Waiting for the last task to finish when the others tasks might have finished much earlier unnecessarily increases the program running time. A better solution would be to process the data as soon it becomes available.