moller
Last updated: June 3, 2026
moller is a tool designed to support high-throughput computing. It generates batch job scripts that allow multiple programs to run in parallel under various computational conditions, such as parameter parallelism.
Usage Environment
- Node Group A, B, C, Login Node
| Version | Kyushu University Members | Academic Research Institution Members | Private Users |
|---|---|---|---|
| 1.0.1 | ○ | ○ | ○ |
How to Use
Environment Settings
$ module load moller/1.0.1
Sample Input File
Sample input files can be found at /home/rist/moller-1.0.1/sample. Below is a simple example input file (input.yaml) that outputs "hello world":
name: testjob
description: Sample task file
platform:
system: genkai
queue: a-batch
node: 1
elapsed: 00:05:00
prologue:
code: |
module load intel impi
jobs:
start:
parallel: false
run: |
echo "start..."
hello:
description: hello world
node: [1,1]
run: |
echo "hello world." > result.txt
sleep 2
epilogue:
code: |
echo "done."
date
Execution Example
Run moller with the input file input.yaml as a command-line argument to generate the job script job.sh:
$ moller -o job.sh input.yaml
Create a working directory for each job to be run in bulk (e.g., dataset-0001 through dataset-0020), and list the directories in a file called list.dat.
Submit the batch job:
$ pjsub job.sh
References
The moller user manual is located at the following path on the Genkai frontend node. Please refer to the manual for detailed usage instructions:
- /home/rist/moller-1.0.1/moller-usersguide-1.0.1-genkai-en.pdf (English)
- /home/rist/moller-1.0.1/moller-usersguide-1.0.1-genkai.pdf (Japanese)
For more information about moller: