-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathcreate-study.bash
More file actions
45 lines (37 loc) · 8.4 KB
/
create-study.bash
File metadata and controls
45 lines (37 loc) · 8.4 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
#!/bin/bash
# This function writes a SLURM script. We can call it with different parameter
# settings to create different experiments
function write_script
{
STUDY_NAME=$(printf 'n%d' ${N})
DIR_NAME=$(printf '%s/n%02dppn%02d' ${STUDY_NAME} ${NODES} ${NPERNODE})
echo $DIR_NAME
mkdir -p $DIR_NAME
cat <<_EOF_ > ${DIR_NAME}/run.slurm
#!/bin/bash
#SBATCH --job-name=poisson
#SBATCH --output=slurm.out
#SBATCH --error=slurm.err
#SBATCH --partition=high_mem
#SBATCH --qos=short+
#SBATCH --constraint=hpcf2018
#SBATCH --nodes=${NODES}
#SBATCH --ntasks-per-node=${NPERNODE}
#SBATCH --time=00:55:00
#SBATCH --mem=10G
#SBATCH --exclusive
mpirun ../../poisson ${N} 1.0e-6 50000
_EOF_
}
# For each problem size, we'll set up the experiment with
# 1, 2, 4, 8, 16, and 32 processes on 1 and 2 nodes
for N in 16384
do
for NPERNODE in 1 2 4 8 16 32
do
for NODES in 8 16
do
write_script
done
done
done