ScienceCluster (incl. ScienceApps)

With a cluster you can access a set of machines that work together so that, in many respects, they can be viewed as a single system. ScienceCluster gives you the possibility to choose the most suitable computational resources for each application you need to run: CPU intensive workloads or workloads requiring a large amount of shared memory as well as workloads requiring a GPU accelerator or a high speed interconnect.

• Batch-queue cluster system service for research
• Targeted to address large-scale computational research
• Cluster-type of service based on
  - Slurm workload manager running on Linux operating system
  - Ceph storage platform
  - Open OnDemand web-based portal for HPC
• Provides access to CPUs (also with high speed interconnects) and GPUs
• Centrally managed by Science IT
• Selection of pre-installed compilers, apps, and container runtime(s) available

Typically ScienceCluster is accessed using the command line (e.g. with ssh and bash), which gives you the full range of flexibility when using the service.

ScienceApps

As part of the ScienceCluster service you can also get access to the ScienceApps platform. It provides users with a point-and-click, browser based portal to access a subset of ScienceCluster's functionality via a suite of interactive computational environments. The following environments are available:

• Jupyter Notebooks (including the option for custom kernels)
• RStudio
• Matlab

The ScienceCluster service is accompanied by support from Science IT experts, removing not only the overhead of setting up your local hardware systems and servers, but also providing you access to new skills and expertise.

For help in using the ScienceCluster and ScienceApps contact Science IT.

For introductory courses see Trainings.

If your group already has a Science IT Computing Services agreement in place:

• Contact your supervisor to learn who is the project manager within your group
• The project manager will then initiate your access

If your group is starting to use Science IT Computing services:

• To get onboarded contact Science IT to schedule an initial meeting
• Topics discussed in this meeting include: Your use cases, services overview, terms and conditions (incl. costs), trainings, support provided, next steps

• Framework agreement covering all Science IT Computing Services
• Pay-per-use cost contribution model
• Usage within limits of regulations related to use of IT at UZH
• Access to computing services restricted to UZH researchers
• More details (incl. costs & FAQs) are available in the UZH Intranet