- Pull your image
export SINGULARITY_DOCKER_ARCHITECTURE=ppc64le
singularity pull docker://centos:7- Get an allocation, for example:
srun --gres=gpu:4 -N 1 --mem=1T --time 1:00:00 -l --pty /bin/bash - Submit your payload
singularity exec centos_7.sif ls /- Pull your image
- If you are using an existing image, confirm that arch is ppc64le
- Get an allocation (for interactive allocation):
- Consider the size of your workload, do you need CPU? memory? GPU?
- These are all specified in the bsub request, here is a good1
- For example, here is a bsub to request allocation for 4 GPUS, via the normal queue, launching an interactive session using bash as terminal
srun --gres=gpu:4 -N 1 --mem=1T --time 1:00:00 -l --pty /bin/bash - Submit your payload
- Using interactive allocation, we are allocated a node - your interactions with singularity will be generally standard , just like you would in any environments
- Remember you do not have sudo access on this cluster, this might impact your workflow
singularity exec centos_7.sif ls /- Workflow is similar to interactive allocation, however we need to create an application profile prior to submitting the workload
- This mode might be better for headless / non-interactive workloads as the resources can be more easily allocated to the workload
- Create an application profile2
- For example
Begin Application
NAME = singapp
CONTAINER = singularity[image(centos_7.sif)]
DESCRIPTION = Singularity User Service
End Application- Submit the workload via the application profile3
- For example
sbatch singapp ./myjob.sh(https://www.ibm.com/support/knowledgecenter/SSWRJV_10.1.0/lsf_command_ref/bsub.man_top.1.html)↩
(https://www.ibm.com/support/knowledgecenter/en/SSWRJV_10.1.0/lsf_container/lsf_singularity_config.html).↩
(https://www.ibm.com/support/knowledgecenter/en/SSWRJV_10.1.0/lsf_container/lsf_singularity_use.html).↩