How lolzio5/jabbal works
Project type and purpose
• This is a hardware-accelerated full-stack system for simulating Conway’s Game of Life at very high frame rates.
• It combines three domains:
• Software UI on a laptop (hand-gesture detection and control)
• A PYNQ-Z1 board running Python/Jupyter to orchestrate the FPGA overlay
• FPGA logic (written in SystemVerilog) to compute next generations and drive an HDMI-out display.
File structure
• Top level:
• UI/hand_detection → Python scripts and Keras model for detecting gestures.
• Notebooks → PYNQ Jupyter notebooks to load overlays and send commands to FPGA.
• Generated bit and hwh files → prebuilt FPGA bitstream and hardware description for PYNQ.
• SystemVerilog & VerilogSource → hardware source code for next-state logic, line buffers, etc.
• PythonGameofLife & C++GameOfLife → CPU-only reference implementations (fallbacks/benchmarks).
• Testbenches → verification suites for the SystemVerilog modules (development artifacts).
README for insights
• Describes a client-server interaction:
• PYNQ loads the bitstream into the FPGA, writes the initial pattern, steps the hardware accelerator, and streams video frames over HDMI.
• Hardware toolchain: Vivado → generates .bit/.hwh → loaded by PYNQ Python API.
Detail
• Every major block: client app, network link, server, overlay loader, memory, compute pipeline, video out.
• Don’t over-abstract—show separate modules in the FPGA overlay (line_buffer, parallel_next_state, single_next_state, pixel_generator).
• Configuration flows (Vivado → .bit/.hwh → PYNQ API).