Resources per team
Teams will be given access to Spider, a versatile high-throughput data-processing platform designed for processing large, structured data sets. Spider is a combined CPU and GPU system
Teams will be granted a maximum of 100,000 CPU core hours
Limited access to GPUs on request
2 TB of persistent storage per team will be available for the duration of the data challenge
Access to Spider is provided by SSH.
Data cube access
Data will be made available in read-only mode on a shared filesystem.
Spider provides extensive monitoring and resource management functionality.
Spider provides an extensive range of centrally-managed system software.
Users are able to install software in their user or project space, and have it made available across the compute nodes.
Spider also supports user software provided in Singularity containers.
Teams will have access to the SURF Helpdesk.
Spider is hosted at SURF, the Netherlands national e-infrastructure provider.
Detailed information about Spider is available from SURF.
10,000+ CPU cores; new hardware regularly being added
Mix of Intel and AMD
At least 8 GB RAM and 80 GB scratch space per core
Per user resource
100,000 CPU hours available
2 TB persistent (for the duration of the challenge) storage per team
A wide range of system software as described in the documentation.
User software can be directly installed in home and project spaces.
Singularity containers are supported.
GPUs if any
Access to Nvidia A100 GPUs on request.
Specified role, as defined here (optional)
Please refer to the extensive documentation provided by SURF for further information.
Compute jobs are managed through SLURM.
Spider provides extensive monitoring and resource tracking capabilities.
Credits and acknowledgements
Publications or data releases should be accompanied by the acknowledgement:
This work made use of the prototype Netherlands SKA Regional Centre, provided by ASTRON and SURF, supported by the Dutch Research Council (NWO) under National Roadmap grant number 184.035.004 (FuSE: Fundamental Sciences E-infrastructure).