and temperature instantly. Get live updates for accurate performance tracking. Detect issues before they affect operations.

Workload Management

View all active GPU processes in one place. Optimize workloads to maximize performance. Reduce system operation delays. Enhance task execution speed.

Metrics & Analytics

Access in-depth usage and performance metrics. Analyze historical trends to identify bottlenecks. Make data-driven decisions for system efficiency. Make data-driven decisions for system efficiency.

Customizable Dashboards

Design dashboards that match your workflow. Add widgets, charts, and alerts for key metrics. Visualize GPU performance clearly at a glance. Tailor views for teams.

Cluster Support

Monitor multiple GPUs across workstations or servers. Manage clusters seamlessly from a single interface. Scale operations without losing performance insight.

High Performance

Runs efficiently without slowing your system. Consumes minimal resources for continuous monitoring. Ensures reliable performance even under heavy loads.

Live Demo Preview

Experience NVITOP in action with our live demo. See real-time GPU monitoring, performance metrics, and workload
management, all in one interface

nvitop - Live GPU Monitor

Fan: 51%

Temp: 68°C

Power: 128W/350W

Memory: 14.3/24 GB

Usage:

Fan: 37%

Temp: 52°C

Power: 101W/320W

Memory: 5.8/16 GB

Usage:

PID	USER	GPU	MEM	%CPU	%MEM	TIME	COMMAND
12345	root	0	4.2G	15.2	17.5	02:34:56	python train.py
12346	user	0	2.1G	8.5	8.8	01:12:33	jupyter-lab
12347	root	1	3.8G	12.1	15.8	00:45:12	inference.py

Why Choose NVITOP?

NVITOP stands out as the ultimate solution for GPU monitoring and performance management. Its real-time insights, detailed metrics, and process-level visibility empower teams to optimize workloads, prevent bottlenecks, and maintain peak system efficiency effortlessly.

Designed for developers, data scientists, and IT professionals, NVITOP combines lightweight performance with scalability, allowing seamless monitoring across single devices or large GPU clusters. Its customizable dashboards and actionable analytics make decision-making faster, smarter, and more reliable ensuring you always stay ahead of performance challenges.

Real-Time Visibility: Instantly track GPU usage, memory, and processes for complete control.
Optimized Performance: Maximize resource efficiency and prevent system slowdowns.
Scalable & Flexible: Works seamlessly on single machines or large GPU clusters.
Actionable Insights: Get clear analytics to make informed decisions quickly.

Installation Guide

Learn how to quickly install NVITOP on your system, configure GPU settings, and start monitoring
performance efficiently and reliably.

$ pip install nvitop

$ pip install --upgrade nvitop

Conda nvitop Install Steps

$ conda config --add channels conda-forge

$ conda update nvitop

Install nvitop from Source – Responsive

$ git clone https://github.com/XuehaiPan/nvitop.git

$ pip install -e .

System Requirements

NVITOP runs on NVIDIA GPU-powered systems with minimal overhead, requiring modern drivers, Python support, and sufficient memory for seamless, high-performance real-time monitoring.

Operating System

Python Version

NVIDIA Requirements

Dependencies

Command Reference

The NVITOP Command Reference provides a complete guide to all commands and options, enabling users to monitor, manage, and optimize NVIDIA GPU performance efficiently with precision and control.

Basic

nvitop Commands

Command	Description
nvitop	Launch the interactive GPU monitor 📋
nvitop -m	Start in monitor mode (read-only) 📋
nvitop -1	One-shot mode, print and exit 📋

Display

Command	Description
nvitop --colorful	Enable colored output 📋
nvitop --light	Use light color scheme 📋
nvitop -o compact	Use compact display mode 📋

Sorting

Command	Description
nvitop -s memory	Sort processes by memory usage 📋
nvitop -s sm	Sort by SM (streaming multiprocessor) usage 📋

GPU Metrics Explained

Nvitop shows GPU usage details like memory, temperature, and running processes, helping you understand how busy and healthy your GPU is at a glance.

GPU Usage

GPU usage shows how much of the GPU’s computing power is currently being used. High usage means the GPU is busy processing tasks like training models, rendering graphics, or running heavy computations.

Memory Usage

Memory usage explains how much GPU memory is occupied by applications. It helps you know whether programs fit in memory or if errors happen due to insufficient GPU memory availability.

Temperature

Temperature tells how hot the GPU is while working. Monitoring it helps prevent overheating, which can slow performance or damage hardware if the GPU stays too hot for long periods.

Power Consumption

Power consumption shows how much electrical power the GPU is using. This metric helps understand energy efficiency and whether the GPU is running at low, normal, maximum power levels.

Running Processes

Running processes list all applications currently using the GPU. It shows process IDs, users, and memory usage, making it easier to identify which programs are consuming GPU resources.

