Slow submissions

[jthorton]

When generating tasks for large datasets this can be quite slow and maybe it would be best to chuck up the dataset and call compute using threading this would allow us to submit the datasets faster.

[jthorton]

Submission threading is now possible and for a 3000 entry optimization dataset we see around a 1.4x speed up during the submission process. We might be able to gain more speedups using some combination of threading and multiprocessing, Ill leave this open to collect ideas.