Description: Scripting data-oriented computational workflows for parallel execution requires careful consideration of program logic, interfacing tools, hardware and HPC environments. Parsl (Parallel Scripting Library) is a scripting API written in Python that provides a common interface to abstract various parallel scripting models that can be easily scaled and executed on various data center resources. Some common patterns are multi-threading, multi-processing, and coarse-grained parallelism at the level of the resource manager. Parsl enhances parallel execution capabilities within Python and its rich ecosystem of open-source libraries by allowing these models to scale to large, multi-node computer systems.

This seminar will introduce HPC users to the Parsl scripting library and how it can be used to abstract, parallelize, execute and monitor computational workflows. Examples representative of common use cases on the HPC’s will be discussed to demonstrate parallel programming models supported in Parsl and how they can be used to scale and orchestrate computation on various data center resources.

Presenter(s): Andrew Simin, GDIT / PET Location: Webcast Date & Time: January 14, 2025, 2:00p - 3:30p ET

Controlled by: DoD HPCMP Controlled by: PET Program CUI Category: OPSEC Limited Dissemination Control: FEDCON POC: Mr. Ronald Hedgepeth, pet@hpc.mil

---

CUI