We’re using cylc to manage the transfer of data from our old archive to JASMIN. Each cycle is responsible for transferring a chunk of data and running several other tasks to verify the data transfer. The amount of data each cycle is dealing with varies so later cycles may run quicker that previous cycles. Tasks have no dependency on their previous task instance, however they are only spawned in later cycles once the task instance from the previous cycle has been submitted. This is causing a bottleneck. I understand there is a way to force cylc to spawn all tasks for a cycle upon cycle initialisation, but this is not documented. Please could you send me instructions on how to do this?