Seminar: Erik Saule: 'Towards Achieving Performance and Parallelism for Graph Analysis'

Wednesday, February 15, 2023 - 5:00pm to 6:00pm
Event Description: 

Graph Analysis is taking a larger role in our scientific and social
models, there is a need to enable application scientists to obtain
quick solutions to their graph problems. At the same time,
computational machines are getting ever more complicated and their new
complexity requires additional consideration when designing codes to
leverage them. In this talk, I walk you through my journey towards
achieving performance and parallelism for graph problems.

We will talk about algorithmic work that is widely applicable to
parallel processing and can be used to inform on the performance of
graph analyses. We will delve in applications both within and outside
the realm of graph analysis. We will take a deep dive in theAdvisor, a
recommender system for academics which leverages graph analysis. We
will see how modern computing systems can be used to process problems,
including centrality metrics. Finally, we will delve into the most
important strategy to achieve wide adoption of high performance and
parallel computing: helping students learn the material in the most
effective way.

Short Bio:
Erik Saule (BS'03 University of Versailles, France, MS'05, PhD'08
Grenoble Institute of Technology, France) is an Associate Professor in
the Computer Science department of UNC Charlotte. He has been with UNC
Charlotte since August 2013. Dr. Saule has been a post-doctoral
researcher in the Department of Biomedical Informatics at The Ohio
State University from 2009 to 2013. His research interests revolves
around the efficient use of modern computing platform for compute
intensive or data intensive application. Naturally, Dr. Saule works
span from theoretical subject such as multi objective optimization,
approximation algorithm, scheduling, graph algorithms to practical
ones such as parallel computing, middleware and the use of
accelerators. He authored and coauthored over 60 refereed publications
and has been awarded over $1M of federal funding, including the
prestigious NSF CAREER award. Dr. Saule currently serves as PhD program
coordinator and associate chair for the CS department at UNC

Dr. Saule is a frequent reviewer for various journals including IEEE
Transactions on Parallel and Distributed Systems (IEEE TPDS), Parallel
Computing (ParCo) and Journal of Parallel and Distributed Computing
(JPDC) and a regular program committee member of various conferences
including the International Conference on Parallel Processing (ICPP),
of the International Conference for High Performance Computing,
Networking, Storage, and Analysis (SC) and of the IEEE International
Parallel and Distributed Processing Symposium (IPDPS). He served in
the past as Program Chair and General Chair of HCW, was the chair of
the Experiencing HPC for Undergraduate program at SC17, and served as
workshop chair of IPDPS for multiple years. He currently is the
algorithm for computational science co track chair for IPDPS 23.

Computed Event Type: 
Event Title: 
Seminar: Erik Saule: 'Towards Achieving Performance and Parallelism for Graph Analysis'