Invited Talk 1

Accelerating Irregular Algorithms on GPUs

Martin Burtscher
(Texas State University)

GPU architectures are highly tuned for regular codes, making them a seemingly poor fit for irregular applications with non-uniform control flow and memory access patterns. This perception is often reenforced by disappointing performance when porting irregular CPU programs to GPUs. In this talk, we highlight important algorithmic, code-optimization, data-structure, and implementation-style differences that make irregular algorithms effective on GPUs. We show on several examples that, by moving away from CPU-centric thinking and coding directly for the GPU, the latter typically turns out to be the more suitable hardware for running irregular applications.
Mobirise

Martin Burtscher is a Professor in the Department of Computer Science at Texas State University. He received the BS/MS degree from ETH Zurich and the PhD degree from the University of Colorado at Boulder. Martin's current research focuses on the parallelization of irregular graph algorithms and other complex programs for GPUs as well as on the synthesis of high-speed lossy and lossless data-compression algorithms. He has co-authored about 130 peer-reviewed scientific publications, which have been cited around 7000 times. Martin is a distinguished member of the ACM and a senior member of the IEEE.

Invited Talk 2

Discussion on Hash-Table Approaches for Efficient Sparse Tensor Contraction

Jiajia Li
(North Carolina State University)

Sparse tensor contraction (SpTC) is a crucial operation in high-performance applications, particularly in computational chemistry, high-order tensor decompositions, and quantum sciences. This talk will explore the performance challenges associated with SpTC and review current state-of-the-art solutions. We will focus on hash-table approaches, discussing the key features of hash table design that significantly impact performance. Additionally, a novel hash method will be introduced, featuring a fast hash function with guaranteed collision-free operations to efficiently support SpTC computations. 
Mobirise

Jiajia Li is an Assistant Professor in the Department of Computer Science at North Carolina State University (NCSU). Her research focuses on high-performance computing, emphasizing the interaction among applications, numerical methods, data structures, algorithms, automatic performance tuning, and computer architectures. She is particularly interested in high-performance sparse (multi-)linear algebra, solvers, and tensor decompositions for large-scale data analytics and domain applications on diverse computer architectures.  Before joining NCSU, Jiajia Li was an Assistant Professor in the Department of Computer Science at the College of William & Mary (W&M) and a Research Scientist at the High-Performance Computing group of Pacific Northwest National Laboratory (PNNL). She earned her Ph.D. in Computational Science & Engineering at Georgia Institute of Technology, advised by Professor Richard Vuduc. She has received several awards, including the Rising Stars in Computational and Data Sciences, Best Paper Award, Best Student Paper Award, and IBM PhD Fellowship. Previously, she also earned a Ph.D. degree from the Institute of Computing Technology at Chinese Academy of Sciences and received her B.S. in Computational Mathematics from Dalian University of Technology.

‌

AI Website Generator