High Performance Computing LTAT.06.026
The goal is to provide an understanding of high performance computing, how it works, and introduce its best practices and use cases to enable efficient calculations with large complex data.
Prerequisites
You are expected to know the basics of Linux, networking and the internet, and to be able to use any preferred search engine.
Course info
This course is thought in two parts - weekly prerecorded lectures followed by practical sessions.
Lectures are there to establish basic knowledge of topics, give explanations and examples, and create an overall understanding of the topic.
Practical sessions are built to be highly hands-on but independent - students are given guides with thorough explanations how to do the lab, which is then supervised by teachers in person during the physical classroom lessons, online labs in Zoom or through Slack at other times.
Learning outcomes:
- .. Understanding of HPC architecture
- .. Understanding of computing resource management
- .. Ability to use HPC environments for scientific computing at will
- .. Beginners level at parallel computing
- .. Skills in working with command line
- .. Skills in commonly used software solutions (OnDemand: Jupyter, RStudio)
Related keywords: hpc, data management, quotas, slurm, running jobs, monitoring jobs, queue, cpu computing, gpu comuting, data storage, python, MPI, OpenMP, Jupyter, RStudio, ETAIS, containers in HPC, modules, software
Taught by: Sander Kuusemets, Ott Oopkaup, Mait Tenslind, Sander Kütisaar, Pille Pärnalaas, Rasmus Soome.
Lecture recordings can be found on panopto
Practicals will take place in room 2034.
Timetable:
- Lecture
- Prerecorded video. Mandatory to watch the lecture before the practical session.
- Practical
- Thursday: 14:15 - 16:00: the instructor changes week-by-week.
- Online Practical
- Friday: 14:15 - 16:00: instructors are Mait Tenslind and Mihkel Tiks.
Lab zoom links
(log into courses to see the Zoom link)
Communication
The course instructors themselves support two means of communication to ease communication between students, or with instructors - either in class, or using the course Slack.
Discussion is welcome and recommended between students, either during labs or Slack. No helping each other during the exam though.
Critical messages will be announced personally using SIS (Study information System messages) and Slack.
(log into courses to see the Zoom link)
Assessment
- Grading: Non-differentiated (pass, fail, not present)
- Course is graded positively (pass) if you complete all the lab exercises
There are 8 labs in total. Completion of labs will be checked by an automatic system.
Lab environment
- Labs are accessed over the network, required software:
- A VPN Client
- An SSH Client
- A Web browser
Each student will be given access to a HPC cluster on the first course week, where they will complete their labs. Preliminary Linux skills are highly recommended.