March 2022 OLCF User Conference Call: Machine Learning for HPC simulations: Using PyTorch, TensorFlow in Fortran, C, and C++ with SmartSim

The OLCF hosts monthly User Conference Calls. These calls are your opportunity to speak with center personnel to get the latest updates, express any concerns you may have, etc. No registration is required for this event.

Monthly Topic: Machine Learning for HPC simulations: Using PyTorch, TensorFlow in Fortran, C, and C++ with SmartSim

Speaker: Sam Partee, HPE

Abstract: While HPC support for AI has grown quickly over the past few years, AI support for HPC has not. Integrating Machine Learning into HPC simulations represents a challenging engineering task. SmartSim was created to ease this burden on scientists. SmartSim is an open source library that facilities the usage of popular ML libraries like PyTorch and TensorFlow/Keras in HPC simulations. SmartSim allows ML to be written in Python, where it belongs, and used in Fortran, C, and C++ applications at HPC scale.

In this talk, we will detail the SmartSim architecture, cover best practices, and provide examples of its usage across scientific domains: Climate/Weather, CFD, MD, and more. A container with built-in tutorials will be provided to users who wish to try SmartSim on their own system.

For more information about SmartSim prior to the talk, please see:
– GitHub: https://github.com/CrayLabs
– Documentation: https://www.craylabs.org/docs/overview.html

Presentation: ( slides | video )