remove ADMCoupling from test cases, use ML_ADMConstraints instead of (nonexistent) ADMConstraints
Roland Haas
Branch: rhaas/noadmcoupling
Branch: master
Merged
Merged pull request
GRHydro: regenerate constraint data for tests that used ADMConstraints
GRHydro: remove ADMCoupling from tests
GRHydro_InitData: remove ADMCoupling from tests
GRHydro: remove ADMCoupling from tests
GRHydro: use ML_ADMConstraints in tests
GRHydro: regenerate constraint data for tests that used ADMConstraints
ADMConstraints and ML_ADMConstraints differ in detail:
ML_ADMConstraints is always 4th order accurate FD, ADMConstraints depends on ADMMacros's order
ML_ADMConstraints uses a "flat" boundary condition with boundary width width, ADMConstraints uses "static" (zero in practise) and 1 point wide (at least in this set of tests)
I have verified that the test using ML_ADMConstraints pass with the old ADMConstraints data if one modifies ML_ADMConstraints to
use 2nd order finite differencing
pretends the boundary is 1 point wide
Â