Currently the pool first schedules all the jobs on the master node, and only after calculating all the jobs do they start running. If the jobs were scheduled on a thread, we could already commence execution. We would have to keep the pool open until the thread finishes, to make sure there's no race condition when the pool finishes all work before it has finished scheduling all work.
Show the title/phase of the pool
Multiple pools can run in succession, it would be nice if we encapsulate a single pool into a group of pools which we could call a "workflow".
Experiment with scheduling jobs on a thread
Currently the pool first schedules all the jobs on the master node, and only after calculating all the jobs do they start running. If the jobs were scheduled on a thread, we could already commence execution. We would have to keep the pool open until the thread finishes, to make sure there's no race condition when the pool finishes all work before it has finished scheduling all work.
Show the title/phase of the pool
Multiple pools can run in succession, it would be nice if we encapsulate a single pool into a group of pools which we could call a "workflow".