Microsoft VP Scott Hanselman has launched gputaskman, an open-source GPU task manager for Windows, Linux, and macOS. The tool reveals per-process GPU usage, VRAM consumption, and provides detailed breakdowns of AI and LLM workloads including model weights and architecture.

Microsoft VP Launches Open-Source GPU Task Manager

Microsoft VP Scott Hanselman has released gputaskman, an open-source GPU task manager that addresses a critical gap in system monitoring tools

1

2

. Available for Windows, Linux, and macOS, this Terminal-based application delivers granular insights into GPU activity that Windows Task Manager cannot provide. Hanselman explained his motivation succinctly: "There's a lot going on in my GPU. I'm making a GPU Taskman to get insights to where my GPU memory is going."

1

Source: Tom's Hardware

Source: Tom's Hardware

Comprehensive Monitoring Beyond Traditional Task Managers

The tool displays per-process GPU usage, VRAM consumption, active runtimes, GPU utilization, clock speeds, temperatures, and page file usage in a three-section interface

1

2

. The top window features hardware monitors including VRAM usage charts. The middle section lists processes with their runtime details and memory footprint. The bottom window reveals detailed app information when users select a process. Screenshots from Hanselman's RTX 4080S system showed the desktop windows manager consuming the most VRAM, alongside AI and LLM workloads and standard Windows applications

1

. Live-streaming charts track GPU load and temperature in real-time, providing power users with the visibility they need to monitor GPU processes effectively.

AI-Focused Features Set gputaskman Apart

What distinguishes this open-source GPU task manager is its model inspector capability, specifically designed for AI and LLM workloads

1

2

. When llama-server.exe was selected in demonstration screenshots, the tool identified the Qwen3-0.6B model and broke down critical details including model weights, KV cache, parameter count, architecture, quantization, and layer count

1

2

. This level of detail proves invaluable as the computing industry increasingly processes data on GPUs rather than CPUs, particularly for users running multiple AI models simultaneously.

Immediate Availability and Future Implications

Hanselman released gputaskman version 0.3.0 under the permissive MIT license, making the source code and portable Windows downloads freely available on GitHub

1

. The developer describes it as "experimental software" but hints at broader implications. In social media comments, Hanselman suggested that expanded GPU monitoring capabilities could be integrated into Windows as standard

1

. If gputaskman gains traction, it could incentivize Microsoft to build more comprehensive monitoring systems in Windows Task Manager

2

. For now, users seeking to monitor GPU processes with unprecedented detail have access to a tool that fills a significant void in system monitoring capabilities.

Today's Top Stories

© 2026 TheOutpost.AI All rights reserved