Scalability - Smart Use of HPC Resources

Europe/Warsaw
online

online

Oskar Klimas (ACK Cyfronet AGH), Klemens Noga (ACC Cyfronet AGH)
Description

The training focuses on running and analyzing the scalability tests.

Scalability tests allow to tune the amount of resources for the job being run. During this training you will learn what scalability tests are and how to interpret their results. Then how to apply that knowledge to your calculations.

Level

Target audience

Training is intended for users who started running their own calculations and need to learn how to give the right amount of resources to the job at ACC Cyfronet AGH. 

Agenda

  1. Chasing perfect scaling using embarrassingly parallel code,
  2. multinode scaling and parallelization overhead,
  3. hands-on scaling tests:
    1. Python simple examples,
    2. GROMACS scaling,
    3. Do It Yourself exercise;
  4. CPU vs GPU power and multiple GPU scaling,
  5. real life case: specific Ares to Athena transfer issue analyzed,
  6. Q&A (15 min)

Requirements

We recommend participating in this training after completing Running Calculations the HPC way training provided at ACC Cyfronet AGH.

Summary of prerequisites
  1. PLGrid account,
  2. active affiliation,
  3. accepting invitation to resources at least a day before training starts,
  4. attending the Dry-run,
  5. computer equipped with SSH connection.
PLGrid Account

Specify your PLGrid username in the registration form. If you don't have an account, please follow one of our guides:
* https://guide.plgrid.pl/en/portal-for-science/profile,
* https://guide.plgrid.pl/en/portal-for-business/account.

Affiliation
Invitation to resources
  • Couple of days before the training, Training Team will be sending invitations to resources in PLGrid Portal at https://portal.plgrid.pl/training_invitations.
  • Accepting the invitation allows users to use the resources of the training grant.
  • Make sure to activate all the required services too (they are listed on the page of the invitation). Use "participation in a training" as motivation if need be.
Dry-run
  • On 26th Oct 2026 at 10:00-11:00 CET, a Dry-run will be held. Attendance will be marked based on Zoom participant list.
  • Participants will learn how to log in to the Ares supercomputer before the training and resolve any problems with the Training Team.
  • In case you cannot come to the dry-run, please inform us directly at training@cyfronet.pl, otherwise your registration will be cancelled.
  • Please attend the dry-run even if your affiliation is not active.
SSH connection

Participants' computers should be equipped with software that allows SSH connections. On Windows, it is recommended to use Windows Terminal (not to be confused with Powershell after checking if OpenSSH is installed (run ssh -V in the terminal). We advise against using Putty.

Venue

The workshop will be conducted online on the Zoom platform. The meeting link will be sent to registered participants.

Language

English

Duration

3 hours 15 minutes

Registration and eligibility

The Registration and the Waiting list close automatically at 23:59 CEST on 22nd Oct 2026. The Registration may close prematurely if the limit number of participants is hit beforehand, but the Waiting List will be available until the deadline above.

Please use your institution's e-mail address. Registrations using generic or private email addresses cannot be accepted. At that e-mail address, you will be contacted with further information.

This course is co-funded by the European Union. Therefore, the course is open for participants from academia, industry, and public administration from the Member States (MS) of the European Union (EU) and Associated/Other Countries to the Horizon Europe programme.

Organized by

Funded by

Acknowledgements
EuroCC 3 has received funding from the European High-Performance Computing Joint Undertaking (JU) under Grant Agreement No. 101306701. The JU receives support from the European Union‘s Digital Europe Programme and Germany, Albania, Austria, Belgium, Bosnia and Herzegovina, Bulgaria, Croatia, Cyprus, Czechia, Denmark, Estonia, Finland, France, Greece, Hungary, Iceland, Ireland, Italy, Latvia, Lithuania, Luxembourg, Malta, Montenegro, the Netherlands, North Macedonia, Norway, Poland, Portugal, Romania, Serbia, Slovakia, Slovenia, Spain, Sweden, Türkiye, and Kosovo.

Organised by

ACC Cyfronet AGH

Registration
Registration
0 / 100
Waiting List
    • 1
      Introduction
      Speaker: Oskar Klimas (ACK Cyfronet AGH)
    • 2
      Intranode scaling

      Intranode and perfect scaling an how it is unattainable in real calculations

      Speaker: Oskar Klimas (ACK Cyfronet AGH)
    • 3
      Multinode scaling

      Multinode scaling and parallelization overhead,

      Speaker: Oskar Klimas (ACK Cyfronet AGH)
    • 10:55
      Short Break

      5 min break

    • 4
      Hands-on scaling tests

      Running some scaling tests yourself

      Speaker: Oskar Klimas (ACK Cyfronet AGH)
    • 11:55
      Short Break

      5 min break

    • 5
      CPU vs GPU

      CPU vs GPU power and multiple GPU scaling,

      Speaker: Oskar Klimas (ACK Cyfronet AGH)
    • 6
      Real life case

      Real life case: specific Ares to Athena transfer issue analyzed

      Speaker: Oskar Klimas (ACK Cyfronet AGH)
    • 7
      Q&A

      Answering any and all questions from you