Solutions
Technologies we offer our clients in the field of High Performance Computing.
Batch Computing
The CLIP Batch Environment (CBE) provides terminal access to a 200 node SLURM cluster with various hardware configurations.
Provided Services
Interactive Computing
Several interactive, web-based services allow to leverage CBE’s computational resources for interactive data exploration.im.
Grid Computing
Part of CLIP is Austria’s only tier 2 site for Grid Computing in international Particle Physics collaborations.
Cloud Batch Environment – CBE
The CLIP Batch Environment (CBE) utilizes SLURM, an “open source, fault-tolerant, and highly scalable cluster management and job scheduling system”.
SSH
The cluster’s login nodes are reachable via SSH from othe VBC network and parts of the ÖAW network, otherwise a VPN connection is needed.
SLURM batch cluster
- HTC (high throughput computing)
We are fulfilling the requirements for a high-throughput computing (HTC) environment. Many projects depending on a large amount of computing cycles to quantify their solutions and here is the difference to high performance computing (HPC). HPC needs a lot of computer power to solve problems in a short period, while HTC needs a lot of computer power either, but for a longer period.
- Scheduling policy
Backfill strategy applied to ensure optimal usage of hardware resources
- Easybuild/ Lmod
used for standardized software installations and creation of environmental modules for the software stack
- commercial software: Matlab/ Mathematica (bring your own license)
- Singularity/Apptainer
containers, may also be built via in-house Jenkins build pipeline
Noteworthy application
- Nextflow: allows to deploy workflows on a variety of platforms including local, HPC schedulers, AWS, Batch, Google Cloud Life Science and Kubernetes by using software containers. This enables scientists to create portable, reproducible, and scalable workflows. It simplifies the implementation and the deployment of complex parallel and reactive workflows on clouds and clusters.
- Alphafold: using Alphabet’s/Google AlphaFold AI for scientific discovery to predict protein structures with incredible speed and precision. This software is designed as a deep learning system that incorporate physical and biological knowledge about protein structure, leveraging multi-sequence alignments, into the design of the deep learning algorithm.
- R/ Bioconductor: a free, open-source life-science software project for analysis and comprehension of genomic data. Bioconductor mainly based on R programming language but contain contributions of other programing languages as well
- TensorFlow: an end-to-end free and open-source framework for machine learning and artificial intelligence. It has a comprehensive, flexible ecosystem of tools, libraries and community resources that let researchers and developers push the state-of-the-art in ML and AI.
- Python/ Numpy/ Scipy/ …
Data management
GRID FTP to stage data in/ out
Interactive Computing
JupyterHub
Our JupyterHub service allows you to spawn one or more JupyterLab sessions that are executed on the CBE compute nodes. Inside this JupyterLab session the user can start Jupyter notebooks that either run an R or Python kernel and execute the code that is stored inside the notebooks.
RStudio Server
Some people have already been using RStudio on CBE through the XPRA service. We now also support the native RStudio integration into CBE using the RStudio Server Pro solution.
XPRA
The CBE compute cluster can also be accessed from a web-based graphical interface called XPRA that is available on VBC Jupyter Hub. The idea of a web based cluster access is to run compute intense jobs that require a graphical user interface (GUI) – like Rstudio and Fiji – directly on a cluster node and interact with the application through your web browser.
Contact us
Please provide affiliated Organisation, and your use-case or resource use estimation (CPU-hours, GPU-hours, Memory)