The documentation vaguely implies that Simba (run with Gizmo) uses SPH with a cubic spline kernel, when actually it uses a cubic spline kernel in an MFM hydro solver, formally giving different smoothing of the particles (e.g. Fig. 1 in Hopkins 2015). I think that the practical suggestion is to just approximate the MFM domains as SPH particles (like approximating the Voronoi cells in Arepo for IllustrisTNG), but the documentation should be clearer that this is a suggested approximation rather than the "correct" approach as it would be for an SPH simulation.
The documentation vaguely implies that Simba (run with Gizmo) uses SPH with a cubic spline kernel, when actually it uses a cubic spline kernel in an MFM hydro solver, formally giving different smoothing of the particles (e.g. Fig. 1 in Hopkins 2015). I think that the practical suggestion is to just approximate the MFM domains as SPH particles (like approximating the Voronoi cells in Arepo for IllustrisTNG), but the documentation should be clearer that this is a suggested approximation rather than the "correct" approach as it would be for an SPH simulation.
Thanks to Omphile Rabyang for pointing this out!