#1261: Suspicious logic in NewRad thorn ------------------------------------------+--------------------------------- Reporter: wolfgang.kastaun@… | Owner: Type: defect | Status: review Priority: major | Milestone: ET_2012_11 Component: EinsteinToolkit thorn | Version: ET_2012_11 Resolution: | Keywords: backport ------------------------------------------+---------------------------------
Comment (by hinder):
McLachlan has a "native" radiative boundary condition in the McLachlan_BSSN.m script which is not used, as most people use NewRad. This native implementation is wrong in the master branch, and I corrected it in the cakernel branch some time ago. At the same time, I created a test case for it, and cross-checked the test results with those from NewRad, and the results were the same to machine precision (3D gridfunctions checked). I then added the NewRad test to the master branch. My point is that the existing NewRad test in the master branch gives identical results to the (corrected) native implementation, so I trust it. So I agree with comment:3 that this bug might not be observable. Of course, this is only for one particular grid setup, set of parameters, number of processes etc.
The corrected native implementation should probably be merged into the master branch.