Extending the Python Ecosystem to Xilinx
summary
PYNQ extends Python to Xilinx SoCs by shipping a browser IDE and JupyterLab running on PYNQ that loads FPGA overlays as Python-accessible modules. The workflow produces hybrid Python+FPGA packages—pip-installable repos that include bitstreams, C drivers, and Python APIs. Demonstrations use a Pink Z2 board and an image-resizer overlay that measures about 3–4× speedup (≈4 ms) versus pure‑Python, showing the system is real, reproducible, and immediately usable.
AI-written summary · about
chapters (21)
- 0:00Intro
- 0:55Assumptions
- 1:27Outline
- 2:07Platform Evolution
- 4:13Example of a Raspberry Pi and FPGA Hat
- 5:01Field Programmable Gate Arrays (FPGAS)
- 6:58Zyng: Integrating Microprocessors, Microcontrollers and Programmable Logic
- 8:59Zynq in Embedded/Edge Applications
- 10:51PYNO Python for Embedded and Edge Applications
- 12:22Python for Embedded/Edge Systems
- 15:57Keys to the Success of Python's Package Ecosystem
- 17:36IPython Notebooks to Jupyter Notebooks
- 19:27Jupyter Notebooks to JupyterLab IDE
- 20:28Jupyter's Client-Server-Kernel Architecture
- 23:07PYNQ Wraps FPGA Circuit & Drivers in Python Package
- 26:35Installing Software and Hardware with PIP! Enabled by new hond packages
- 27:59Load the resizer Hybrid Package
- 29:13Image Resizer
- 32:06A Pythonic Framework for Programmable Platforms
- 34:31PYNQ Community
- 34:42Explore FPGA Design with Free Xilinx Vivado Tools
related talks
34mBuilding FPGA-based Machine Learning Accelerators in Python
19mPyha: Python overlay for OOP-VHDL
22mPyPy and the future of the Python ecosystem
22mPyPy hands-on
11mDoing Hard Things in a Soft Way
15mExtending Python Into the Future
19mPython as a language for testing and code generation
40mBuild Wireless Products Faster with Zephyr and MicroPython