- [ ] Adjoint cubic spline source #227 - [x] Expand `Simulation.gradient` for tri-axial anisotropy - [x] `Interpolation`: Change `nearest` for properties to actual cell value - [x] #235 Currently, a Simulation is copied to every parallel process, including all models, fields, etc. This should change to save RAM. - [ ] Implement *rigorous interpolation*: Shantsev and Maaø (2015), GJI. - [x] `compute1d` in `Simulation`: get fast 1D results from an approximation of the 3D model (#302) - [ ] #233 properly into docs (https://github.com/emsig/emg3d-gallery/issues/2); e.g. - build on GHA, push to RTD; - build all docs an GHA and host on gh-pages (need to add a version thing) - ... . - [ ] Add time-domain computations to `Simulation` - [ ] Additional receiver types: - `RxFluxPoint` (magnetic flux density) - `RxCurrentPoint` (electric current density) - [ ] Expand solver to handle Octree meshes - [ ] ...
Simulation.gradientfor tri-axial anisotropyInterpolation: Changenearestfor properties to actual cell valuecompute1dinSimulation: get fast 1D results from an approximation of the 3D model (Layered computation using empymod. #302)emg3demg3d-gallery#2); e.g.SimulationRxFluxPoint(magnetic flux density)RxCurrentPoint(electric current density)