PyPI

A non-compiled version of the tool can be found in PyPI. That usually points to the latest release of the tool and might not necessarily contain all the upgrades of the development branch.

To install from PyPI simply use:

pip install pyLowOrder

GPU support

A GPU supported version of the tool can be installed directly from PyPI using the following syntax:

pip install pyLowOrder[cuda12x]

Note that the cuda12x can be exchanged by:

• cuda in order to trigger the non-compiled version of cupy so that it builds specifically for your system.
• rocm-4-0 in order to deploy a version of cupy tailored for AMD GPUs using ROCM. ROCM is still experimental. Please be aware that the CUDA/ROCM version should match the ones in your system.

Optional modules

Additionally, the following optional modules can be installed from PyPI using the standard syntax of:

pip install pyLowOrder[module]

where module can be exchanged by:

• NN will install the dependencies related with the NN module of the tool.
• GPR will install the dependencies related with the GPR module of the tool.
• optionals will install some optional dependencies.

MPI possible installation issues

pyLOM uses mpi4py for MPI bindings. From mpi4py's PyPI page:

Installing mpi4py from its source distribution (available at PyPI) or Git source code repository (available at GitHub) requires a C compiler and a working MPI implementation with development headers and libraries.

If your machine doesn't have these (and it happens commonly in Windows systems), you need to install them. One valid approach through conda:

conda install -c conda-forge compilers mpi4py
pip install pyLowOrder