
Welcome to our info page on the use of supercomputers! Here you will find information on various topics in Cosmos High Performance Computing. Use our search or jump to the different areas via the tiles.
Please note: This page collects links and information on specific topics. If you have individual questions or need more basic information, please visit our page on training or use our contact form.
HPC ACCESS
SOFTWARE
DATA MANAGEMENT
VISUALISATION & POSTPROCESSING
HPC ACCESS
- Requirements (computing time project, JUDOOR account), instructions SSH login, GUI via X-forwarding, Putty, VNC: JUWELS, JURECA or via JupyterLabFolien and Video)
- Job-Monitoring und Nutzer-Portal
- Nutzungsbeispiel: Deep Learning on Supercomputers – An Introduction (MNIST Tutorial mit Keras/Tensorflow, Anpassung für Distributed Training (Horovod), (GIT repo)
Projekt
- At LRZ, only projects from academic institutions can apply, but industry can be included as a project partner.
- Proposal:
- What comes after an approval?
SSH
Overview and Status of HPC-Systems
Graphisch aufbereitetes HPC-Job-Monitoring: Portal, Documentation
Working environments
- Compute Cloud: Information, Overview
VNCUsage of a pre-configured remote desktop on dedicated nodes on Linux cluster or supercomputer
- HPC ACCESS
- SSH login via classic console with fixed IP address
- HPC ACCESS
- SSH login via classic console with fixed IP address
#HPE Apollo (#Hawk) #NEC Cluster access (#vulcan) #SSH-Login
Software
No High Performance Computing, AI or Cloud Computing without software! Modern HPC systems differ significantly from consumer computers in terms of their architecture (e.g. type and number of computing cores, size of main memory), storage media and, above all, the usually complex network topology. Therefore, operating systems adapted to the HPC systems are used. Based on these, a broad software portfolio is made available to users. On the one hand, this includes basic software adapted to the HPC system, such as compilers or communication libraries for parallel computing, and on the other hand, HPC applications, such as simulation software and scientific libraries. The software offer is usually available to users via a modular system. Software components in the desired versions can be assembled and used in computing jobs.
The offer includes open source packages and community applications as well as commercial software. Furthermore, users are free to install software packages or self-developed programmes in their private directory. On request, software can also be included in the portfolio or support provided for installation and configuration on the HPC system.
Explanations and examples of software are given below.
#Anwendersoftware #Bibliotheken #Entwicklungsumgebung #Jupyter #Modulsystem #Paketmanager #Remote Access #Spack #Userspace
- Provided software environment: Module system and Spack: Documentation, Overview
- Spack in the user space for own installation through user
- Jupyter on HPC systems
- Software with remote-access-functionality, e.g.: Visual Studio Code, Matlab
- Application software packages – HLRS Platforms
Datenmanagement
HPC projects typically generate large amounts of data. Storage systems with features that meet requirements ensure optimal performance when handling this data. For example, parallel file systems enable fast, parallel reading and writing of data during an HPC job, which can be moved to long-term archives for permanent backup. Explanations of the different storage systems and their use, as well as the possibilities of data transfer between different file systems are linked in this category.
#Archiv #Dateisysteme #Datenkurator #Datentransfer #DSS #GIT #GlobusOnline #HPSS Archiv #Langzeitspeicher #LinuxCluster #NEC Cluster Disk Storage #parallel #Ressourcen #scp #Speicher #Storage #SuperMUC-NG #WorkSpace
- Properties and Usage of file systems Supercomputer, Linux Cluster
- Data Science Storage
- Data Science Archive
- Datatranfer between centres or user-workspaces and centre via globus online or scp
Visualisation & Postprocessing
We support engineers and scientists, industrial users but also the public sector in the visual analysis of data typically generated by simulations on high-performance computers.
Different visualisation methods and technologies from our portfolio can be used for this purpose. For the visualisation of results, separate access to designated visualisation nodes is necessary.
Access is via Remote Desktop (via a web browser) or Client-Server (software required). Another option is via VNC viewer (software required) or the use of JupyterLab (web browser). Further information and documentation can be found here.
#Browser #Desktop #LinuxCluster #OpenGL #Paraview #Remote #SuperMUC-NG #VirtualGL #VisIt #Visualisierung #VNC #Web-Frontend #Xpra
- Remote Visualisation 2021: web-based frontend, work on a remote desktop. Portal, Documentation
- Paraview in jobs (batch-mode) or for interactive wrk on nodes (client-server-nodes)
- ParaView -> Login via Client: Hawk PrePostProcessing – HLRS Platforms
- Graphic Environment: VNC-Zugang: Graphic Environment – HLRS Platforms
- COVISE tutorial.pdf (hlrs.de)
- COVISE Online Documentation (hlrs.de)