May 27, 2019
Golem splits up tasks into smaller pieces to spread it around the network so multiple people can work on a different part of your task at the same time. This is called a subtask because it's a split up piece of the original task.
Splitting the task up into multiple subtasks will almost always result in faster computation, but that's primary for larger tasks.
When Golem notices that all subtasks have finished computing, then all pieces are going to get stitched back together and you'll receive back your requested computation in it's final form.