Efficient usage of HPC systems in scientific computing
online
The "Efficient usage of HPC systems in scientific computing" course is aimed at users who want to expand their knowledge and skills in working with HPC systems available at ACC Cyfronet AGH.
Agenda
- HPC systems usage
- login and data transfer
- command line interface for remote access to servers
- remote visualisation and remote desktops
- Available computational and storage resources of Ares, Athena and Helios supercomputers
- Scientific software environment management (Modules/Lmod)
- compilation of users' software
- usage of users' own modules
- Working with queuing systems
- batch scripts
- sequential, parallel and distributed computations
- srun
- OpenMP
- MPI
- hybrid OpenMP + MPI
- jobs with GPGPUs
- array jobs
- jobs' dependencies
- usage of additional node features
- sequential, parallel and distributed computations
- interactive jobs
- attaching terminal session to running jobs
- batch scripts
- efficient usage of the SLURM queuing system
- job monitoring
- efficient usage of HPC filesystems
- Python environments management
- HPC containers
- Working with big datasets
- Documentation and user support
- Ares and Athena as an integral part of PLGrid Infrastructure
- EuroHPC (LUMI) - computational opportunities
Duration
5 hours
Requirements
Each user should have a PLGrid account, active affiliation and access to the Ares cluster. Accounts and access to supercomputers could be obtained in Portal PLGrid. If somebody has no active affiliation, please contact organisers at training@cyfronet.pl.
That is more advanced training for people familiar with basic usage of HPC resources. Basic training for people who want to start using HPC resources at ACC Cyfronet AGH is Introduction to Scientific High-Performance Computing.
Venue
Training will be conducted via Zoom through the link provided for registered participants.
Language
English/Polish - dependent on the participants.
Level
intermediate
ACC Cyfronet AGH