Running Jupyter on Qubernetes backend

Introduction

This document describes how to execute Jupyter notebooks on the Qubernetes backend. The Qubernetes backend is a Kubernetes cluster that is configured to run quantum circuits on GPU-backed simulators, e.g. Qiskit or PennyLane.

Prerequisites

You need the following tools to be able to run Jupyter with the Qubernetes backend:

• Docker

The following tools are optional but recommended:

• Make

Getting started

Python dependencies

To be able to run Jupyter notebooks on the Qubernetes backend, you need to configure your Python dependencies in the requirements.txt file. The template including in this repository contains the following dependencies:

qiskit
qiskit-aer-gpu

You can add more dependencies to this file if needed.

Build the Docker image

To build the Docker image that contains your dependencies, run the following command:

make build

You will need to adjust the username and the image name in the Makefile to match your Docker Hub username and the name of the image you want to create.

Start Jupyter

To start Jupyter, run the following command:

make run config=/path/to/kubeconfig

You will need to adjust the path to the kubeconfig file that you use to connect to the Qubernetes cluster.

Access Jupyter

To access Jupyter, open the URL that is displayed in the terminal. You will need to copy the token from the terminal to authenticate.