Seeding of BHs over MPI

@jwillis if you have some time, I'd appreciate a quick second look at my theory below.

(Sorry, it has likely been a while)

I am looking at the seeding of BHs in FOF. More specifically, I am looking at the identification of the densest particle. I worry in particular about this line: https://gitlab.cosma.dur.ac.uk/swift/swiftsim/-/blob/master/src/fof.c#L1548 Should we really check the mass threshold here? We are looking at groups that have a local root. We then check whether their mass is above the threshold mass for seeding. However, the array group_mass is not done yet. We still need to include the masses of the fragments that may exist on other nodes. This is only done here: https://gitlab.cosma.dur.ac.uk/swift/swiftsim/-/blob/master/src/fof.c#L1637 My understanding is that we can only start to look at full group masses after that point.

The loop around 1548 may then be missing some haloes, and hence not account for the fact that they may already have a BH for instance. Or not identify the correct densest particle.

Am I looking at this correctly or am I missing something here?