Installation Instructions

Introduction

I have a Lenovo Legion Y520 Windows 10 laptop with Ubuntu 20.04 installed on a Samsung portable SSD T5 disk. I start my machine from this disk.

All installation instructions in this course are based on Ubuntu 20.04.

Install Julia

Julia is a language that is fast, dynamic, easy to use, and open source.

Prerequisites

Your computer OS is Ubuntu 20.04 64 bit.

Step	Action	Comment
1	Download Julia	Click on the link and find the row `Generic Linux Binaries for x86`. Click on `64-bit`. Download the file eg. julia-1.5.3-linux-x86_64.tar.gz
2	Ctrl+Alt-T	Open a terminal window.
2	$ mkdir julia	Create a folder
3	$ cd julia	Enter folder
4	$ mv ~/Downloads/julia-1.5.3-linux-x86_64.tar.gz .	Move the downloaded file to folder `julia`.
5	$ tar -zxvf julia-1.5.3-linux-x86_64.tar.gz	Extract the file.
6	$ cd julia/julia/julia-1.5.3/bin/
7	$ ./julia	Start Julia.
8	julia> Ctrl-D	Close Julia.
9	cd ../..
10	$ sudo ln -s /home/rob/julia/julia-1.5.3/bin/julia /usr/local/bin/julia	Create a symbolic link.
11	$ julia	Start Julia.
12	Ctrl-D	Close Julia.

Info

You can remove the symbolic link with: sudo rm /usr/local/bin/julia.

Install Git

Git is a free and open source distributed version control system designed to handle everything from small to very large projects with speed and efficiency.

Prerequisites

Your computer OS is Ubuntu 20.04 64 bit.
You have a GitHub account.

Step	Action	Comment
1	sudo apt install git
2	git config –global user.email "<your email address>"
3	git config –global user.name "<your first and last name>"	E.g. "Rob Bontekoe"
4	git config –global github.user "<yout git hub name>"	E.g. "rbontekoe"
5	cat .gitconfig	Show your git data.

Install Julia for VSCode

Juno will not be further expanded with new features. It looks like VSCode is going to be the future development platform for Julia. Watch the video JuliaCon 2020 | (Juno 1.0) VSCode for Julia 1.0 | Sebastian P., Shuhei K., David A..

Prerequisites

Your computer OS is Ubuntu 20.04 64 bit.
You have installed Julia 1.5+.
You have installed Git.

In this activity, you will:

Install VSCode.
Install the Julia Extension.
Optional, Show Inline Results.

Step 1: Install VSCode.

The installation of VSCode runs most smoothly if you install Julia and Git first.

Step	Action	Comment
1	Go to the VSCode site
2	Click on: Download	Select the Download button.
3	Right of .deb, select: 64 bit	Click on the 64 bit button.
4	Select: Save	Select the Download folder and click on Save.
5	Go to your Download folder
6	Type: sudo apt install ./<downloaded .deb file>	E.g. sudo apt install ./code1.50.0-1602051089amd64.deb. VSCode will be installed.
7	cd ~	Go to your home folder.
8	Type: code .	VSCode is started.

Step 2: Install the Julia Extension.

You won't be able to work with Julia until you install the plugin.

Step	Action	Comment
1	Ctrl+Shift-X	Open the Extensions pane. You can also hover over the most left icon column to find the option.
2	At the top of the pane, in the field `Search Extensions in Marketplace` type: julia	Find the Julia extension: Jul1a, Julia Language Support.
3	Click on: Install	Install the Julia extension.
4	Clear the field `Search Extensions in Marketplace`	Erease julia.

Step 3 (optional): Show Inline Results.

One of the great things about Juno is that you could see the result of a statement on the same line when you executed the code in a file step by step using Shift-<Enter>. This feature is now also available in the Julia extension. However, you step by step through your code in VSCode with Alt-<Enter>.

Step	Action	Comment
1	Ctrl+Shift-X	Open the Extensions pane. You can also hover over the most left icon column to find the option.
2	Type: julia	Type `julia` in the `Search Extensions in Marketplace` field when the extension is not visible.
3	Click on the: Sprocket symbol	The `Install` button has been replaced by the sprocket symbol. It opens the menu of the julia extension.
4	Select: Extension Settings
5	Click in the adjacent pane on the tab: User	The tab User is the default selected tab!
6	Search for the option: julia > Execution: ResultType
7	Choose: inline	The default is REPL, which meanns that the results are printed in the terminal.

