Nyuzi Processor

Nyuzi is an experimental GPGPU processor hardware design focused on compute intensive tasks. It is optimized for use cases like deep learning and image processing.

This project includes a synthesizable hardware design written in System Verilog, an instruction set emulator, an LLVM based C/C++ compiler, software libraries, and tests. It can be used to experiment with microarchitectural and instruction set design tradeoffs.

Documentation: https://github.com/jbush001/NyuziProcessor/wiki
License: Apache 2.0

The following instructions explain how to set up the Nyuzi development environment. This includes an emulator and cycle-accurate hardware simulator, which allow hardware and software development without an FPGA, as well as scripts and components to run on FPGA.

Install Prerequisites

Linux (Ubuntu)

This requires Ubuntu 16 (Xenial Xeres) or later to get the proper package versions. It should work for other distributions, but you will probably need to change some package names. From a terminal, execute the following:

sudo apt-get -y install autoconf cmake make ninja gcc g++ bison flex python \
    python3 perl emacs openjdk-8-jdk swig zlib1g-dev python-dev \
    libxml2-dev libedit-dev libncurses5-dev libsdl2-dev gtkwave python3-pip
pip3 install pillow

Emacs is used for verilog-mode AUTO macros. The makefile executes this operation in batch mode

MacOS

These instruction assume OSX Mavericks or later.

Install XCode from the AppStore application (Click Here). Then install the command line compiler tools by opening Terminal and typing the following:

xcode-select --install

Install python libraries:

pip3 install pillow

Install Homebrew from https://brew.sh/, then use it to install the remaining packages from the terminal (do not use sudo):

brew install cmake bison swig sdl2 emacs ninja

Alternatively, you could use MacPorts if that is installed on your system, but you will need to change some of the package names

You may optionally install GTKWave for analyzing waveform files.

Windows

I have not tested this on Windows. Many of the libraries are cross platform, so it should be possible to port it. But the easiest route is probably to run Linux under a virtual machine like VirtualBox.

Build (Linux & MacOS)

The following script will download and install the Nyuzi toolchain and Verilator Verilog simulator. Although some Linux package managers have Verilator, they have old versions. It will ask for your root password a few times to install stuff.

./scripts/setup_tools.sh

Build everything else:

cmake .
make

Run tests:

make tests

If you are on a Linux distribution that defaults to python3, you may run into build problems with the compiler. In tools/NyuziToolchain/tools/CMakeLists.txt, comment out the following line:

add_llvm_external_project(lldb)

Occasionally a change will require a new version of the compiler. To rebuild:

git submodule update
cd tools/NyuziToolchain/build
make
sudo make install

Next Steps

Sample applications are available in software/apps. You can run these in the emulator by switching to the build directory and using the run_emulator script (some need 3rd party data files, details are in the READMEs in those directories).

For example, this will render a 3D model in the emulator:

cd software/apps/sceneview
./run_emulator

To run on FPGA, see instructions in hardware/fpga/de2-115/README.md