sharm294 / sonar Goto Github PK

Service	Master	Dev
Build
Coverage
Quality
Documentation

sonar is a simulation/testbenching and project management infrastructure for Vivado projects.

This was developed during my masters and as I no longer use/need it, further work on sonar is not planned. While the current test cases pass (in my limited test setup [Ubuntu 16.04, Vivado 2017.2] at least), there are certainly bugs and incomplete features should you to deviate from the examples.

sonar can be imported into a Python script. Then, the user can define the ports of the device-under-test (DUT) and test vectors. It will generate a .sv testbench and an associated .dat file containing the user specified test vectors. It can also optionally generate a C++ data file and testbench for use with HLS though that is not quite as sophisticated.

All generated files are placed in the specified directory. To simulate the SV file, add the TB, the .dat file and the DUT file(s) to the simulator of your choice. To simulate the C++ file, make a testbench executable using the generated TB file and run it.

sonar's project management and associated scripts show an example workflow where all these files are automatically added to the created Vivado projects.

At present, sonar can create directories for repositories, and within these directories, create IPs through its command line interface. For IPs, a skeleton directory structure is created for the user to fill in as needed. It provides a set of scripts and a Makefile to compile HLS IPs, create sonar testbenches, and pull them all into Vivado projects. The goal with management is to provide a consistent feel to hardware projects, making them easier to navigate, build and share with others. The reach goal is to add hardware IP dependency information so sonar can go fetch dependencies from other repositories without duplicating code.

For now, this is experimental and used primarily to run the test suite.

To install, clone this repository, go to the cloned directory and run:

$ pip install .

# start a new shell or login again

# initialize with local Xilinx path
$ sonar init vivado /path/to/Xilinx/directory
# then use "sonar activate vivado_{version}" to change versions
$ sonar activate vivado_2017.2
# for now, this should print "sourced file" if successful

The best example for usage can be found in docs/projects/hello_world. This example shows a sample C++ source code and an associated testbench (hello_world.py). This testbench file shows the syntax and some of the features that the sonar testbenches can have.

# to use the testbench, import sonar in python e.g.
from sonar.testbench import Testbench, Module, TestVector, Thread

The test_hello_world.sh script in tests/shell is part of the sonar's internal testing. It shows an example of sonar's command line interface and the commands used to create a new repository, add an IP to it and simulate it.

Installing the package is sufficient. It is recommended to install and setup argcomplete to autocomplete sonar's CLI commands. If the package exists, sonar will use it.

pytest and coverage is used for internal testing.

For development, it is HIGHLY recommended to use a virtual env such as conda or docker. sonar uses the pre-commit package to enforce style checks for every commit. Conda instructions to set up the development environment are below:

# install sonar as editable
$ pip install -e .

# install optional but RECOMMENDED packages
conda install argcomplete

# install testing dependencies
$ conda install pytest coverage

# install pre-commit and pre-commit hooks
$ conda install pylint
$ conda install -c conda-forge pre-commit cpplint cppcheck shellcheck
$ conda install -c sarcasm clang-format

# activate argcomplete globally for your user if it's not otherwise activated
# note: make sure user bash completion scripts are picked up by .bashrc!
$ activate-global-python-argcomplete --user

# install pre-commit if not installed for this repository
$ pre-commit install