DSE with the Python API

LightningSim also features a Python API, which is great for automation and large-scale testing, such as for design space exploration (DSE). In this example, we will use a design where the latency is sensitive to FIFO buffer depths, and we will show how we can use LightningSim’s Python API to quickly perform DSE across a large range of FIFO depths.

The example project is pre-synthesized for your convenience, and we have provided a Jupyter notebook to walk through the steps.

Walkthrough

Step 1. We first need to install JupyterLab to open the provided notebook. Within the terminal where you activated the lightningsim environment, run:

conda install -y -c conda-forge jupyterlab

Step 2. Change directories into the example folder and start JupyterLab:

cd lightningsim-doc/examples/example-5
jupyter lab

This will automatically open your browser to JupyterLab.

Step 3. In the file list, double-click example-5.ipynb to open the notebook.

Follow the instructions in the notebook to learn how to use LightningSim’s Python API and run the DSE experiment.

Step 4. When you are finished, shut down JupyterLab by selecting File menu > Shut Down and confirming by clicking the Shut Down button.

Recap

In this example, you learned about LightningSim’s Python API, and how it can be used for advanced programmatic usage such as for performing design space exploration.