System Design on Zynq using SDSoC

After completing this workshop, you will be able to:

• Introduce the concept of “software-defined” systems on chip (SDSoC)
• Understand the capabilities and limitations of the SDSoC development environment
• Get hands-on experience creating application-specific systems on chip from C/C++ programs using the SDSoC
• Gain practical understanding of the SDSoC design flow
  • How the SDSoC compiler maps programs to HW/SW systems
    • Structure of generated hardware systems
    • Structure of the generated software
  • How to control the compilation and generation process
    • Modifying program source
    • Using #pragmas

Day 1:

• Zynq SoC architecture and Vivado IPI
• SDSoC tool overview
• Lab 1: Getting started with SDSoC design flow
  • Go through the process of using SDSoC to create a new project using available templates.
• Data motion networks
• Lab 2: Pragmas and data motion networks
  • Handling data movements between the software and hardware accelerators using various pragmas and SDSoC API.
• Coding Considerations
• Profiling
• Lab 3: Profiling application and create accelerators
  • Profiling an application, analyzing the results, identifying function(s) for hardware implementation.

Day 2:

• Estimation
• Lab 4: Estimating accelerator performance
  • Estimating the expected performance of an application when functions are targeted in hardware, without going through the entire build cycle.
• Debugging
• Lab 5: Debugging software application
  • Debugging software application targeting Standalone and Linux OS in SDSoC.
• Using C-callable libraries and multiple accelerators
• Improving performance with Vivado HLS
• Lab 6: Fine-tuning with Vivado HLS
  • Using various techniques and directives of Vivado HLS which can be used in SDSoC to improve design performance. 
• Creating SDSoC platform
• Lab 6: Creating and using platform for a custom application
  • Creating a custom platform for a matrix multiply application (2018x version)
  • Creating a custom platform for an audio application (2017x and prior versions)