User Guide

The graph L3 layer provides an asynchronous and easy-to-integrate framework. By using the Xilinx FPGA Resource Manager (XRM) and this completely original and innovative L3 asynchronous framework, users can easily call the L3 asynchronous APIs in their pure software (heterochronous or synchronous) codes without the need of considering any hardware related things. The innovative L3 asynchronous framework can automatically acquire current platform’s available resources (available FPGA boards, boards’ serie numbers, available compute units, kernel names, etc.). In addition, this framework seperates FPGA deployments with L3 API calling, so the graph L3 layer can be easily deployed in the Cloud and can be easily used by pure software developers.

• Getting Started
  • Software Requirements
  • Hardware Requirements
  • Environment Setup
  • Building the dynamic library
• Running Examples
  • Basic Flow
  • Example
  • Asynchronous Execution
    • Example of using multiple requests
• Louvain Paritition Demo
  • Linear Louvain Partition Flow
  • BFS Louvain Partition Flow
• Louvain Modularity Launch Demo
  • Launch u50 Flow
  • Launch u55c Flow