Links:

Install IJulia

IJulia is a browser back-end for the Jupyter interactive notebook environment.

Prerequisites

Julia 1.0+ has been installed.

Step	Action	Comment
1	julia	Start Julia.
2	]	Activate package manager.
3	add IJulia	Install package.
4	BackSpace	Back to the julia prompt.
5	using IJulia	Load package.
6	notebook(detached=true, dir=".")	Start IJulia.

Only the first time you get the next question:

install Jupyter via Conda, y/n? [y]:

Type: y <Enter>

Your browser window opens.

Step	Action	Comment
7	Select: New > Julia 1.5.0-rc2	Create a new Julia notebook.
8	Type: 1 + 2	Enter a formula in a cell.
9	Shift-Enter	Execute formula. The result 3 will be shown.

Install Docker

Docker is an open container platform for developing, shipping, and running applications.

Prerequisites

Your computer OS is Ubuntu 20.04 or higher.

Step	Action	Comment
1	sudo apt-get update	Update Software Repositories.
2	sudo apt-get remove docker docker-engine docker.io	Uninstall Old Versions of Docker
3	sudo apt install docker.io	Install Docker
4	sudo systemctl start docker
5	sudo systemctl enable docker
6	sudo docker version	Check Docker Version

Install NVidia Driver

See: How to install the NVIDIA drivers on Ubuntu 20.04 Focal Fossa Linux

Step	Action	Comment
1	$ ubuntu-drivers devices	Check drivers

== /sys/devices/pci0000:00/0000:00:01.0/0000:01:00.0 ==
modalias : pci:v000010DEd00001C8Dsv000017AAsd000039D1bc03sc02i00
vendor   : NVIDIA Corporation
model    : GP107M [GeForce GTX 1050 Mobile]
driver   : nvidia-driver-390 - distro non-free
driver   : nvidia-driver-450-server - distro non-free
driver   : nvidia-driver-435 - distro non-free
driver   : nvidia-driver-418-server - distro non-free
driver   : nvidia-driver-440-server - distro non-free
driver   : nvidia-driver-450 - distro non-free recommended
driver   : xserver-xorg-video-nouveau - distro free builtin

Step	Action	Comment
2	$ sudo ubuntu-drivers autoinstall	Install recommended driver.
3	$ sudo reboot

Install CUDA for Docker Containers

See: NIVIDIA User guide

Prerequisites

You have Docker installed
Julia version 1.5.0+

Step	Action	Comment
1	Run the following code below:	Enable GPU support for Docker.

sudo mkdir -p /etc/systemd/system/docker.service.d
sudo tee /etc/systemd/system/docker.service.d/override.conf <<EOF
[Service]
ExecStart=
ExecStart=/usr/bin/dockerd --host=fd:// --add-runtime=nvidia=/usr/bin/nvidia-container-runtime
EOF
sudo systemctl daemon-reload
sudo systemctl restart docker

Step	Action	Comment
2	sudo docker start test_sshd	Start the container.
2	ssh rob@172.17.0.2	Enter the container.
3	julia	Start Julia.
4	using Pkg	Start package manager
5	Pkg.add(["BenchmarkTools", "CUDA", "Flux"]))	Load CUDA, etc.
6	<Backspace>	Return to Jullia.
7	using CUDA; has_cuda()	Test whether CUDA is enabled.

TODO: It didn't work for me. Troubleshoot!

https://github.com/JuliaGPU/docker

Important Docker Commands

Command	Comment
sudo docker images	Display all Docker Images.
sudo docker ps	Display all running containers.
sudo docker ps - a	Display all containers.
sudo docker start <name or id>	E.g `docker start test_sshd`.
sudo docker exec -it <name or id> <command>	E.g `docker exec -it test_sshd bash`.
sudo docker exec -it <name or id> <command>	E.g `docker exec -it test_sshd julia`.
Ctrl-D	Exit the container
sudo docker stop <name or id>	E.g `docker stop test_sshd`.
sudo docker rm -f <name or id>	Delete a container.
sudo docker rmi <image id>	Delete an image.

Install Ubuntu on Samsung T5 and Lemovo Legion Y520 with Windows 10

The steps I have done.

