[yt-users] corrupt enzo restart file?

[Matthew Turk]

Hi Britton,

> Sorry for chiming in so late.  I didn't look at this email initially because
> of the subject.  We could either be smarter with the Overdensity field and
> check if Density is in the field_list, or we could make a DM_Overdensity
> field and then tell the VirialFilter to use that instead with the
> overdensity_field flag.

I think that the second solution would be simpler.  The new optimized
hierarchy (below) also explicitly knows if the simulation is DM-only,
so we can have that be a check in the halo profiler rather than the
Overdensity field.  Having the overdensity_field flag explicitly be
required to be set if it's not to use baryon fields would also be
alright, I think.

> Maybe we should come back to that, but I think Matt's
> hierarchy optimization stuff is supposed to do away with that all together.

It know does do away with that -- and it is working well, and has been
tested quite a bit over the last couple months in the hg repo.  I'll
try to roll it in to trunk sometime in the next day or two, and it
should bring with it a substantial increase in the speed of parallel
projections.

-Matt