OpenSeesMP App

OpenSeesMP App#

Parallel OpenSees Runner for High-Performance Computing

The OpenSeesMP app is a Tapis-powered, SLURM-submitted application that runs OpenSees in MPI-parallel mode on the Stampede3 cluster at TACC. Unlike OpenSees-express, this app is built for large-scale, multi-core structural simulations that require distributed memory processing, job queuing, and access to advanced hardware.

You can find the template for this app in the TACC Repo

Key Features#

  • Runs OpenSeesMP — the MPI-enabled version of OpenSees

  • Executes on Stampede3 via SLURM scheduler

  • Leverages Tapis API for structured job submission and tracking

  • Accepts directory-based input with customizable .tcl scripts

  • Supports MPI execution using ibrun across 2+ nodes and 48+ cores per node

This app is composed of three main files:

  1. A Tapis app.json file defining the HPC job structure

  2. A profile.json that loads Stampede3 modules

  3. A tapisjob_app.sh script that launches the MPI-enabled binary

Summary: OpenSees at Scale with OpenSeesMP#

Component

Purpose

app.json

Describes inputs, parameters, HPC config, and output archiving

profile.json

Loads required software modules on Stampede3

tapisjob_app.sh

Executes OpenSeesMP using ibrun and user-defined .tcl script

This app is ideal for:

  • Large-scale simulations requiring distributed memory

  • Parallel analyses using OpenSeesMP

  • Users familiar with MPI and SLURM workflows

Tip

To debug or test your script before submitting to Stampede3, you can use the lightweight OpenSees-express or a JupyterHub terminal session with a smaller .tcl input file.

Contents