Step	Action/Response	Comment
1	Download ISO-image from Ubuntu 18.04. website
2	Copy with Rufus to USB-stick
3	Start machine and press F2	Change the Bios Lenovo

Under Boot tab:
Fast Boot disabled
Under Security tab:
Secureboot disabled
See:
https://github.com/kfechter/LegionY530Ubuntu/blob/master/Sections/InstallUbuntu.md

Step	Action/Response	Comment
4	Start Y520 with Windows 10 and press F12
5	Choose Ubuntu	Restart via USB.
6	After restart choose for install Ubuntu on 500GB portable disk
7	sudo apt update
8	sudo do-release-upgrade -d	Upgrade to Ubuntu 20.04.

To be reconsidered, I am not sure to deliver on-line courses this way.

Dockerfile

Prerequisites

Your computer OS is Ubuntu 20.04.
You have installed Dockerversion version 0.18.0 or higher.

If you have installed Docker, you can create a Docker Image from a Dockerfile. From an image you can create Docker containers.

FROM "jupyter/minimal-notebook"

USER root

ENV JULIA_VERSION=1.1.0

RUN mkdir /opt/julia-${JULIA_VERSION} && \
    cd /tmp && \
    wget -q https://julialang-s3.julialang.org/bin/linux/x64/`echo ${JULIA_VERSION} | cut -d. -f 1,2`/julia-${JULIA_VERSION}-linux-x86_64.tar.gz && \
    tar xzf julia-${JULIA_VERSION}-linux-x86_64.tar.gz -C /opt/julia-${JULIA_VERSION} --strip-components=1 && \
    rm /tmp/julia-${JULIA_VERSION}-linux-x86_64.tar.gz

RUN ln -fs /opt/julia-*/bin/julia /usr/local/bin/julia

USER $NB_UID

# Add packages and precompile

RUN julia -e 'import Pkg; Pkg.update()' && \
    julia -e 'import Pkg; Pkg.add("Plots"); using Plots' && \
    # julia -e 'import Pkg; Pkg.add("PlotlyJS"); using PlotlyJS' && \
    julia -e 'import Pkg; Pkg.add("Distributions"); using Distributions' && \
    julia -e 'import Pkg; Pkg.add("Optim"); using Optim' && \
    julia -e 'import Pkg; Pkg.add("StatsPlots"); using StatsPlots' && \
    julia -e 'import Pkg; Pkg.add("IJulia"); using IJulia' && \
    fix-permissions /home/$NB_USER

Install IJulia

Prerequisites

Your computer OS is Ubuntu 20.04.
You have installed Dockerversion version 0.18.0 or higher.

Note

The instruction are based on Ubuntu 18.04. We use the command line in a terminal window.

First check whether Docker is installed on your system, by typing docker --version. If the command docker is not recognized, install Docker.

Step	Action	Comment
1	Open a terminal window on Ubuntu	`$` is the prompt of your OS.
2	$ mkdir work	Create the folder `work`. Optionally, create sub-directories with the names of your students.
3	$ mkdir julia	Create another directory `julia'.
4	$ cd julia	Goto the directory julia.
5	Copy the code of the section Dockerfile to the clipboard
6	$ nano Dockerfile	Open the text editor nano.
7	Ctrl-Shift-V	Paste the text on the clipboard into the text editor.
8	Ctrl-O	Save the Dockerfile.
9	Ctrl-X	Exit nano
10	$ ls	list the content of the directory julia.
11	$ cat Dockerfile	Display the content of the Dockerfile.
12	$ docker build -t julia-image .	Create a Docker IJulia image. The name of the image is `julia-image`.
13	$ docker run -v ~/work:/home/jovjan/work -p 8888:8888 --name julia julia-image	Create a Docker container, and the IJulia Notebook server starts.
14	Ctrl-C	Stop the server.

Using IJulia

Instructor

Step	Action/Response	Comment
1	$ docker start julia	Start the `IJulia`.
2	$ docker exec -it julia bash	Start the Docker client.
3	jovyan@40f8c3bcf0c1:~$ jupyter notebook list	Display the token.
	`jovyan@40f8c3bcf0c1:~$ jupyter notebook list Currently running servers: http://0.0.0.0:8888/?token=f394613a12c103a2de92e70aa34715b2183d58e3d5580a2e :: /home/jovyan`
4	Select the token	In this case: `f394613a12c103a2de92e70aa34715b2183d58e3d5580a2e`.
5	Ctrl-Shift-C	Copy the token to the clipboard.
6	Store the key somewhere, so you can email it to your students.
7	$ Ctrl-D	Exit Docker client.