AIsaAIsaStudio English 打开 AIsa Studio

shanselman 开源 gputaskman:可按进程查看显存占用与 LLM KV cache 分解的 GPU 任务管理器

原标题:shanselman/gputaskman: A GPU task manager for Windows, Linux, and WSL: per-process VRAM, GPU activity, and LLM model/KV-cache insights (Windows AMD and NVIDIA).

shanselman 发布开源工具 gputaskman,定位为面向 GPU 内部的任务管理器,覆盖 Windows、Linux、WSL 和 Apple Silicon 的 macOS,可显示按进程的显存占用、GPU 引擎活动以及 LLM 模型权重与 KV cache 的缓冲区分解。可通过 PyPI 用 uv 或 pipx 安装,Windows 另提供免 Python 的便携 ZIP;作者强调这是早期实验软件,分类与 GGUF 分解是估算,CPU-only、NPU-only 和云端负载不在范围内。

为什么值得看

README 明确界定了测量与估算的边界,适合想弄清显存被谁占用的本地 LLM 用户参考。

阅读原文 ↗

本页的摘要、推荐理由与解读出自 AIsa Studio 的 AI 风向智能体,原文版权归原作者所有。

原文

GitHub 上的新仓库 shanselman/gputaskman,创建 11 天,已有 3094 星(平均每天 278 星)。 A GPU task manager for Windows, Linux, and WSL: per-process VRAM, GPU activity, and LLM model/KV-cache insights (Windows AMD and NVIDIA). 语言:Python

README:

gputaskman

An htop for what's inside your GPU, not just how full it is.

Windows with NVIDIA/AMD telemetry and Intel WDDM activity, plus Linux and WSL support (NVIDIA everywhere, AMD and Intel on native Linux; see Linux and WSL) and macOS on Apple Silicon (see macOS). A terminal UI that connects per-process GPU memory and engine activity with loaded runtimes, model files, and LLM buffer breakdowns.

Task Manager says "15 GB used." This tries to answer: by what? Model weights? KV cache? Your browser? The desktop compositor? And why is so much GPU memory in system RAM?

Scope: local GPU workloads, not cloud services. Decision models, vision, speech, LLM inference, training/fine-tuning (including Unsloth), graphics, and video all belong here. Visibility varies: process memory and engine activity are measured; identifying individual models and their work requires supporting evidence. A PyTorch/CUDA DLL alone cannot identify a model or prove training, and a mapped file does not prove that model is resident on the GPU or being exercised. CPU-only, NPU-only, and cloud workloads are out of scope.

Live GPU overview with an LLM's weights, KV cache, and compute buffer breakdown

Early, experimental software. Device and process counters are measurements; classifications and GGUF-derived breakdowns are estimates. This does not dump VRAM or enumerate a game's textures.

Install from PyPI

gputaskman is on PyPI. No repository clone is needed:

uv tool install gputaskman
gputaskman

Or, with pipx:

pipx install gputaskman
gputaskman

Works on Windows, Linux/WSL, and macOS with Python 3.11+ and the supported GPU/driver configuration described below. Both installers keep the app in an isolated environment. The shorter gtm command is also available. Upgrade with uv tool upgrade gputaskman or pipx upgrade gputaskman. For Windows without Python, use the portable download below.

Portable Windows download

The Releases page is where tagged builds publish two portable ZIPs: gputaskman-VERSION-windows-x64.zip for Intel/AMD Windows PCs, and gputaskman-VERSION-windows-arm64.zip for native Windows ARM64. These are separate native builds, not a universal executable.

Extract the entire ZIP, keeping _internal beside gputaskman.exe. Open Windows Terminal or PowerShell in the extracted folder:

.\gputaskman.exe
.\gputaskman.exe --once
.\gputaskman.exe --json

No Python, uv, CUDA Toolkit, model download, or installation is needed. A supported GPU with a current Windows WDDM driver is required. Builds are currently unsigned; code signing and an installer are not part of this initial distribution. Each ZIP includes licenses and a .zip.sha256 companion file is published for integrity checks.

Clone and run

Use 64-bit Windows 10/11 with a current WDDM driver and a terminal (Windows Terminal is recommended). Live workload checks have included an RTX 4080 SUPER and an RTX Spark N1X on Windows ARM64. CUDA workloads are supported on ARM64 with a compatible NVIDIA driver and ARM64 CUDA runtime.

AMD live check: Radeon 860M Graphics on Windows 11 Enterprise build 26300, AMD driver 32.0.22050.1003. During desktop/browser activity, WDDM reported per-process VRAM and AMD ADL reported GPU utilization, clocks, and ASIC power; temperature, memory activity, and PCIe readings were unavailable. Vulkan/ROCm workloads were not tested on this machine.

Intel and Qualcomm adapters on Windows can use WDDM's per-process memory and engine counters. When those counters are available, device GPU activity is the busiest WDDM engine after combining its process counters. An Intel or Qualcomm adapter reporting at most 512 MiB of dedicated memory and more shared system memory is treated as integrated: per-process local usage is shown as Shared RAM, not VRAM, while device dedicated and shared counters remain separate. This is a DXGI-capacity heuristic. Qualcomm GPU activity uses the WDDM adapter counter; temperature, power, clocks, and memory activity are not currently available on Windows. Arc B580 behavior is covered by mocked counters but has not been validated on Intel hardware. When a Windows telemetry provider does not supply a version, gputaskman reads the selected adapter's DriverVersion from its exact PnP device's driver registry key. If that value is unavailable, the driver remains unknown.

Install Git for Windows and uv if needed. With WinGet:

winget install --id Git.Git --exact
winget install --id astral-sh.uv --exact

Open a new PowerShell window so those tools are on PATH, then:

git clone https://github.com/shanselman/gputaskman.git
cd gputaskman
uv run --locked gputaskman

That's it. uv obtains the pinned Python version if needed, creates .venv, and installs the locked dependencies. The first run needs internet access. There is no CUDA Toolkit, Visual Studio, administrator session, API key, or model download required to monitor the GPU. NVML ships with the NVIDIA driver; AMD ADL is supplied by supported AMD Windows drivers.

Already have Python 3.11+ and prefer pip?

git clone https://github.com/shanselman/gputaskman.git
cd gputaskman
py -m venv .venv
.\.venv\Scripts\python.exe -m pip install .
.\.venv\Scripts\gputaskman.exe

No PowerShell activation script or execution-policy change is needed. The pip route resolves compatible dependencies; the uv route uses the checked-in lockfile.

If an alternate package index makes --locked report a stale lockfile on an unchanged checkout, uv run --frozen gputaskman uses the checked-in lockfile without re-resolving it. --frozen skips

原文版权归 GitHub 爆发仓库 及原作者所有,本站转载仅供阅读。如需删除,请联系 support@aisa.one。