Assassyn: Asynchronous Semantics for Architectural Simulation & Synthesis

Assassyn is aimed at developing a new programming paradigm for hardware development. The ultimate goal is to unify the hardware modeling (simulation), implementation (RTL), and verfication.

---

TL;DR

If you already have Docker setup, you can use the Dockerfile in the root folder to quickly set up the development environment. Here’s how:

Build the Docker Image

# Build the Docker image
$ docker build -t assassyn-dev .

Run the Docker Container

# Run the container and mount your local project directory into the container
$ docker run -it --rm -v /path/to/your/project:/app assassyn-dev

Note: Replace /path/to/your/project with the path to your local Assassyn project folder.
For example, if your project is located at /home/user/assassyn, the command will be:

docker run -it --rm -v /home/user/assassyn:/app assassyn-dev

Inside the container, the project will be accessible at /app. You can run commands like python or make as needed.

---

Getting Started

Users: The language backend is implemented in Rust, while for easy-to-use sake, the frontend is implemented in Python. You can initialize this package by running the command below.

# Install the required dpendences for Rust and Python >= 3.6
# Make sure you have: zsh, Rust toolchain, and pip
$ ./init.sh
$ source setup.sh # Add assassyn's python package to your PYTHONPATH

All the test cases are located in python/unit-tests. These cases are not only testing features, but also self-examplified to demonstrate the usage of each feature. Refer python/unit-tests/README.md for more details on the feature-case correspondence.

Advanced Users: If you would like to simulate the generated RTL, you can use verilator. To simply setup verilator, you can run the following command.

$ ./scripts/build-verilator.sh
$ source setup.sh # Re-run the script since it has a different environment variable for verilator

Developers: All the test cases are located in python/tests, you can just run them like all other Python scripts.

$ python python/unit-tests/test_driver.py 

Refer our developer doc for more details on tooling.

Why yet another RTL generator?

Designing circuits using RTL exposes excessive low-level details to users, from the behavioral semantics, to timing, and placement. Our insight of simplifying this programming paradigm is to designate a set of simple and synthesizable primitives while retaining enough expressiveness to program a massively concurrent programming system.

Refer language manual for more details.