Used to explain the evolution of Automatic splitting in CRAB3…
The general handling of a task in CRAB3, from user submission to handling on the server, and processing via the schedd:
A first implementation of automatic splitting, with a single probe and processing done via a dependent SubDAG:
Latest implementation of the automatic splitting, using several probe jobs and a few tail DAGs to catch unprocessed data:
