Closed rbooth200 closed 7 years ago
Seems fine. I guess this is all independent of the other Mfv/radws/eos pull requests so I can merge now?
yes
I am happy for this to be merged. I assume that the bug with brute force was only affecting the first step anyway?
Yeah, I think so. I've got some simulations with the gadget mac running fine at the moment. They are certainly faster than brute force.
This pull request fixes some errors relating to the meshless and self-gravity with the gadget mac. Before these changes atree was being zero'd in the initializaiton resulting in a brute-force gravity calculation.
There was also a mistake for the imported cells, which has been fixed: we were trying to recompute the acceleration tolerance, amin, for imported cells, which does not work because atree is zerod.