Need to do this as a pre-init stage to task marking which will unskip tasks that have active cells

In count_and_link_tasks task counts were missed.

In make_extra_hydroloop_tasks_serial,_make_hydro_loops_dependencies was not
called (this broke the SedovBlast test as some tasks were called out of
sequence giving bad initial values).

In marktasks_mapper/serial the send/recv tasks were missing.

Add entropy to the energy diagnostic file. Use an MPI_Reduce instead of MPI_Allreduce to get the values

Renamed 'h' in cell struct to 'width' and renamed 'h' and 'ih' in the space struct to 'width' and 'iwidth'.

engine_step now writes to a file called timesteps.txt that contains timestep information along with a header.

no need to recursively collect the drift data as it is now all computed in the top-level cells directly.

do the drifts with the threadpool instead of with the scheduler, makes the simulation ~10% faster overall.

still torn between doing engine_marktasks in parallel or not, it's faster on my laptop but not sure how this fares on a larger machine.