InfiniteGPU: Revolutionizing AI Workloads with Compute Resource Sharing

InfiniteGPU is a robust platform designed to streamline the exchange of compute resources for artificial intelligence workloads. By allowing requestors to offload heavy AI inference tasks to a connected network of providers, this platform effectively transforms how computing power is accessed and utilized. Providers can monetize their idle computational capacity across NPUs, GPUs, and CPUs through an intuitive web interface and a dedicated desktop application.

Project Goals

The primary aim of InfiniteGPU is to create a frictionless marketplace whereby:

• Requestors can execute AI inference tasks without the need for expensive infrastructure.
• Providers can earn passive income by sharing the compute resources of their devices.
• The platform manages task distribution, execution monitoring, and automated payment processes seamlessly.

Key Features

• 🧠 ONNX Model Execution: Easily run AI inference tasks with industry-standard ONNX models.
• 𓇲 Neural Processing Units: Leverage NPUs to accelerate AI inference.
• 📁 Multiple Input/Output Formats: Support for plain text, images, videos, and numpy tensors.
• 💰 Automated Payments: Integrated Stripe payment processing ensures centralized, automated transactions.
• ⚡ Real-time Updates: Utilize SignalR for instant task status and progress updates.
• 🖥️ Native Desktop Client: A WinUI 3 application provides seamless background execution of compute tasks.
• 🔐 Secure Authentication: Implement JWT-based authentication with comprehensive user management features.
• 📊 Financial Dashboard: Monitor earnings, settlements, and payment history easily.
• 🎨 Modern UI: Enjoy a streamlined, responsive interface built with React and TailwindCSS.

Architecture Overview

Backend

• Implementation: Built on ASP.NET Core 8.0, utilizing clean architecture patterns.
• Key Technologies:
  • CQRS Pattern with MediatR for command/query separation.
  • Entity Framework Core for data management and SQL Server for persistence.
  • ASP.NET Identity for secure user management.
  • SignalR Hubs for real-time communication.
  • FluentValidation for reliable input validation.
  • Azure Blob Storage for storing task data and models.
  • Stripe API for handling payment processes.

Frontend

• Framework: Developed in React 19, ensuring a modern user experience.
• Key Tools:
  • Vite for development and optimized production builds.
  • TailwindCSS for efficient, utility-first styling.
  • TanStack Query for effective asynchronous state management.

Desktop Application

• Technology: Built using WinUI 3, allowing high-performance execution of AI tasks.
• Key Components:
  • ONNX Runtime and OpenCV Sharp for advanced AI task execution and image processing.

Application Flow

• Requestors can register, upload ONNX models, configure tasks, fund their wallets, and track task progress in real-time.
• Providers install the desktop application and earn credits by executing tasks, with the ability to withdraw funds once a set threshold is reached.

InfiniteGPU stands out by balancing powerful functionality with user-friendly design, making it a leading solution for those looking to share or access computing power for AI workloads efficiently.