Fan Speed

Fan speed indicates how fast the GPU cooling fan is spinning. Higher fan speed usually means the GPU is hotter and working harder, helping keep the temperature within safe limits.

Community & Contributions

Nvitop is supported by a friendly community. People share ideas, report problems, improve features, and help each other make the tool better together.

5.2K+

Stars

30+

Contributors

280+

Forks

500+

Commits

How to Contribute

Understand the Project

Read Nvitop documentation, learn project goals, and review existing issues carefully first.

Set Up and Develop

Fork the repository, create a feature branch, follow standards, and write clean commits.

Test Your Changes

Test updates locally, add useful tests, and ensure compatibility across environments.

Submit and Collaborate

Open a pull request, explain changes clearly, address feedback and collaborate.

Frequently Asked Questions

What is Nvitop?

Nvitop is a terminal-based tool used to monitor NVIDIA GPU usage. It provides real-time details about GPU performance and running processes.

Why is Nvitop popular among developers?

It offers a clean, interactive interface for GPU monitoring. Developers prefer it over basic tools for its detailed insights.

Is Nvitop open source?

Yes, Nvitop is an open-source project hosted on GitHub. Anyone can use, modify, and contribute to its development.

Which programming language is Nvitop written in?

Nvitop is written entirely in Python. This makes it easy to extend and customize.

What platforms does Nvitop support?

Nvitop mainly supports Linux and Windows systems. It works wherever NVIDIA drivers and NVML are available.

How does Nvitop differ from nvidia-smi?

Nvitop provides a continuously updating and interactive terminal interface. Unlike nvidia-smi, it allows sorting, filtering, and managing GPU processes in real time with better visual clarity.

Can Nvitop monitor multiple GPUs?

Yes, Nvitop can monitor multiple NVIDIA GPUs simultaneously. It displays individual GPU utilization, temperature, memory usage, and running processes, making it suitable for multi-GPU systems and servers.

Does Nvitop show GPU memory usage?

Nvitop shows detailed GPU memory usage for each process. Users can easily identify which applications are consuming excessive memory and take action to optimize or terminate those processes.

Can users sort or filter processes in Nvitop?

Nvitop allows users to sort and filter GPU processes by memory usage, process ID, or user. This feature helps quickly locate specific tasks, especially on shared or high-load systems.

How do you exit Nvitop safely?

To exit Nvitop safely, users simply press the q key. This closes the monitoring interface gracefully without interrupting running GPU processes or affecting system performance.

How can Nvitop be installed?

Nvitop can be installed easily using Python’s package manager pip. By running a simple installation command, users can quickly set up the tool on supported systems.

Does Nvitop require NVIDIA drivers?

Yes, Nvitop requires properly installed NVIDIA GPU drivers. It uses NVIDIA’s NVML library to collect real-time GPU statistics, which are unavailable without official driver support.

Can Nvitop run inside Docker containers?

Nvitop can run inside Docker containers when NVIDIA Container Toolkit is enabled. This allows GPU monitoring within containerized environments commonly used in machine learning and cloud-based workflows.

Is Nvitop lightweight?

Nvitop is considered lightweight compared to graphical monitoring tools. It uses minimal CPU and memory resources while still providing detailed and continuously updated GPU performance information.

Does Nvitop support remote monitoring?

Yes, Nvitop supports remote monitoring through SSH connections. Users can monitor GPU usage on remote servers, clusters, or cloud machines directly from their local terminal.

Can Nvitop terminate GPU processes?

Nvitop allows users to send termination signals to GPU processes. This is useful for stopping unresponsive or unnecessary jobs that consume GPU resources on shared systems.

What is nvisel in Nvitop?

Nvisel is a command-line utility included with Nvitop. It helps users select and manage CUDA-visible devices, making GPU assignment easier in multi-GPU environments.

Is Nvitop suitable for machine learning workloads?

Yes, Nvitop is widely used in machine learning and AI workflows. It helps developers track GPU usage during training, debugging, and inference to ensure efficient resource utilization.

Are there alternatives to Nvitop?

Several alternatives to Nvitop exist, such as nvtop and gpustat. However, Nvitop is often preferred due to its interactive interface, detailed metrics, and advanced process management features.

How can users contribute to Nvitop?

Users can contribute to Nvitop by reporting issues, improving documentation, or submitting pull requests. The open-source community actively welcomes new features, bug fixes, and collaborative development efforts.

NVITOP for NVIDIA GPUs – Monitor AI & ML GPU Usage Live

Monitor NVIDIA GPUs in real time with NVITOP. View GPU processes, memory usage, utilization, and performance on Linux and Python systems.

Price: Free

Price Currency: $

Operating System: Windows

Application Category: Software

Editor's Rating:
4.7

NVITOP Where Power Meets Precision