Running samples

Puhti

Setting up the repository and getting slurm scripts

Clone pencil-code and check out a version that compiles:

git clone https://github.com/pencil-code/pencil-code.git pencil-code-samples
cd pencil-code-samples
git checkout 900fd0187d67ecd318f6a846cd17d8f6fe1052dc
source sourceme.sh

Fix import pencil to import pencil_old in create_emftensors:

cd $PENCIL_HOME/samples/meanfield_special_e_tensor
sed -i 's/import pencil/import pencil_old/g' create_emftensors.py

Download run_sample_puhti.sh and start_sample_puhti.sh into the samples directory:

wget https://raw.githubusercontent.com/simo-tuomisto/meanfield_examples_doc/master/run_sample_puhti.sh
wget https://raw.githubusercontent.com/simo-tuomisto/meanfield_examples_doc/master/start_sample_puhti.sh

Before running the samples

All samples can be run in the same fashion. You can use the commands below, but first set the $SAMPLE_NAME environment variable:

• Isotropic alpha - export SAMPLE_NAME=isotropic_alpha
• Jouve 2008 Benchmark A - export SAMPLE_NAME=Jouve-2008-benchmarkA
• Jouve 2008 Benchmark B - export SAMPLE_NAME=Jouve-2008-benchmarkB
• Jouve 2008 Benchmark B - export SAMPLE_NAME=Jouve-2008-benchmarkB
• Steenbeck-Krause 1969 model1 - export SAMPLE_NAME=Steenbeck-Krause-1969-model1

Running the sample

Compile sample:

cd $PENCIL_HOME/samples/meanfield_special_e_tensor/$SAMPLE_NAME
module purge
module load StdEnv
module load hdf5/1.10.4-mpi 
pc_setupsrc
pc_build -f compilers/Intel_MPI REAL_PRECISION=double

Run start:

mkdir data
sbatch ../start_sample_puhti.sh

Create emftensors:

module purge
module load python-env
python ../create_emftensors.py .

Run sample:

module purge
sbatch ../run_sample_puhti.sh

diff data/time_series.dat reference.out.double