February 2025 / Intro to Triton and HPC / HPC Winter Kickstart
Quick links
News and important links (20/02/2025):
Link to register: https://link.webropol.com/ep/hpcwinter25
See the time-line below
Livestream (morning): https://twitch.tv/coderefinery
Video playlist: https://www.youtube.com/playlist?list=PLZLVmS9rf3nOeuqXNa8tS-tDtdQrES2We
Exercises (afternoon): Link sent to registered participants
Notes document: Link sent to registered participants
For winter 2025, we are running an intensive version (one-day-only) of our Kickstart course. The longer version of the course will be held on the first week of June and will last for 3 half days
Learning goals
Learn the basics of High Performance Computing with slurm
Watch a step-by-step example of the typical data analysis workflow with Aalto Triton HPC cluster
Engage with hands-on exercises to make sure you are able to run your analysis on Triton
This course is part of Scientific Computing in Practice lecture series at Aalto University, supported by many others outside Aalto, and offered to others as part of CodeRefinery.
Practical information
The course happens on Wed 26 February 2025 and is divided in two parts:
Morning lectures (9:45 - 12:00 EET): This is the livestream demo part of the course. Everyone may attend the livestream at https://twitch.tv/coderefinery, no registration needed. This is done so that we get a higher quality recording without any personal data from course participants
Lunch (12:00 - 13:00 EET): Lunch on your own
Afternoon hands-on session (13:00 - 16:00): This is the practical part of the course. We will be connected to the same zoom room and do the exercises together.
Cost: Free!
Language: English
Additional course info at: scip@aalto.fi
Schedule
All times are EEST (Europe/Helsinki time)!
The daily schedule will be adjusted based on the audience’s questions. There will be frequent breaks and continuous questions time going on, this is the mass equivalent of an informal help session to get you started with the computing resources.
Subject to change
Schedule may still have minor updates as it happens.
09:45–10:00: Joining time/icebreaker
10:00–10:15 Introduction: what are HPC systems? Enrico Glerean and other staff
10:15–11:45: A day in the life of an HPC user Richard Darst and Simo Tuomisto
Datasets/projects that could be used for the demo:
Loading an application? Not needed but good to remeber (e.g. python env): Modules
-
5.4 More advanced parallelisation - discussion only - (multithreading/multiprocessing, MPI, GPUs
Visualising the results (e.g OOD)
Moving the data away from the cluster
11:45–12:00: Where to go from here and how to ask for help (Susanne Merz and Enrico Glerean)
12:00–13:00: Lunch break (on your own)
13:00–16:00: Hands-on exercises with Triton HPC cluster
Main room: Lobby and Generic questions (SM)
Room 1: Hands-on with exercises from the morning, Slurm and Triton basics (RD, EG)
Room 2: Connecting questions? (TP)
Room 3: GPUs and parallelization (ST)
Room 4: AI / LLMs (YT)
Room 5: Speech2Text (TR)
Preparation
We strongly recommend you are familiar with the Linux command line. Browsing the following material is sufficient:
Command line/shell basics [ref] [video] - Important background knowledge which we won’t go over again.
A more detailed version of the above, for those who automate a lot of analysis, is `[Basic Linux shell and scripting]<https://www.youtube.com/watch?v=ESXLbtaxpdI&list=PLZLVmS9rf3nN_tMPgqoUQac9bTjZw8JYc&index=3>`__
Or read/watch the shorter crash course / video.
Watch this background info about why we use computer clusters. This is important information for why we are in this course, which we won’t cover directly. The most important videos are the intro (what is a cluster and why?), storage hierarchy (how the data looks), and the Slurm job scheduler (how the cluster runs things).
How to attend: Online workshops can be a productive format, but it takes some effort to get ready. Browse these resources:
Attending a livestream workshop, good to read in detail (ignore the CodeRefinery-specific parts).
How to use HackMD to take answer questions and hold discussions.
Technical prerequisites
Software installation
SSH client to connect to the cluster (+ be able to connect, see next point)
Zoom (if attending breakout rooms)
Cluster account and connection verification:
Access to your computer cluster.
Aalto: if you do not yet have access to Triton, request an account in advance.
Then, connect and get it working
Aalto (and possibly useful to others): try to connect to Triton to be ready. Come to the Wednesday session for help connecting (required).
Next steps / follow-up courses
Keep the Triton quick reference close (or equivalent for your cluster), or print this cheatsheet if that’s your thing.
Each year the first day has varying topics presented. We don’t repeat these every year, but we strongly recommend that you watch some of these videos yourself as preparation.
Very strongly recommended:
Installing Python packages with Conda (Note that conda on new-Triton has changed. See Python Environments with Conda for details)
Git intro (useful)
Other useful material in previous versions of this course:
Scientific Computing workflows at Aalto - concepts apply to other sites, too (optional): lecture notes and video, reference material.
Tools of scientific computing (optional): lecture notes and video
While not an official part of this course, we suggest these videos (co-produced by our staff) as a follow-up perspective:
Attend a CodeRefinery workshop, which teaches more useful tools for scientific software development.
Look at Hands-on Scientific Computing for an online course to either browse or take for credits.
Cluster Etiquette (in Research Software Hour): The Summer Kickstart teaches what you can do from this course, but what should you do to be a good user.
How to tame the cluster (in Research Software Hour). This mostly repeats the contents of this course, with a bit more discussion, and working one example from start to parallel.
Community standards
We hope to make a good learning environment for everyone, and expect everyone to do their part for this. If there is anything we can do to support that, let us know.
If there is anything wrong, tell us right away - if you need to contact us privately, you can message the host on Zoom or contact us outside the course. This could be as simple as “speak louder / text on screen is unreadable / go slower” or as complex as “someone is distracting our group by discussing too advanced things”.
Material
See the schedule