Name: Justin Wozniak
Pronouns:
Biography:
Justin M Wozniak is a Computer Scientist at Argonne National Laboratory and a Scientist-At-Large at the University of Chicago. He received a PhD in Computer Science & Engineering from the University of Notre Dame in 2008. He currently participates in the DOE Exascale Computing Project on multiple projects in cancer applications, deep learning, and large-scale workflows. His workflow system, Swift/T, won an R&D 100 award in 2018, and he was a team member for the CANDLE project which won an R&D 100 award in 2023. He is the co-founder and co-organizer of the XLOOP workshop at SC and the ERROR workshop at eScience.
Institution/Lab: Argonne National Laboratory
Website: https://www.anl.gov/profile/justin-m-wozniak
SRP Collaboration Topic/Title: HPC Workflow Performance
Field or research area: HPC, Workflows
Please select all the topical areas that apply to your project:
Computer Science (i.e., architectures, compilers/languages, networks, workflow/edge, experiment automation, containers, neuromorphic computing, programming models, operating systems, sustainable software); High-Performance Computing
Brief Abstract:
High Performance Computing (HPC) workflows combine physical simulations with machine learning and other analysis to address scientific problems in modeling the spread of COVID-19, designing precision cancer treatments, and other areas. The Swift/T workflow language is designed to run on the largest available supercomputers and integrate complex, parallel software packages under the control of a high-level dataflow language. Optimizing and debugging such workflows is difficult. There are range of tools developed by the computer science community to analyze the performance and correctness of such software systems. In this project, the team will apply one or more such systems to Swift/T workflows in a scientific application area. The team will run Swift/T in a simulated environment using a distributed computing simulator such as SimGrid/SMPI to evaluate correctness and performance. Technical details: Swift/T is a parallel programming language that uses a Java-based programming language translator to generate code for a C and MPI -based runtime. Swift/T can distributed and run user scientific codes written in any language, but Python and R modules are a popular choice. The system is an excellent environment in which to learn about programming language design/implementation, distributed computing, and systems software in general.
Desired relevant skills, background, or interests:
C language and Linux environment.
Other comments:
Do any special requirements apply? Permanent Resident OK; International OK
Other, specify:
Keywords:
workflows scripting computer languages programming distributed computing simulation exascale
Lightning Talk Title: Programming Workflows at Exascale