Program

8:50 - 9:00

Opening Remarks   

9:00 - 10:00

Keynote Talk

Title: AI @ the Edge  

Speaker: Pete Beckman, Co-Director, Northeastern / Argonne Inst. for Science and Eng.; Senior Fellow, University of Chicago Computation Institute

Abstract: The number of network-connected devices (sensors, actuators, instruments, computers, and data stores) now substantially exceeds the number of humans on this planet. Billions of things that sense, think, and act are connected to a planet-spanning network of cloud and high-performance computing centers that contain more computers than the entire Internet did just a few years ago. Parallel computation and machine learning are providing the basis for this new computing continuum that analyses data in-situ, and uses HPC to model, predict, and learn. This new paradigm is giving rise to intelligent cities, smart agriculture, and advanced manufacturing. The Amazon Inc. DeepLens system is an example of this new model that links machine learning at the edge to the cloud. Another example is the Waggle Platform, developed at Argonne National Laboratory. The Array of Things project at the University of Chicago is deploying hundreds of Waggle-based nodes with advanced wireless sensors in Chicago and other cities. Each of the nodes support parallel edge computing. This presentation will explore the computing continuum, and how artificial intelligence at the edge is now firmly connected to supercomputing.

Bio: Pete Beckman is the co-director of the Northweastern University/Argonne Institute for Science and Engineering and is a recognized global expert in high-end computing systems. During the past 25 years, his research has been focused on software and architectures for large-scale parallel and distributed computing systems. For the DOE’s Exascale Computing Project, Beckman leads the Argonne team focused on extreme-scale operating systems and run-time software. He is the founder and leader of the Waggle project for smart sensors and edge computing that is used by the Array of Things project. Beckman also coordinates the collaborative technical research activities in extreme-scale computing between the US Department of Energy and Japan’s ministry of education, science, and technology and helps lead the BDEC (Big Data and Extreme Computing) series of international workshops. Beckman leads the extreme computing research activities at Argonne National Laboratory. He received his PhD in computer science from Indiana University.

10:00 - 10:30

Coffee Break

10:30 - 12:00

Regular Paper Session I: High-Performance Data Analytics and Management

Session Chair: Dr. Shantenu Jha

Load Imbalance Mitigation Optimizations for GPU-Accelerated Similarity Joins    (Best Paper Award Winner!) Slides

Benoit Gallet and Michael Gowanlock

A Performance Analysis of Large Scale Scientific Computing Applications from Log Archive    Slides

Liqiang Cao, Xu Liu, Xiaowen Xu, and Zhanjun Liu

Mnemo: Boosting Memory Cost Efficiency in Hybrid Memory Systems    Slides

Thaleia Dimitra Doudali and Ada Gavrilovska

12:00 - 13:30

Lunch Break

13:30 - 15:00

Regular Paper Session II: High-Performance Machine Learning

Session Chair: Dr. Ada Gavrilovska

Learning Everywhere: Pervasive Machine Learning for Effective High-Performance Computation    Slides

Shantenu Jha and Geoffrey C. Fox

Green, Yellow, Yield: End-Host Traffic Scheduling for Distributed Deep Learning with TensorLights   Slides

Xin Huang, Ang Chen, and Eugene Ng

A GPU Inference System Scheduling Algorithm with Asynchronous Data Transfer   Slides

Qin Zhang, Li Zha, and Boqun Cheng

15:00 – 15:30

Coffee Break

15:30 - 16:10

Short Paper Session: High-Performance Data Processing Algorithms (20 mins each)

Session Chair: Dr. Liqiang Cao

It Can Understand the Logs, Literally   Slides

Aidi Pi, Wei Chen, Will Zeller, and Xiaobo Zhou

A Partition-centric Distributed Algorithm for Identifying Euler Circuits in Large Graphs   Slides

Siddharth D Jaiswal and Yogesh Simmhan

16:10 – 16:15

Panel Preparation

16:15 – 17:45

Panel: HPC and Cloud Convergence. What about HPC and Edge?   

Panel Moderator: Dr. Ada Gavrilovska, Georgia Tech

Panel Members:

Much has been said in recent years about the convergence of HPC and Cloud technologies. But many cloud applications are pushing the boundaries of current datacenter-based solutions and driving the development of a new tier of technologies for the “Edge”. The goal of this panel is to discuss the use cases from the scientific and high-performance community where the edge plays an important role, the technologies that will shape the landscape in this domain, and how and whether the HPC and enterprise communities can leverage each others efforts to accelerate progress.

17:45 - 18:00

Closing Remarks