The HPC Ecosystems Project will be hosting the next ...
OpenHPC2.x virtual training lab.
This FREE training lab is hosted as a virtual training workshop to gain hands-on experience with deploying a three-node virtual cluster using the OpenHPC2.x software stack.
How
Virtual Resources
Using Vagrant to automate the deployment and VirtualBox to host the virtual machines, the virtual cluster will be deployed on your own machine and will remain available to you for future development and practice.
Training Material
All of the material will be presented to you from the start (all you need is the online guide), and you will work at your own pace to complete the material. Upon successful completion, you will have a functional 3-node OpenHPC virtual cluster.
In addition to the online guide for the learning material, all additional interaction (questions, discussions, networking) is facilitated through a virtual classroom.
The course material is estimated to take most people between 15-20 hours of hands-on time to complete, with the course being open for a period of just over two weeks (straddling three weekends to give most people more than enough time to pace themselves comfortably).
A little bit more about the asynchronous content delivery:
Pre-Requisites
Experience
It is suggested that participants have experience and a basic understanding of:
The workshop will concentrate on a guided hands-on OpenHPC deployment that is officially supported for the HPC Ecosystems and SADC communities.
At the end of the deployment process, a fully functional Management Server will be operational as a VM, and can be migrated to a physical system, or used as-is to manage a physical HPC cluster, or to perform future HPC-related training and learning through the virtual cluster environment.
Computational Resources
The Virtual Lab is designed to operate on x86_64 systems with 8GB of RAM. Although other environments have been successful, these may require some tweaks and customisation to the virtual training (e.g. reducing the number of compute nodes, using Arm workarounds, etc.).
The format of the programme is designed to accommodate all partner institutions in any timezone, with participants of any experience or background (provided the foundational reference material is reviewed in the case of complete newcomers to HPC) with any time availability or workload commitments.
You do not need an operational HPC system to perform this training! Everything is designed to run on your local machine (desktop or laptop). There are many participants who have successfully run the 3-node virtual cluster on their local machines, but if you experience resource constraints, we will do our best to investigate and assist.
In extreme circumstances of low or unreliable internet connectivity, there are mechanisms to perform the lab entirely offline in your own time, although we will unfortunately not be able to facilitate this during the official virtual lab - please contact us for further information and assistance if you are faced with connectivity issues that prevent you from participating fully in the official virtual lab.
Please note: there is no cost to participate.
We look forward to sharing this learning journey with you!