Skip to content

Latest commit

 

History

History
216 lines (166 loc) · 8.95 KB

README.md

File metadata and controls

216 lines (166 loc) · 8.95 KB
Raw

T8code.jl

All Tests Status Aqua QA License: MIT

T8code.jl is a Julia package that wraps t8code, a C/C++ library to manage multiple connected adaptive quadtrees or octrees in parallel and with possibly mixed element types.

Installation

If you have not yet installed Julia, please follow the instructions for your operating system. T8code.jl works with Julia v1.6 and later.

T8code.jl is a registered Julia package. Hence, you can install it by executing the following commands in the Julia REPL:

julia> import Pkg; Pkg.add(["T8code", "MPI"])

With this command, you install both T8code.jl and MPI.jl. Currently, T8code.jl supports only builds of the C/C++ library t8code with MPI support, so you need to initialize MPI appropriately as described in the Usage section below.

T8code.jl depends on the binary distribution of the t8code library, which is available in the Julia package T8code_jll.jl and which is automatically installed as a dependency. The binaries provided by T8code_jll.jl support MPI and are compiled against the default MPI binaries of MPI.jl. At the time of writing, these are the binaries provided by MicrosoftMPI_jll.jl on Windows and MPICH_jll.jl on all other platforms.

By default, T8code.jl provides pre-generated Julia bindings to all exported C interface functions of the underlying t8code library. If you want/need to generate new bindings, please follow the instructions in the dev folder and copy the generated files to the appropriate places in src.

Using a custom version of MPI and/or t8code

t8code needs to be compiled against the same MPI implementation used by MPI.jl. Thus, if you want to configure MPI.jl to not use the default MPI binary provided by JLL wrappers, you also need to build t8code locally using the same MPI implementation. This is typically the situation on HPC clusters. If you are just using a single workstation, the default installation instructions should be sufficient.

T8code.jl allows using a t8code binary different from the default one provided by T8code_jll.jl. To enable this, you first need to obtain a local binary installation of t8code. Next, you need to configure MPI.jl to use the same MPI implementation used to build your local installation of t8code, see the documentation of MPI.jl. At the time of writing, this can be done by first setting up the Preferences.jl setting containing the path to your local build of the shared library of t8code.

julia> using Preferences, UUIDs

julia> set_preferences!(
           UUID("d0cc0030-9a40-4274-8435-baadcfd54fa1"), # UUID of T8code.jl
           "libt8" => "/path/to/your/libt8.so", force = true)

julia> set_preferences!(
           UUID("d0cc0030-9a40-4274-8435-baadcfd54fa1"), # UUID of T8code.jl
           "libp4est" => "/path/to/your/libp4est.so", force = true)

julia> set_preferences!(
           UUID("d0cc0030-9a40-4274-8435-baadcfd54fa1"), # UUID of T8code.jl
           "libsc" => "/path/to/your/libsc.so", force = true)

Alternatively, you can use the convenience functions set_library_t8code!, set_library_p4est! and set_library_sc! to set the paths:

julia> using T8code

julia> T8code.set_library_t8code!("/path/to/your/libt8.so")
[ Info: Please restart Julia and reload T8code.jl for the library changes to take effect

julia> T8code.set_library_p4est!("/path/to/your/libp4est.so")
[ Info: Please restart Julia and reload T8code.jl for the library changes to take effect

julia> T8code.set_library_sc!("/path/to/your/libsc.so")
[ Info: Please restart Julia and reload T8code.jl for the library changes to take effect

If all libraries are located in the same directory and have the default names (libt8.so, libp4est.so and libsc.so or other file endings according to your system), you can also set all three preferences by only specifying the directory:

julia> using T8code

julia> T8code.set_libraries_path!("/path/to/your/lib/directory/")
[ Info: Please restart Julia and reload T8code.jl for the library changes to take effect
[ Info: Please restart Julia and reload T8code.jl for the library changes to take effect
[ Info: Please restart Julia and reload T8code.jl for the library changes to take effect

To delete the preferences again, you can call T8code.set_libraries_path!() or for each library T8code.set_library_t8code!(), T8code.set_library_p4est!() and T8code.set_library_sc!(), respectively.

Note that you should restart your Julia session after changing the preferences.

Next, you need to set up the preferences for MPI, which can be done by

julia> using MPIPreferences

julia> MPIPreferences.use_system_binary()

if you use the default system MPI binary installation to build p4est.

Currently, custom builds of t8code without MPI support are not supported.

Usage

T8code.jl provides Julia bindings for three libraries:

The bindings for t8code are imported by default to the global namespace. They all start wit the prefixes t8_* or T8_*. The bindings for p4est and libsc are not exported and must be qualified by their full namespace. I.e.:

T8code.Libt8.sc_some_function(...)
T8code.Libt8.p4est_some_function(...)
T8code.Libt8.p8est_some_function(...)

A typical initialization of T8code.jl may look as follows:

using MPI
using T8code

using T8code.Libt8: sc_init
using T8code.Libt8: sc_finalize
using T8code.Libt8: SC_LP_ESSENTIAL
using T8code.Libt8: SC_LP_PRODUCTION

# Initialize MPI. This has to happen before we initialize sc or t8code.
mpiret = MPI.Init()

comm = MPI.COMM_WORLD

# Initialize the sc library, has to happen before we initialize t8code.
sc_init(comm, 1, 1, C_NULL, SC_LP_ESSENTIAL)

# Initialize t8code with log level SC_LP_PRODUCTION. See sc.h for more info on the log levels.
t8_init(SC_LP_PRODUCTION)

# Some more code [...]

At the end of the application finalize in order to check for unbalanced references resp. unfreed memory.

sc_finalize()

See examples/ for more information on how to use T8code.jl.

Issues

It is highly recommended to use Julia v1.9 or newer. Older versions showed some issues like extremely long computing times (basically hanging) when initializing some structs defined by t8code. Nevertheless, most functionality has been tested with older Julia versions and should work just fine.

The compat entry is set to julia = "1.6" for broader support of other packages using T8code.jl. In the future, this workaround will be dropped when Julia v1.9 (or newer) is more widely established in the community.

Authors

T8code.jl is mainly maintained by Johannes Markert (German Aerospace Center (DLR), Germany) and Joshua Lampert (University of Hamburg, Germany). It is an adapted fork from P4est.jl maintained by Joshua Lampert, Michael Schlottke-Lakemper (RWTH Aachen University, Germany), and Hendrik Ranocha (Johannes Gutenberg University Mainz, Germany).

The t8code library itself is written by Johannes Holke and Carsten Burstedde, and others.

License and contributing

T8code.jl is licensed under the MIT license (see LICENSE.md). t8code itself is licensed under the GNU General Public License, version 2.