Introduction to High-Performance Parallel Distributed Computing using Chapel, UPC++ and Coarray Fortran

July 26 - July 27

Calendar

When:

July 26, 2023 – July 27, 2023 all-day

2023-07-26T00:00:00-04:00

2023-07-28T00:00:00-04:00

Contact:

Paul Hargrove

Training Event

training ideas IDEAS Productivity Agile Project Management Kanban

Schedule:

This two-day tutorial will run July 26-27, 12:00pm-3:25pm ET.

Abstract:

A majority of HPC system users use scripting languages such as Python to prototype their computations, coordinate their large executions, and analyze the data resulting from their computations. Python is great for these many uses, but it frequently falls short when significantly scaling up the amount of data and computation, as required to fully leverage HPC system resources. In this tutorial, we show how example computations such as heat diffusion, k-mer counting, file processing, and distributed maps can be written to efficiently leverage distributed computing resources in the Chapel, UPC++, and Fortran parallel programming models. This tutorial should be accessible to users with little-to-no parallel programming experience, and everyone is welcome. A partial differential equation problem will be shown in all three programming models along with performance and scaling results on big machines. Attendees will be shown how to compile and run these programming examples, and provided opportunities to experiment with different parameters and code alternatives while being able to ask questions and share their own observations. Come join us to learn about some productive and performant parallel programming models!

Current OLCF users with access to Frontier will be able to access a reservation on Frontier to work the examples. Current NERSC users will be able to use Perlmutter. Training accounts on Perlmutter are available for participants who do not have access to either Frontier or are not NERSC users. The examples will also be available in a Docker container and a cloud-based virtual desktop environment for access by any attendee.

Keywords:

Basic and introductory topics for expanding broader engagement
Software engineering for portable performance and scalability
Parallel programming methods, models, languages and environments
Clusters and distributed systems

Presenters

Dr. Amir Kamil is a Lecturer in Electrical Engineering and Computer Science at the University of Michigan and Visiting Faculty at Lawrence Berkeley National Laboratory. He earned his Ph.D in Computer Science from UC Berkeley. Prior to joining the University of Michigan faculty, Dr. Kamil was a postdoctoral scholar and lecturer at UC Berkeley and a postdoctoral researcher at Lawrence Berkeley National Laboratory. He remains affiliated with the Computer Languages and Systems Software group at LBNL. His research interests include compiler analysis, optimization, and programming models for high-performance computing, diversity and inclusion in computing, and software tools for computer-science education.

Dr. Damian Rouson is the Group Lead for the Computer Languages and Systems Software Group at Lawrence Berkeley National Laboratory. His research focuses on accelerating computational science and engineering applications using partitioned global address space programming models and emerging hardware, including GPUs and AI accelerators. He was awarded a 2020 Better Scientific Software Fellowship by the Exascale Computing Project. He leads the development of the OpenCoarrays parallel runtime library. He co-authored the textbook Scientific Software Design: The Object-Oriented Way (Cambridge Press 2011) and has taught related courses and tutorials at universities, labs and conferences. He has held staff and faculty appointments at universities and labs in the U.S. and Europe. He received a B.S. from Howard University and M.S./Ph.D. from Stanford University, all in Mechanical Engineering. He founded the software consultancy Sourcery, Inc. and the California not-for-profit Sourcery Institute.

Prof. Michelle Mills Strout has led the Chapel project at Hewlett Packard Enterprise since January 2021. She is also an Affiliate Professor at the University of Arizona in Computer Science, where she was a full Professor for 2015 through 2022. Michelle has been teaching in Computer Science since 2000. Michelle’s main research area is high performance computing and her research interests include compilers, scientific computing, and software engineering. She especially enjoys collaborating with domain scientists to understand their high performance computing needs and how parallel programming models can help. Dr. Strout earned her Ph.D. at the University of California, San Diego in 2003 with Jeanne Ferrante and Larry Carter as co-advisors. Michelle has two adult children who she hangs out with whenever possible. In her free time, Michelle enjoys hiking, boating, downhill skiing, and reading.