next up previous contents
Next: When we do not Up: Sequence parallelization Previous: Sequence empty   Contents

In case we already have parallel tasks

If we depend on data in an active task, that is T was found and it currently exists in an active task, or, if our number of active tasks is equal to the maximum size of our sub-cluster, then we decide not to spawn more tasks. Otherwise, we will go on seeking to exploit more parallelism.