Kaspa FPGA Miner

This project implements a hardware-accelerated Kaspa cryptocurrency miner using High-Level Synthesis (HLS). The miner is written in C++ and designed to be synthesized for FPGA platforms such as Xilinx Zynq-7000 using Vivado HLS. It includes a testbench for functional simulation, a Docker development environment, and HLS synthesizable modules for the HeavyHash algorithm.

---

Overview

Kaspa uses a unique proof-of-work algorithm known as HeavyHash. This project implements a parallelized and hardware-efficient version of that algorithm using:

• Custom deterministic PRNG (xoshiro256++)
• Lightweight BLAKE3 stub (for demonstration or replacement)
• Matrix-based hash mixing
• AXI-compatible FPGA interfacing

---

Architecture

---

Project Structure

File	Purpose
kaspa_miner.cpp	Top-level miner kernel for HLS
heavyhash_hls.cpp	Core HeavyHash implementation
blake3_hls.h	Placeholder BLAKE3 hash (replace with real hash for production)
xoshiro256pp.h	High-speed PRNG used for matrix generation
kaspa_miner.tcl	Vivado HLS build script
sim_ap_int.h	CPU-simulation version of HLS integer types
heavyhash_testbench.cpp	Testbench for validating functionality
Dockerfile / run.sh	Portable build environment setup

---

Architecture

Getting Started

Prerequisites

• Xilinx Vivado HLS 2022.1
• Docker (for portable development)
• A compatible FPGA board (e.g., Zynq-7000 series)

Build Instructions

# Launch development container
./run.sh

# Inside container:
vivado_hls -f build_kaspa_miner.tcl

---

Test

A basic testbench is provided:

g++ heavyhash_testbench.cpp -o test && ./test

Expected digest output should match known Kaspa hash outputs. For more robust validation, integrate a reference Kaspa node or hash oracle.

---

Notes

• This project includes a simplified BLAKE3 hash function for HLS testing. In production, integrate a verified BLAKE3 core or pre-process the input on CPU.
• Real nonce sweeping and mining requires persistent execution and a full networking stack.