Julia language on triton

The Julia programming language is a high-level, high-performance dynamic programming language for technical computing, in the same space as e.g. MATLAB, Scientific Python, or R. For more details, see the web page https://julialang.org/ .

Interactive usage

Julia is available in the module system. By default the latest stable release is loaded:

module load julia
julia

Batch usage

Running Julia scripts as batch jobs is also possible. An example batch job is

Batch script for Julia job:

#!/bin/sh
#SBATCH --time=00:01:00
#SBATCH --mem=1G
module load julia
srun julia juliascript.jl

Number of threads to use

By default Julia uses up to 16 threads for linear algebra (BLAS) computations. However, the julia module on triton sets the environment variable OMP_NUM_THREADS to 1, so only a single thread is used. If you wish to use more threads than that (e.g. you have launched a batch job with multiple threads per task with the “#SBATCH –cpus-per-task=N” option), you can set the OMP_NUM_THREADS environment variable to some other value, or alternatively inside julia you can use the blas_set_num_threads() function.