Also what might work on Skylake, but these cpuids are hard to comfirm

Stats after kick

Implements the change proposed in #193. 

Triple benefit:
 - Stats are computed at the correct point in time and are hence more accurate.
 - The drift does not compute stats anymore so this might help with the bandwidth.
 - The cell structure becomes 96 bytes lighter which will also help with the bandwidth.

The stats are now computed when requested by the used using a threadpoolized loop over the particles. This happens rarely and is hence not a burden on the whole system any more. The interesting stuff is in `src/statistics.[ch]`.

This is built on top of `mark_task_in_drift2` so should merge smoothly with master very soon.

See merge request !269

Conflicts:
	src/engine.c

Otherwise until engine_init we see [0000]

Mark tasks in drift2

This "works" as in "I'm going to be now, it runs 100 steps of CosmoVolume without barfing".

@pdraper, if you have any spare cycles, can you run it on anything you've got until it breaks? Otherwise, I'll just keep on hacking at it as soon as I've got another hour or two. Thanks!

See merge request !267

Conflicts:
   src/engine.c

Add cell links for gravity tasks

Add links for gravity tasks so that we can discover these from cells.

Also expunge external_gravity from cell.h.

See merge request !268

Expunge external_gravity from cell.h