Installation

FEMTISE.jl requires Julia 1.10 or newer. It is not yet registered in the General registry, so it is installed directly from GitHub.

Install a released version (recommended)

Pinning a release guarantees that your scripts keep producing the same results when the package evolves. In the Julia REPL press ] to enter the package manager, then:

(@v1.10) pkg> activate .                                          # project of your calculation
(my_project) pkg> add https://github.com/mendzmartin/FEMTISE.jl#v0.1.1

The #v0.1.1 suffix selects the release tag. The exact version is recorded in the Manifest.toml of your project; keep that file next to your scripts. The list of releases and what changed in each one is in the Changelog.

Equivalently, from a script:

using Pkg
Pkg.activate(".")
Pkg.add(url = "https://github.com/mendzmartin/FEMTISE.jl", rev = "v0.1.1")

Install the development version

To follow the main branch (latest changes, possibly unreleased):

(my_project) pkg> add https://github.com/mendzmartin/FEMTISE.jl

pkg> update then fetches the newest commit. Read the Unreleased section of the Changelog before updating.

Work on the package itself

If you want to modify FEMTISE.jl, clone it and add it in development mode:

git clone https://github.com/mendzmartin/FEMTISE.jl.git ~/src/FEMTISE.jl
(my_project) pkg> dev ~/src/FEMTISE.jl

Changes to the source are picked up the next time using FEMTISE is executed. The test suite runs with pkg> test FEMTISE.

Running scripts and using threads

Save your calculation in a file, e.g. my_calculation.jl:

using Pkg
Pkg.activate(@__DIR__)      # the project where FEMTISE was added
using FEMTISE
# ... your calculation ...

and run it with

julia my_calculation.jl          # serial
julia -t 4 my_calculation.jl     # 4 threads

Threads are used by the post-processing functions and by the conversion of finite element functions to arrays. The eigensolver itself is serial except for the multithreaded BLAS.

Checking the installation

julia> using FEMTISE
julia> ?eigen_values_and_eigen_vectors     # help mode shows the docstring

Then follow the Quick start.