Julia

Julia is a dynamic programming language that offers high performance while being easy to learn and develop code in.

This section contains demos (in the form of scripts and notebooks) and how-to guides for doing various things with Julia on NREL HPC environments.

Available modules

Swift	Vermilion	Kestrel (CPU)	Kestrel (GPU)
julia/1.6.2-ocsfign
julia/1.7.2-gdp7a25	julia/1.7.2	julia/1.7.2
		julia/1.8.5-generic-linux
		julia/1.10.0-gcc
			julia/1.10.4

Julia 1.9.x does not work well on Sapphire Rapids

We advise against installing and using Julia 1.9.x on Kestrel as packages can fail to precompile and result in a segmentation fault. This is a known issue with Julia 1.9.x on Sapphire Rapids processors, possibly due to an LLVM issue. Julia 1.10 will be installed as a module once a stable release is available. Until then, please use Julia 1.7 or 1.8.

Contents

1. Installing Julia
2. Tour of Julia
3. Parallel Computing in Julia
4. Calling Python, C, and FORTRAN from Julia

Demo Scripts and Notebooks

The following scripts and notebooks are available on the master branch of NREL/HPC to download and run,

• Julia Tour
• Julia Parallel Computing
• Calling Python, C, and FORTRAN from Julia
• PyJulia -- calling Julia from Python (PyJulia_Demo.ipynb)
• Integrating mpi4py and MPI.jl
  • Hello World
  • Self contained approximation of pi
  • Approximation of pi as "library" call
  • Comparing different Control Variates to approximate pi -- uses MPI Split
  • Example batch script for all of the above

Requirements and Installation

Running the demos requires the python modules mpi4py and julia. For details on installing these modules, see the 'Environment Setup' section of the README found in the demos/scripts directory.

For more information on mpi4py, see the mpi4py documentation

For more information on PyJulia, see the PyJulia documentation.