Data analytics are an integral part of data science. The introduction of new scalable and high-performance data analytics frameworks, including the use of accelerators such as GPUs, has enabled data scientists to inspect their data and extract insights previously unattainable. In the last decade, the GPU has trailblazed this new exciting field of high-performance data analytics. This session will cover recent developments that have helped scale data analytics on the GPU to new heights. We will focus on analytics in social network analysis, bioinformatics, and cybersecurity, to name a few. We will also focus on other foundational components of scalable data analytics such as communication, data structures, algorithms, and ETL.
NVIDIA and Georgia Institute of Technology, U.S.
David A. Bader
New Jersey Institute of Technology, U.S.
February 24, 2022
- 3:20-3:40 Graph Analytics in the Exascale Era
Mahantesh Halappanavar, Marco Minutoli, and Sayan Ghosh, Pacific Northwest National Laboratory, U.S.; Aydin Buluc, Lawrence Berkeley National Laboratory, U.S.; Erik G. Boman, Sandia National Laboratories, U.S.; Alex Pothen, Purdue University, U.S.
- 3:45-4:05 Inverse-Deletion BFS - Revisiting Static Graph BFS Traversals with Dynamic Graph Operations
Oded Green, NVIDIA and Georgia Institute of Technology, U.S.
- 4:10-4:30 Parallel Graph Algorithms by Blocks for Heterogeneous Systems
Abdurrahman Yasar, Georgia Institute of Technology, U.S.; Siva Rajamanickam and Jonathan Berry, Sandia National Laboratories, U.S.; Umit V. Catalyurek, Georgia Institute of Technology, U.S.
- 4:35-4:55 GPU Graph Processing using Modern C++
Andrew Lumsdaine, University of Washington, U.S.
February 25, 2022
- 11:10-11:30 Kernel Specialization for Faster Segmented Sort on GPUs
Georgii Evtushenko, Nvidia Corporation, U.S.
- 11:35-11:55 Improving the Speed and Quality of Parallel Graph Coloring
Ghadeer Alabandi and Martin Burtscher, Texas State University, U.S.
- 12:00-12:20 Delegate Centric Top-K Computation on GPUs
Hang Liu, Stevens Institute of Technology, U.S.
- 12:25-12:45 GPU Accelerated Online Search for Gravitational Waves
Zhihui Du, New Jersey Institute of Technology, U.S.