Run DS Agnostic App#
Notebook Demo: Submitting OpenSees HPC Jobs with the Agnostic App
This notebook demonstrates how to submit OpenSees analyses on DesignSafe using the designsafe-agnostic-app, a flexible Tapis application that supports OpenSees, OpenSeesPy, and MPI-enabled workflows.
While OpenSees-specific apps exist on DesignSafe, this demo focuses on the agnostic execution path to help you understand exactly how your analysis is launched on HPC systems — without hiding details behind presets.
Why This Demo Exists#
This notebook represents a typical structural engineering workflow on DesignSafe:
OpenSees (Tcl or OpenSeesPy)
Optional MPI execution
Script-driven analysis
Structured output handling
The purpose is not to teach OpenSees modeling, but to show how an OpenSees analysis fits into the general DesignSafe execution model.
By using the agnostic app, the demo makes all assumptions explicit — from the executable, to MPI usage, to directory layout.
What to Watch For#
As you go through the notebook, pay close attention to:
How the input directory defines the OpenSees run
How the job command mirrors a local OpenSees invocation:
[ibrun] OpenSeesMP <input.tcl>
or
python OpenSeesPy_script.py
How MPI behavior is explicitly enabled or disabled
How environment modules and Python packages are set up
How outputs are organized for:
large simulations
post-processing
batch or parametric studies
You’ll notice that only a small number of inputs differ from a non-OpenSees job — the execution flow itself is identical.
What This Demo Is (and Is Not)#
This notebook is:
✔ A complete OpenSees job submission example
✔ A transparent view of how OpenSees runs on HPC systems
✔ A reusable template for advanced OpenSees workflows
It is not:
❌ A full OpenSees modeling tutorial
❌ An MPI theory lesson
❌ A benchmark or performance comparison
The focus is on execution mechanics, not structural modeling theory.
Key Takeaway#
OpenSees on DesignSafe follows the same execution model as any other HPC job. Once you understand this notebook, you can scale from a single analysis to large automated studies with confidence.