Because RASCIL, as example, needs for one step a lot of memory (~15 GB in my example), which is more than what a worker gets. -> Investigate how those cases should be handled, especially if we design that each core should get one worker.
This is required if we want to use Dask to parallelize single-threaded algorithms on a node.
Because RASCIL, as example, needs for one step a lot of memory (~15 GB in my example), which is more than what a worker gets. -> Investigate how those cases should be handled, especially if we design that each core should get one worker.
This is required if we want to use Dask to parallelize single-threaded algorithms on a node.