Should we include an information of the ability of a modelscript to use multiple cores?

[llavall]

I would like to raise the question if that information might be useful in the future and we therefore should consider to include it, either in the Excel sheet/metadata itself provided by the modeler or at least in the FSK repository.

One use case for us could be the option to automatically decide in what environment (number of cores available) a model should run later, assuming we will have the option to address server with different core numbers.

If we don't want the modeler to provide that information we could run the model during the curation step on a low core server vs. a high core server and assume the ability of multicore use via the time saving.

The main key here is the precondition that the model script is able to use multiple cores with inbuilt features only, which means R packages for now, in order to distinguish from our latest experience of a model optimized for a 100 core HPC, but with the need of special software installed to successfully distribute the processes to the cores.

[mfilter]

It's a good and valid point. For sure we could add a free-text field that allow the modeller to describe this feature. Still this would then have to go into the GoogleSheet with metadata proposals. The other side is how we handle this on the server - here I suggest an update to the description of FSK-Lab and the model repository upload welcome page that relates to this point. In the end we have to look at this point during the curation step.

[schuelet]

@mfilter This is reaching for the stars. I put it on Ice if you want to keep it.

[llavall]

I think if we have that information from the modeler we can think about how to make use of that in the future. At the moment I am only aware of the Norovirus model which is able to do that, because I was working on the parallelisation. Modelers should be motivated to use multicore features if possible, in order to speed up the calculation time.