Quick-start guide
Whether you’re developing or running simulation software, it can be challenging to figure out where to start when joining a new group. This is especially true if you’ve never programmed or used high-performance computing resources (HPC, often colloquially called supercomputing), as the sheer amount of things you need to learn can feel overwhelming. This guide aims to help smooth out this process by providing a one-stop list of useful skills, programs and resources you’ll need while working at CTCMS. A lot of your work will likely be done on HPC clusters (supercomputers) which are accessed via a remote terminal (command-line) connection, so this guide will briefly cover how to connect to and use HPC clusters as well.
Terminology
Below are some useful definitions and terminology which will appear throughout this guide:
- Operating system: software such as Windows, macOS or Linux which underpins the operation of a computer. Manages software interfacing with the computer’s hardware, and handles starting and scheduling user programs.
- Unix: a family of operating systems with similar design and functionality. Linux and macOS are both Unix systems.
- Shell: a program that lets us give commands to a computer and receive output. It is also referred to as the terminal, terminal emulator or command line.
- SSH: stands for secure shell. Program used to connect and login to a remote, network connected computer (such as a HPC cluster).
- File system: program which manages the storage, retrieval and organisation of data (files) on disk.
- File manager: a program which lets us browse files stored on a computer (either a local workstation or a remote file server) using a convenient graphical interface. Examples include Explorer in Windows, Finder in macOS and Nautilus in Ubuntu Linux (other Linux distributions may have different file managers).
- Directory: An abstraction which allows for multiple files to be grouped together in a single “location”. Often also referred to as a folder.
- Cluster: a computer system consisting of multiple smaller, tightly interlinked computers (called nodes), which are capable coordinating to carry out large, computationally intensive calculations in parallel. Often referred to as a supercomputer.
Setting up your computer
Unsurprisingly, you’ll need a computer to do computational science. An important choice to think about before you start is which operating system to use. The operating system (OS) is the layer of software which interacts directly with the computer hardware and serves as an interface layer whenever you need the computer to do something like run a program, write to a file or connect to a network. There are multiple operating systems in common use, including Microsoft Windows, Apple macOS and Linux; which one you choose will affect the software and tools available to you, as well as how you interface with the computer.
This guide will focus on the “Big Three” operating systems which are most commonly used in scientific computing: Windows, Mac and Linux. You can get by with any of these, but Linux will provide you with the best and most complete environment to work in (even if it has a steeper learning curve if you’re used to Windows). It is also the most complete “out-of-the-box” experience and will include (almost) all the software you’ll need by default, with no additional work required beyond installing the operating system.
What follows are some quick-start guides to get you up and running, whichever operating system you decide to use. These guides assume you are using either your own computer (such as a laptop you bring to work), or a university computer on which you have local administrator access (your local IT department will be able to tell you if you have local admin). If you’re using a shared computer at UQ then most of the required programs should already be installed (contact UQ ITS if there’s anything in this guide which you need but isn’t installed on the shared computers - you won’t be able to do it yourself).
Windows
Windows users can get by with relatively little setup, as long as you’re willing to do most of your work on remote HPC clusters. You’ll also need to get comfortable with using the command-line to do things, which has a learning-curve but is well worth the effort.
In order to connect to a remote server like a cluster you’ll need two programs: an SSH client
and an SFTP client. Windows 10 has both of these built in, but they must be accessed from the command
line. To use them, launch either cmd.exe or Powershell and run the commands ssh.exe or sftp.exe,
following the instructions in the CTCMS introduction to
Linux.
It might also be useful to install the Python programming language on your computer, as this is a very commonly-used tool in computational science. Follow the installation instructions on the official Python website, or use the versions available on the clusters.
Finally, if you need to do any development work (i.e. writing code), you’ll need to install a development environment (sort of like a word-processor for code). The easiest such program to install and use on Windows is Microsoft Visual Studio, which is freely available under a community license. Visual Studio has support for the all programming languages you’re likely to use, such as C, C++, Fortran and Python.
MacOS
MacOS already comes with SSH and SFTP clients, which you run by opening a terminal (by
opening Finder then typing terminal and opening the Terminal application) and running the command
ssh or sftp (respectively).
For additional software, it’s a good idea to install the Homebrew software manager since this will
allow you to install almost everything you’re likely to need. Follow the instructions on the Homebrew
website and you should be able to install software by running the command
brew install \<pkg\> in the terminal (where \<pkg\> is the name of the package you want to
install).
MacOS comes bundled with an old version of Python (2.7), which may be enough for your needs if you’re
not doing anything complicated. If you want a newer version (i.e. version 3 or greater), the easiest way
is to run brew install python in a terminal to install it via Homebrew.
