Sparse cholesky (#72)

* [wip] initial supernodal cholesky

* catch up with main branch

* optimize allocations in supernodal ldlt

* add test for supernodal cholesky

* use amd ordering for benches

* implement amd ordering

* make slice groups more ergonomic

* add amd tests

* [wip] higher level api for sparse cholesky

* fix amd bug

* fix complex mul bug

* fix/refactor cholesky regularization

* hide ghost stuff

* optimize unnecessary temporaries

* rename FaerSparseError to FaerError

* add missing conj in permute function

* add full solver tests

* implement solve function and dynamic regularization

* upgrade pulp dep

* minor fixes

* impl Error for FaerError, faer bugfix

* reexport regularization

---------

Co-authored-by: sarah <>