Run DS OpenSeesPy App

Run DS OpenSeesPy App#

Notebook Demonstration Using the ‘designsafe-openseespy-s3’ App

This Jupyter notebook demonstrates how to submit, configure, and monitor batch OpenSeesPy jobs on DesignSafe using the ‘designsafe-openseespy-s3’ Tapis application.

Rather than focusing on OpenSeesPy modeling itself, the goal here is to help you understand the job-submission workflow:

  • how your files are staged,

  • how execution happens on Stampede3 under SLURM,

  • how portal inputs map to runtime behavior, and

  • how to diagnose and debug jobs after submission.

By the end of this notebook, you should feel comfortable adapting the same pattern to your own OpenSeesPy scripts, whether they are serial, MPI-enabled, or part of a small parameter study.

What This Notebook Covers#

This notebook walks through a minimal but complete workflow for running OpenSeesPy on DesignSafe:

  • Preparing an input directory containing a Python script and supporting files

  • Selecting appropriate portal inputs (script name, MPI flag, resources)

  • Submitting a job programmatically from Jupyter

  • Understanding how the app:

    • stages inputs,

    • loads modules,

    • installs Python packages,

    • injects the TACC-compiled OpenSeesPy library, and

    • launches your script (serial or MPI)

  • Locating outputs and interpreting the job summary log

The emphasis is on execution mechanics, not OpenSees theory or model formulation.

What This Notebook Does Not Cover#

To keep the focus clear, this notebook intentionally does not cover:

  • Writing or validating OpenSeesPy models

  • Advanced MPI domain decomposition strategies

  • Large-scale production workflows

  • Custom SLURM scripting

Those topics are addressed elsewhere in the OpenSees-on-DesignSafe documentation.

When to Use This Pattern#

The workflow demonstrated here is well suited for:

  • Training and tutorials

  • Small-to-moderate OpenSeesPy analyses

  • MPI-enabled OpenSeesPy scripts using ‘mpi4py’

  • Parameter studies and lightweight automation

  • Users who want no direct interaction with SLURM scripts

If you later need more control over execution logic or scaling behavior, the same concepts shown here apply when moving to more specialized Tapis apps or custom workflows.

How to Read This Notebook#

You can follow this notebook in two ways:

  • Linearly, as a step-by-step tutorial, or

  • Selectively, using it as a reference when submitting your own jobs

Each section explains why a setting exists, not just what value to use, so you can confidently adapt the workflow to new problems.

Contents