Container Singularity Ubuntu with Yade.

Dir structure:

.
|-- README.md
|-- def
|   |-- yade_2018.02b.def  	<- Ubuntu 18.04 + Yade 2018
|   `-- yade_2020.01a-12.def	<- Ubuntu 20.04 + Yade 2020
`-- sbatch
    `-- job.sh			<- Sbatch Sample

2 directories, 4 files

Building...

Clone this repository:

git clone https://github.com/lmagdanello/yade_singularity.git

Access the main directory:

cd yade_singularity

And build w/ singularity:

sudo singularity build --notest def/yade.sif def/yade_<version>.def

This process will generate the image of the container with the name of yade.sif.

The examples below were made using def 2018.02b with Ubuntu 18.04 and Yade 2018, but the same is true for version 2020 with Ubuntu 20.04.

Running...

You can run the image itself, without using singularity:

$ def/yade.sif
Container was created
Arguments received:
Welcome to Yade 2018.02b
Warning: no X rendering available (see https://bbs.archlinux.org/viewtopic.php?id=13189)
TCP python prompt on localhost:9000, auth cookie `ueasks'
XMLRPC info provider on http://localhost:21000
[[ ^L clears screen, ^U kills line. F8 plot. ]]

In [1]:

Or using singularity (run, exec and shell):

$ singularity run def/yade.sif
Container was created
Arguments received:
Welcome to Yade 2018.02b
Warning: no X rendering available (see https://bbs.archlinux.org/viewtopic.php?id=13189)
TCP python prompt on localhost:9000, auth cookie `ucysse'
XMLRPC info provider on http://localhost:21000
[[ ^L clears screen, ^U kills line. F8 plot. ]]

In [1]:


$ singularity exec def/yade.sif yade --version
Yade version: 2018.02b


$ singularity shell def/yade.sif
Singularity> yade --version
Yade version: 2018.02b
Singularity> cat /etc/lsb-release
DISTRIB_ID=Ubuntu
DISTRIB_RELEASE=18.04
DISTRIB_CODENAME=bionic
DISTRIB_DESCRIPTION="Ubuntu 18.04.5 LTS"
Singularity>

You can also pass arguments to the container, like other scripts. For example, given a python script that calculates the square root of 144. Let's run it with yade:

Note: -x is to suppress Yade's iterative terminal;

# sqrt.py
import math
print(math.sqrt(144))

$ def/yade.sif -x sqrt.py
Container was created
Arguments received: -x sqrt.py
Welcome to Yade 2018.02b
Warning: no X rendering available (see https://bbs.archlinux.org/viewtopic.php?id=13189)
TCP python prompt on localhost:9000, auth cookie `ycdaku'
XMLRPC info provider on http://localhost:21000
Running script sqrt.py
12.0


$ singularity run def/yade.sif -x sqrt.py
Container was created
Arguments received: -x sqrt.py
Welcome to Yade 2018.02b
Warning: no X rendering available (see https://bbs.archlinux.org/viewtopic.php?id=13189)
TCP python prompt on localhost:9000, auth cookie `eskyua'
XMLRPC info provider on http://localhost:21000
Running script sqrt.py
12.0


$ singularity exec def/yade.sif yade -x sqrt.py
Welcome to Yade 2018.02b
Warning: no X rendering available (see https://bbs.archlinux.org/viewtopic.php?id=13189)
TCP python prompt on localhost:9000, auth cookie `scksau'
XMLRPC info provider on http://localhost:21000
Running script sqrt.py
12.0


$ singularity shell def/yade.sif
Singularity> yade -x sqrt.py
Welcome to Yade 2018.02b
Warning: no X rendering available (see https://bbs.archlinux.org/viewtopic.php?id=13189)
TCP python prompt on localhost:9000, auth cookie `esaudy'
XMLRPC info provider on http://localhost:21000
Running script sqrt.py
12.0

Testing...

Yade has some regression tests built into the software. These same tests are covered by the container, you can run them to validate Yade as follows:

$ singularity test def/yade.sif

Help and Inspect

$ singularity run-help def/yade.sif
        This is a container that uses Ubuntu to prepare the environment for installing Yade (https://yade-dem.org/), which is an extensible open-source framework for discrete numerical models, focused on Discrete Element Method.

$ singularity inspect def/yade_2018.02b.sif
Author: leonardo.araujo@atos.net
Version: v0.0.2
org.label-schema.build-date: Tuesday_24_November_2020_15:4:2_-03
org.label-schema.schema-version: 1.0
org.label-schema.usage: /.singularity.d/runscript.help
org.label-schema.usage.singularity.deffile.bootstrap: docker
org.label-schema.usage.singularity.deffile.from: ubuntu:18.04
org.label-schema.usage.singularity.runscript.help: /.singularity.d/runscript.help
org.label-schema.usage.singularity.version: 3.5.2-1.1.el7


$ singularity inspect def/yade_2020.01a-12.sif
Author: leonardo.araujo@atos.net
Version: v0.0.2
org.label-schema.build-date: Tuesday_24_November_2020_17:13:34_-03
org.label-schema.schema-version: 1.0
org.label-schema.usage: /.singularity.d/runscript.help
org.label-schema.usage.singularity.deffile.bootstrap: docker
org.label-schema.usage.singularity.deffile.from: ubuntu:20.04
org.label-schema.usage.singularity.runscript.help: /.singularity.d/runscript.help
org.label-schema.usage.singularity.version: 3.5.2-1.1.el7

Slurm

This container can be executed inside a sbatch, in the Slurm queue.

E.g .:

# sbatch_file_example.sh:

#!/bin/bash
#SBATCH --nodes=<Total of Nodes>
#SBATCH --ntasks-per-node=<Number of Tasks/Node>
#SBATCH --ntasks=<Total Tasks>
#SBATCH -p <Partition> 
#SBATCH -J <Job Name> 

echo $SLURM_JOB_NODELIST
nodeset -e $SLURM_JOB_NODELIST
echo
singularity run def/yade.sif -x sqrt.py

The output will look like:

node[1068-1071]
node1068 node1069 node1070 node1071

Container was created
Arguments received: -x sqrt.py
TCP python prompt on localhost:9000, auth cookie `yedkua'
Welcome to Yade 2018.02b
Warning: no X rendering available (see https://bbs.archlinux.org/viewtopic.php?id=13189)
XMLRPC info provider on http://localhost:21000
Running script sqrt.py
12.0

This sample sbatch is available at: sbatch/job.sh

leonardo.araujo@atos.net ;)