on CPUs or GPUs, with local or canonical Maxwellians, and the Maxwellian BC. It also works with parallelized species, with one caveat:
Canonical Maxwellians require additional inter-species communication to enforce n_e=n_i, and this step is not yet supported with parallelizeSpecies=true. So one has to run the initialization with parallelizeSpecies=false, and then start another simulation with parallelizeSpecies=true and reading in the ICs with the fromFile options.
We can species parallelize this step in the future, but it's both tricky and not urgent right now.
on CPUs or GPUs, with local or canonical Maxwellians, and the Maxwellian BC. It also works with parallelized species, with one caveat: Canonical Maxwellians require additional inter-species communication to enforce n_e=n_i, and this step is not yet supported with parallelizeSpecies=true. So one has to run the initialization with parallelizeSpecies=false, and then start another simulation with parallelizeSpecies=true and reading in the ICs with the fromFile options.
We can species parallelize this step in the future, but it's both tricky and not urgent right now.