Abstract of ITM-HPCWF Project

Some of the workflows envisaged in the ITM-TF (and by others, including ITER) require significant compute resources for at least some components of the workflow. The approach taken by the ITM involves running a workflow engine on the Gateway (Kepler) and having the HPC level modules initiated from this workflow (using Unicore)..

An alternative approach, pursued by some American groups, is to have a single large job running on the HPC resource with sub-components of this job implementing the various modules in the workflow.

This proposal asks the HLST to implement the infrastructure used by one of the leading USA groups (IPS framework) on the HPC-FF (and/or IFERC) and then evaluate it, in strict collaboration with some key ITM-TF developers.