For software development, I recommend using either the native MacOS XCode editor, or downloading
Microsoft VS Code via this link.
Linux
Unlike the other operating systems in this guide, the software available to you on Linux will depend on which distribution you use - each distribution has different versions of software available, along with slightly different commands needed to install or update software. Many distributions will have a software center (a graphical application, similar to an app store) to let you install software, but some require you to use the command-line. Here is a short list of how to install and update software for the most common distributions (all of these options require superuser/admin permissions. Talk to ITS if you don’t have this on your machine but think you need it):
| Distribution | Install software | Update software | Search for software |
| Ubuntu/Debian | sudo apt install <pkg> |
sudo apt upgrade |
apt search <query> |
| Fedora/CentOS | sudo dnf install <pkg> |
sudo dnf update |
dnf search <query> |
| Arch/Manjaro | sudo pacman -S <pkg> |
sudo pacman -Syu |
pacman -Ss <query> |
Almost every Linux distribution should come with the ssh and sftp utilities, which you
can run in a terminal via the respective commands. Almost every distribution also comes with a version
of Python installed, although some distributions have an old version (e.g. 2.7) while some have newer
versions. You can usually install newer versions via the software center or command-line. You can run
the installed Python interpreter by typing python at the command-line, although some distributions
like CentOS or Debian
may require you to manually specify the version of Python by running e.g. python3 for the newer
versions.
There’s a huge variety of software development tools available for Linux, but at the absolute minimum
you’ll need a development environment, as well as compilers for C, C++ and Fortran. The easiest way to
do this is to install gcc, g++ and gfortran via the command-line (the exact name of the packages to
install will differ between distributions) and to download Microsoft VS Code from this
link.
Getting access to the clusters
You’ll spend a lot of time running simulations on the various HPC clusters the Bernhardt group has access to. There are three organisations which run the clusters we use and you’ll need an account with all three, which you should create using your UQ email address. If you already have an account with one of these organisations then you can still use it, but you’ll need to make a request to be transferred to the correct project groups. Ask your supervisor for the relevant project IDs.
Here’s a table with the organisations, clusters, project ID and signup links for the major clusters we use:
| Organisation | Cluster names | Signup link |
| UQ Research Computing Centre (RCC) | Bunya, Wiener | https://rcc.uq.edu.au/high-performance-computing |
| National Computational Infrastructure (NCI) | Gadi | https://nci.org.au/users/how-access-nci |
| Pawsey | Setonix | https://support.pawsey.org.au/portal/servicedesk/customer/user/signup |
(Note that when signing up for Pawsey clusters, you may need to ask your group leader to invite you to the relevant project/s before you can log in to the clusters.)
Once you have an account, you can log in to the clusters via SSH using the following commands:
| Bunya (RCC) | ssh username@bunya.rcc.uq.edu.au |
| Wiener(RCC) | ssh username@wiener.hpc.dc.uq.edu.au |
| Gadi (NCI) | ssh username@gadi.nci.org.au |
| Setonix (Pawsey) | ssh username@setonix.pawsey.org.au |
(Replace username with your username on the cluster, which would have been sent in an email when you
signed up.)
The next steps
If you already know how to use compute clusters (e.g. navigating, running jobs and important Linux commands) then you’re good to go! If not, CTCMS has guides on introduction to the Linux command-line and advanced Linux command-line usage which should be enough to get you up and running. If you prefer video tutorials, the Queensland Cyber Infrastructure Fund (QCIF)’s Marlies Hankel has produced a series of short instruction videos on the basics of Linux and HPC which are available on the QCIF YouTube page. CTCMS also has a guide on visualising and plotting data with gnuplot, which covers useful tools for presenting your research results in an approachable manner.
Each of the clusters have extensive user support documentation which covers topics including compiling and running code, submitting compute jobs and efficient use of available software. These guides are excellent resources for if you get stuck, and should be your first port of call. If you need help with something which is not covered in the guides, or need something installed or changed on the clusters, then you’ll need to lodge a ticket with the relevant helpdesk. The guides and helpdesks can be found at the following URLs:
| Organisation | Cluster names | User support guide URL | Helpdesk |
| UQ Research Computing Centre (RCC) | Bunya, Wiener | https://github.com/UQ-RCC/hpc-docs | https://rcc.uq.edu.au/support-desk |
| National Computational Infrastructure (NCI) | Gadi | https://opus.nci.org.au/display/Help/Gadi+User+Guide | https://help.nci.org.au |
| Pawsey | Setonix | https://support.pawsey.org.au/documentation/ | https://support.pawsey.org.au/portal/servicedesk/customer/user/login?destination=portals |