[
https://issues.apache.org/jira/browse/AIRAVATA-3578?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Eroma updated AIRAVATA-3578:
----------------------------
Description:
With this new feature, when users are providing inputs for an experiment, we
need to be able to have them added in to the job script as script commands.
e.g. in SLURM job script we should be able to add parameters such as
#SBATCH -c=12 (which is cores per task) and
#SBATCH -n=2 (which is tasks per node. Which is currently the Total Core count
of the experiment) by taking the input values provided by the user at create
experiment.
When user provides number of tasks per node and number of cores per task as
input parameters, we need to get them to the job script as Slurm parameters.
It was discussed to do this through groovy scripts.
was:
With this new feature, when users are providing inputs for an experiment, we
need to be able to have them added in to the job script as script commands.
e.g. in SLURM job script we should be able to add parameters such as
#SBATCH -c=12 (which is cores per task) and
#SBATCH -n=2 (which is tasks per node) by taking the input values provided by
the user at create experiment.
When user provides number of tasks per node and number of cores per task as
input parameters, we need to get them to the job script as Slurm parameters.
It was discussed to do this through groovy scripts.
> Groovy scripts to incorporate with user's input data
> ----------------------------------------------------
>
> Key: AIRAVATA-3578
> URL: https://issues.apache.org/jira/browse/AIRAVATA-3578
> Project: Airavata
> Issue Type: New Feature
> Components: Airavata System, helix implementation
> Reporter: Eroma
> Assignee: Dinuka Subashana De Silva
> Priority: Major
>
> With this new feature, when users are providing inputs for an experiment, we
> need to be able to have them added in to the job script as script commands.
> e.g. in SLURM job script we should be able to add parameters such as
> #SBATCH -c=12 (which is cores per task) and
> #SBATCH -n=2 (which is tasks per node. Which is currently the Total Core
> count of the experiment) by taking the input values provided by the user at
> create experiment.
> When user provides number of tasks per node and number of cores per task as
> input parameters, we need to get them to the job script as Slurm parameters.
>
> It was discussed to do this through groovy scripts.
--
This message was sent by Atlassian Jira
(v8.20.7#820007)
