Fast GPU Code for Graphs

Speaker: Wen-mei W. Hwu , University of Illinois, Urbana-Champaign

Date: Monday, June 15, 2020

Time: 2:00 PM to 3:00 PM Note: all times are in the Eastern Time Zone

Public: Yes

Location: https://mit.zoom.us/j/536883569 (Please email lindalynch@csail.mit.edu for password)

Event Type: Seminar

Room Description: https://mit.zoom.us/j/536883569 (Please email lindalynch@csail.mit.edu for password)

Host: Julian Shun, MIT CSAIL

Contact: Julian Shun, jshun@mit.edu

Relevant URL: http://fast-code.csail.mit.edu/

Speaker URL: https://ece.illinois.edu/about/directory/faculty/w-hwu

Speaker Photo:
None

Reminders to: fast-code-seminar@lists.csail.mit.edu, seminars@csail.mit.edu, pl@csail.mit.edu

Reminder Subject: TALK: Fast GPU Code for Graphs

Abstract: Modern analytics and recommendation systems are increasingly based on graph data that capture the relations between entities being analyzed. At the IBM-Illinois Center for Cognitive Computing (C3SR), we have been developing practical natural language tools such as name-entity resolution and domain-concept identification based on graphs. In practice, graphs come in huge sizes, offer massive parallelism, exhibit a wide variety of sparsity patterns, and are stored in sparse-matrix formats such as compressed sparse row (CSR). The huge sizes and sparsity of practical graphs have led to long execution times of analytics applications. To exploit the massive parallelism and speed up graph analytics, applications developers are increasingly interested in using GPUs for graph analytics and traversal. However, developers of GPU graph code face major challenges such as load imbalance, insufficient GPU memory, and coalesced memory accesses. Recent generations of GPUs have also introduced architectural features such as dynamic parallelism and unified virtual memories that offer new venues for addressing the challenges in creating fast GPU graph code. In this talk, I will present fast GPU algorithms developed at C3SR that have been shown to achieve much higher execution speed than NVIDIA’s graph libraries and explain how the challenges are addressed in these algorithms.
 
Wen-mei W. Hwu is a Professor and holds the Sanders-AMD Endowed Chair in the Department of Electrical and Computer Engineering, University of Illinois at Urbana-Champaign. He is the director of the IMPACT research group (www.crhc.uiuc.edu/Impact). He co-directs the IBM-Illinois Center for Cognitive Computing Systems Research (C3SR)
and serves as one of the principal investigators of the NSF Blue Waters Petascale supercomputer. For his contributions, he received the ACM SigArch Maurice Wilkes Award, the ACM Grace Murray Hopper Award, the IEEE Computer Society Charles Babbage Award, the ISCA Influential Paper Award, the IEEE Computer Society B. R. Rau Award and the Distinguished Alumni Award in Computer Science of the University of California, Berkeley. He is a fellow of IEEE and ACM. Dr. Hwu received his Ph.D. degree in Computer Science from the University of California, Berkeley.

Research Areas:
Algorithms & Theory, Computer Architecture, Programming Languages & Software, Systems & Networking

Impact Areas:
Big Data

This event is not part of a series.

Created by Julian J. Shun at Monday, June 08, 2020 at 2:19 PM.