bovulpes / tuto-qc

Material for hands-on session of the "Quantum Computing Days at IN2P3", December 2-3, 2019, IPNO, France

Material for the hands-on session of the "Quantum Computing Days at IN2P3", December 2-3, 2019, IPNO, France

Before coming to the workshop

In order to do the exercises of the hands-on session, you need a laptop (PC or Mac) with Linux on it. The software for the exercises (mainly QC simulators for quantum circuits) together with all the necessary dependencies will run inside a "docker container" prepared in this purpose. The code sources for the exercises can be obtained from the git repository:

https://github.com/bovulpes/tuto-qc.git

The first step is to install the Docker Engine on your machine, by choosing the Community Edition specific to the Linux flavour of your operating system (you need administrator privileges on your system). Instructions on how to install it can be found here:

for Ubuntu:

https://docs.docker.com/install/linux/docker-ce/ubuntu/

for CentOS:

https://docs.docker.com/install/linux/docker-ce/centos/

for Fedora:

https://docs.docker.com/install/linux/docker-ce/fedora/

for Debian:

https://docs.docker.com/install/linux/docker-ce/debian/

For MacOS it is a bit more complicated, since you have to create first a Docker ID (a sort of account), as you can see here:

https://hub.docker.com/editions/community/docker-ce-desktop-mac

The Docker ID could be useful for your if later on you want to create your own docker images and upload them on the docker hub in order to make them available to a large audience (they are public).

Once you have the Docker Engine CE installed, pull to your machine the image for the tutorial with the command:

docker pull bovulpes/qc-in2p3-tuto:v01

Create the working directory for this tutorial with the name "qctuto" somewhere in your favourite directory:

mkdir <your-favourite-directory>/qctuto

and define the environment variable:

export QCTUTO_HOME=<your-favourite-directory>/qctuto

The next command creates a container named "qctuto" starting from the pulled image "bovulpes/qc-in2p3-tuto:v01", sets the home directory to the value defined above and starts a "bash" shell:

docker run -itd --name qctuto -v $QCTUTO_HOME:$QCTUTO_HOME --net=host bovulpes/qc-in2p3-tuto:v01 /bin/bash

With the next command you add inside the container a user/group having the same identity as your local account:

docker exec qctuto useradd -u $UID $USER

Now you can enter the container in an interactive session:

docker exec -it --user $USER -w $QCTUTO_HOME qctuto /bin/bash

Note: if you get the error about missing option "-w", make sure that the API version of your docker installation is >= 1.35 by checking the output of the command "docker version". If it is not the case, update you docker to a higher version.

You can repeat the command in another terminal and open another session inside the container.

The support material is on git hub and you can get a copy of it with:

git clone https://github.com/bovulpes/tuto-qc

Then go in the directory:

cd tuto-qc/qpp-examples

and compile the example file by running:

make

If successful, try to execute it:

./minimal

and you should see this output:

Hello Quantum++!
This is the |0> state:
1
0

At the end of the work, close all terminal windows from the "qctuto" container and then kill the container with the command:

docker kill qctuto

The simulators

A comprehensive list of (free) simulators can be found here:

https://quantiki.org/wiki/list-qc-simulators

For this tutorial we will use:

Quantum++ (C++11)
https://github.com/vsoftco/qpp

Using the myQLM simulator from Atos Quantum Lab

Below are the instructions on how to download and use the myQLM docker image for the workshop.

The image once installed is around 2.5GB; its export file is about 950MB.

1 – Retrieve the image (equivalent of docker pull if our image was published on docker.io)

~> sftp [email protected]

[email protected]'s password: ...see mail...

Connected to doorway.bull.com.

sftp> cd downloads

sftp> get myqlm-runtime-0.0.5_latest.export

Fetching…

sftp> bye

Note: This can be done from any machines connected to Internet.

2 – Import the image into your docker installation

~> zcat myqlm-runtime-0.0.5_latest.export | docker import - myqlm-runtime-0.0.5:latest

Note: One can use “docker images” to see the list of images

3 – Instantiate a container from the image and connect to it

~> [sudo] docker run -ti -p 8011:8888 myqlm-runtime-0.0.5:latest /bin/bash -l

Notes:

8011 here is an arbitrary port on the host side that will be mapped to the port 8888 inside the container (Jupyter server default port).

From another window, one can use “docker ps” to see the container.

The myQLM wheels are already installed except for myqlm-simulators and are located at /var/myqlm/dist/0.0.5/atos/myqlm-0.0.5-atos/

There is a README file that explains how to install myQLM at /var/myqlm/dist/0.0.5/atos/README.md

4 – Start a Jupyter server once inside the container

~> su – qatuser
~> myqlm_notebook/launch_myqlm_notebook
…

Copy/paste this URL into your browser when you connect for the first time, to login with a token:

http://0.0.0.0:8888/?token=f508e5b2659e107d02b9821fb8a604e9f7dd960dd2fb16f1

Note: You can also work in command line mode

5 – Connect to the server from a browser

http://<localhost_or_host_ip>:8011

Note: Copy/paste the value of the token shown on the Jupyter console above