- This event has passed.
LANS Seminar
Seminar Title: Enabling Graph Analysis at Extreme Scale with Sparse Linear Algebra
Speaker: Md Taufique Hussain, Postdoctoral Research Fellow, Wake Forest University
Date: Thursday, December 4, 2025
Time: 2:30 PM-3:30 PM (In-Person)
Location: Hybrid, Bldg. 240, Conference Room 4301
Host: Srini Eswar
Description: Many scientific challenges within the DOE’s research portfolio generate massive data sets. For example, the Joint Genome Insitute’s metagenomic data contains over 80 billion microbial proteins, with the goal of identifying novel proteins within this extensive database. Many such large-scale problems, like this problem, can be modeled as graph analysis problems. Due to the massive scale of the problems, they require exascale computing power. However, irregular computational and data sparsity patterns present obstacles to the scalability and energy efficiency of algorithms that solve these problems. My research addresses this long standing challenge by developing algorithms that minimize inter-process communication, thereby enhancing both the scalability and energy efficiency of large-scale applications. In this presentation, I will demonstrate communication reduction techniques for sparse matrix algorithms that tackle the scalability challenges faced by real-world graph analysis applications, including traditional graph algorithms and graph machine learning. I will highlight these approaches through several recent projects focused on communication-avoiding sparse matrix multiplication (SpGEMM) and the computational demands associated with their scalability.
Bio: Md Taufique Hussain received his PhD from Indiana University earlier this year and is currently a postdoctoral research fellow at Wake Forest University. His research focuses on designing scalable and parallel linear algebra algorithms essential for graph analysis, scientific computing, and machine learning applications. His work has been published in leading high-performance computing journals and conferences, including IPDPS, SC, TPDS, and others. His work has been sponsored by the DOE ExaBiome, ExaGraph, and DOD NICE projects. Currently, he is a member of Sparsitute, an MMICC center funded by the DOE Advanced Scientific Computing Research program.
Please note that the meeting URL for this event can be seen on the cels-seminars website which requires an Argonne login.
See all upcoming talks at https://www.anl.gov/mcs/lans-seminars