Description
gputaskman is an advanced GPU task manager developed to offer granular control and visibility into graphics processing unit performance across various operating systems, including Windows, Linux, and Windows Subsystem for Linux (WSL). Its primary function is to monitor and display critical GPU metrics, enabling users to understand how their system's graphics hardware is being utilized.
Read More
The tool provides per-process VRAM (Video Random Access Memory) usage, allowing users to identify which applications are consuming the most GPU memory. This is particularly useful for optimizing resource allocation and troubleshooting performance bottlenecks in demanding applications like AI model training, 3D rendering, and gaming. Beyond VRAM, gputaskman tracks overall GPU activity, giving a real-time overview of the GPU's workload.
A standout feature of gputaskman is its specialized insight into Large Language Model (LLM) model and KV-cache usage. This capability is invaluable for AI developers and researchers who need to monitor the memory footprint and performance characteristics of their LLMs. By providing detailed information on how LLMs are utilizing GPU resources, gputaskman helps in optimizing model deployment and inference, ensuring efficient use of high-performance computing resources.
The project is hosted on GitHub, indicating its open-source nature and fostering community contributions. It supports both AMD and NVIDIA GPUs on Windows, ensuring broad compatibility for a wide range of hardware configurations. The tool's focus on detailed, per-process metrics and LLM-specific insights makes it a powerful utility for anyone involved in GPU-intensive tasks, from system administrators to AI engineers.
- CategoryMLOps & Model Deployment
- Pricing ModelOpen source
- DeploymentSelf-hosted
- PlatformsWindows, Linux
gputaskman's Core Features
Per-process VRAM monitoring
GPU activity tracking
LLM model insights
KV-cache insights
Support for Windows operating system
Support for Linux operating system
Support for Windows Subsystem for Linux (WSL)
Compatibility with AMD GPUs (Windows)
Compatibility with NVIDIA GPUs (Windows)
Open-source project hosted on GitHub
Regular releases and updates
Community contributions
Getting Started with gputaskman
Clone Repository: Use `git clone https://github.com/shanselman/gputaskman.git` to get the source code.
Install Dependencies: Follow the project's documentation to install necessary Python packages and system libraries.
Build/Compile: If required, compile the application or its components according to the instructions for your operating system.
Execute: Run the gputaskman application from your terminal to start monitoring GPU activity.
Monitor: Observe per-process VRAM, GPU activity, and LLM-specific metrics through the tool's interface.
gputaskman's Use Cases
- GPU Resource Monitoring
- LLM Performance Optimization
- Troubleshooting GPU Issues
- Developer Tooling
- Cross-Platform GPU Management





