If the limits of your own pc or laptop are inhibiting your research, it is time to scale up to a higher level. There are several options for employees and students who require more computing power than their own desktop or laptop can provide.
SURF Lisa Compute Cluster:
Lisa Compute Cluster is a centrally managed Linux cluster, that is ideal for large-scale computations. It’s a service comprising a wide range of resources, compilers and software, such as R statistics and MATLAB, and libraries, such as the Math Kernel Library (MKL) or Intel. SURFsara continually adjusts the service to the needs of the user community. For example, Lisa Compute Cluster includes accelerators (very fast processors) and high memory nodes (for users who need nodes with extra memory). When processing batch jobs, SURFsara applies a ‘fair share’ mechanism.
BAZIS Compute Cluster
IT for Research (ITvO) offers access to your own Linux computational cluster at the VU. BAZIS is a managed service for high performance computing (HPC). Research groups can add their own compute servers to BAZIS and directly use software and options that are usually only available on a supercomputer. BAZIS also has 7 GPU nodes with 240GB memory and 32 cores for general use, sponsored by the VU HPC Counsel. Compute results are stored in Scistor and can be directly available on your own workstation for further analysing.
SURF Research Capacity Computing Service (RCCS) further extends the solutions available with the on-premise Bazis cluster and SURF Lisa CPU cluster with:
• Access to Lisa GPU island
• Access to SURF Cartesius
• Access to SURF HPC-CLoud
• Project space on SURF systems
Watch this youtube video for detailed information.
To help you to efficiently use scientific software on your workstation or compute cluster there are regular training possibilities. Check in your department, at https://hpc.labs.vu.nl or https://training.prace-ri.eu/
Please contact the IT for Research department in order to discuss your possibilities.
For every step of your data analysis, good documentation is necessary.