HLS & hls4ml Tutorial

• Original files can be downloaded from Xilinx (after agreeing to license) here: ug871-design-files.zip
• Xilinx Vivado HLS Guide: ug871-vivado-high-level-synthesis-tutorial.pdf

Software Requirements

• Vivado/Vitis Design Suite 2017.1 or later release (and source /PATH/TO/Vivado/20XX.X/settings64.sh with correct path/version)
• Miniconda (can be downloaded from https://docs.conda.io/en/latest/miniconda.html)
• Once Miniconda is installed and setup, create the conda environment with hls4ml and qkeras:

conda env create -f environment.yml

• The setup to make a Jupyter kernel with the Xilinx libraries is (assuming Vivado is already set up / sourced)

python setup_kernel.py

Labs

1. Introduction to HLS Lab 1
2. Introduction to HLS Lab 2
3. Pipelining Lab 1
4. Pipelining Lab 2
5. hls4ml Lab
6. hls4ml+QKeras Lab

Other Resources

• Parallel Programming for FPGAs (HLS Book): http://kastner.ucsd.edu/hlsbook/
• HLS Book Labs: https://pp4fpgas.readthedocs.io/en/latest/index.html
• UCSD CSE 237C: http://kastner.ucsd.edu/ryan/cse237c/
• hls4ml tutorial at Fast ML workshop: https://github.com/FPGA4HEP/course_material/tree/fml
• QKeras in hls4ml: https://indico.cern.ch/event/919468/contributions/3863588/attachments/2039929/3416087/sps_qkeras.pdf
• QKeras and hls4ml API functionality notebook: https://gist.github.com/thesps/c7e59cf5597804693b8663ddc9496b64
• QKeras and hls4ml API PR: fastmachinelearning/hls4ml#195
• "Official" hls4ml tutorial: https://github.com/hls-fpga-machine-learning/hls4ml-tutorial