Troubleshooting KHARMA

Troubleshooting KHARMA

Or, "Oh No, My Simulation Blew Up! What now?"

This page aims to catalogue known bugs or unstable parameter combinations in KHARMA. If an instability is unavoidable, eventually it should be moved into the code proper as a warning or overridable error. Many such issues languish for a while as common single problems, but without a known cause or universal remedy. This page is for all those.

Currently these are links to relevant PRs/discussions, which give problem descriptions and workarounds.

"DivB exceeds maximum!" with WENO and face-centered B fields in polar coordinates

• https://github.com/AFD-Illinois/kharma/pull/132, first note (i.e. use weno5_linear, weno5_lower_poles, or ppm reconstructions)

Low-density, high-temperature shockwave from pole through disk

Still lots of possible causes for this. These could help:

• https://github.com/AFD-Illinois/kharma/pull/132, second note (i.e. reconnect_B3_inner/outer_x2)
• https://github.com/AFD-Illinois/kharma/issues/79 provides a general overview
• Lowering bsq_over_u_max, thereby maintaining a higher but consistent funnel temperature less prone to inversion weirdness

$1D_w$/onedw inverter produces infalling material along the pole

This might be inverter weirdness, or an out-and-out bug which allows returning incorrect velocities instead of failing. The solution either way is the kastaun inverter, which is now the default.