Julia example

Assumptions

We will assume that your code and data are organized as follows:

/home/netid/project/
            ├── test.jl
            ├── data.csv
            ├── etc.

Quick run

cd $HOME/project
srun julia test.jl

with

# test.jl
x = rand(1000);
function sum_global()
           s = 0.0
           for i in x
               s += i
           end
           return s
       end;
@time sum_global()
@time sum_global()

which yields

[lv39@cbsulogin project]$ srun  julia test.jl
  0.016342 seconds (7.42 k allocations: 303.106 KiB, 98.58% compilation time)
  0.000143 seconds (3.49 k allocations: 70.156 KiB)

SBATCH script

Create a new file, called main.sbatch, with the following content:

Top

Use the default options from the SBATCH example, and paste them into main.sbatch.

Adjustments

If you are using any parallel processing Julia, adjust the node, tasks-per-node, and cpus-per-task. For instance, if you intend to run with 4 parallel threads, use

#SBATCH --cpus-per-task=4

Bottom

Now for the functional part, which is more or less the same as the srun command above, but with some additional options:

# Modules aren't available yet for Julia, consider using `juliaup`. 

julia --threads ${SLURM_CPUS_PER_TASK} test.jl

You should also consider using Julia-specific project features. This simple page is not the right place to teach that.