Moritz Sallermann

Micromagnetics

2

Implements the micromagnetic Hamiltonian. - Closes #362 Todo: - [ ] implement gradients - [ ] fully implement CUDA backend - [ ] implement unit tests - [ ] implement...

Provide a Tutorial

2

It might be nice to provide a "step-by-step" tutorial (maybe as a separate repository, something like https://github.com/spirit-code/spirit-tutorial). The overlap with our existing documentation should be kept minimal and the tutorial...

Use clang-format

2

Using clang-format seems to be a good way to enforce uniform formatting across the code.

Bug: When a vacancy is set and then removed mu_s stays zero

2

When an atom type different from a vacancy is set the magnetic moment should be restored.

If the geometry is changed too drastically while DDI is activated problems arise. This seems to only happen on the CUDA backend. Steps to reproduce: 1. Perform a few iterations...

When using the newest cuda toolkit (version 10.2.89) together with the `ddi_method fft` the code crashes while calling functions from `Vectormath.cu` (e.g. `Vectormath::fill`). The error message is: ```Vectormath.cu:234 in function...

If additional defects are specified via the `n_defects` option, the magnetic moment gets always set to `0`. If the defect is, however, not a vacancy this is wrong and should...

Benchmark

1

I think it would be good if Spirit included a small benchmarking executable, e.g. `./benchmark` lying next to the `./spirit` executable after compilation. Command line arguments could be used when...

Setting vacancies is not working

1

When setting vacancies in the UI, to e.g. create a disk, the remaining non-vacant spins do not react to changes of the spin configuration and basically become unaccessible. Steps to...

Started to implement parallelization over single spins instead of lattice basis cells. Closes #420.