NVMT Pro - Industrial Hardware Dashboard

NVMT Pro is a high-performance NVIDIA GPU management, CPU management and monitoring ecosystem designed for industrial-scale hardware infrastructure. Written in Rust, it combines low-level hardware interaction with a modern, high-fidelity web dashboard for real-time telemetry and remote orchestration.

Key Features

• Real-Time Telemetry: Instant visualization of GPU clocks, power consumption, thermal metrics, and VBIOS details.
• Hardware Control: Direct interface for core/memory offsets, power limits, and manual fan control.
• Integrated SSH Gateway: Native, high-performance terminal (xterm.js) built directly into the dashboard for immediate remote shell access.
• Multi-Node Architecture: Orchestrate and monitor entire clusters from a single, unified interface.
• Linux Screen Integration: Live monitoring and interaction with background screen sessions.
• Script Orchestration: Powerful batch execution interface with a library for automated deployment and management.
• Premium Industrial UI: Modern dark-mode aesthetic with glassmorphism, JetBrains Mono typography, and high-frequency updates.

Technology Stack

• Backend: Rust (Axum, ssh-rs, nvml-wrapper, tokio).
• Frontend: Vanilla JavaScript, Tailwind CSS, Lucide Icons, xterm.js.
• Performance: Throttled rendering loops and efficient WebSocket data streaming ensure a responsive experience even with high metric frequency.

Getting Started

Prerequisites

• NVIDIA GPU with compatible drivers.
• Rust Toolchain (Cargo).

Installation & Build

Download and run form release

Running the Dashboard

To start the management interface and enable remote access:

./target/release/nvmt --serve --port 10090

The dashboard will be available at http://localhost:10090.

Configuration

NVMT Pro uses a config.json file for node discovery and remote credentials:

{
  "nodes": [
    {
      "name": "Node-01",
      "host": "192.168.1.10",
      "port": 10090,
      "ssh_user": "root"
    }
  ]
}

---

Built by Makertronic.