IPDRM 2016

First Annual Workshop on Emerging Parallel and Distributed Runtime Systems and Middleware

May 27, 2016, Chicago Illinois, USA.


Held in conjunction with the 30th IEEE International Parallel and Distributed Processing Symposium, (IEEE IPDPS 2016), May 23-27, 2016, Chicago, Illinois, USA.

Submission deadline for full papers is January 22, 2016 (Automatic Extension January 29, 2016)


Node architectures of extreme-scale systems are rapidly increasing in complexity. Emerging homogeneous and heterogeneous designs provide massive multi-level parallelism, but developing efficient runtime systems and middleware that allow applications to efficiently and productively exploit these architectures is extremely challenging. Moreover, current state-of-the-art approaches may become unworkable once energy consumption, resilience, and data movement constraints are added. The goal of this workshop is to attract the international research community to share new and bold ideas that will address the challenges of design, implementation, deployment, and evaluation of future runtime systems and middleware.

 Download the IPDRM 16 CFP


This workshop will emphasize novel, disruptive research ideas over incremental advances. We will solicit papers on topics including, but not limited to, the following areas:

Runtime System/Middleware Design, Evaluation and Usage

Constraints and Issues for Runtime Systems and Middleware

Design Principles and Programming Support


We invite two kinds of submissions to this workshop: (1) Full-length research papers (8-page limit); (2) Short papers (4-page limit), which can take the form of position papers, experience reports, or surveys/comparisons of runtime systems and middleware. Papers should not exceed eight (or four) single-spaced pages (including figures, tables and references) using 10-point font on 8.5x11-inch pages. Submissions will be judged on correctness, originality, technical strength, significance, presentation quality, and appropriateness. Submitted papers should not have appeared in or under consideration for another venue. A full peer-review process will be followed with each paper being reviewed by at least 3 members of the program committee

Submission will be accepted through the EasyChair System through this link: (https://easychair.org/conferences/?conf=ipdrm16)


Full Papers (8 Pages Max)

Short Papers (4 Pages Max)

All dates are Anywhere on Earth

General Chairs

Publicity Chair

Proceeding Chair

Web Chair

Program Committee

Workshop Program

09:00 AM to 09:15 AMWelcome and Workshop Introduction
09:15 AM to 10:00 AMKeynote: Hank Hoffmann, University of Chicago
10:00 AM to 10:30 AMBreak
10:30 AM to 12:00 PM

Session 1 -- Chair: Shuaiwen Leon Song, Pacific Northwest National Lab

Non-Intrusive Migration of MPI Processes in OS-bypass Networks Simon Pickartz (RWTH Aachen University), Carsten Clauss (ParTec Cluster Competence Center GmbH), Stefan Lankes (RWTH Aachen University), Stephan Krempel (ParTec Cluster Competence Center GmbH), Thomas Moschny (ParTec Cluster Competence Center GmbH), Antonello Monti (RWTH Aachen University)

Photon: Remote Memory Access Middleware for High-Performance Runtime Systems Ezra Kissel, Martin Swany (Indiana University)

Asynchronous Runtimes in Action: An Introspective Framework for a Next Gen Runtime Joshua Suetterlein (University of Delaware), Joshua Landwehr (University of Delaware), Andres Marquez (Pacific Northwest National Lab), Joseph Manzano (Pacific Northwest National Lab), Guang R Gao (University of Delaware)

12:00 PM to 01:15 PMLunch
01:15 PM to 02:45 PM

Session 2 -- Chair: Todd Gamblin, Lawrence Livermore National Lab

OWBP: Flash-Aware Offline Write Buffer Policy Alireza Haghdoost, David H.C. Du (University of Minnesota)

Topology-Aware Rank Reordering for MPI Collectives Seyed Hessamedin Mirsadeghi, Ahmad Afsahi (Queen’s Univeristy)

GPUShare: Fair-sharing Middleware for GPU Clouds Anshuman Goswami, Jeffrey Young, Karsten Schwan, Naila Farooqui, Ada Gavrilovska, Matthew Wolf, Greg Eisenhauer (Georgia Tech)

02:45 PM to 03:15 PMBreak
03:15 PM to 04:45 PM

Session 3 -- Chair: Joseph Manzano, Pacific Northwest National Lab

Performance Characterization of Hypervisor- and Container-based Virtualization for HPC on SR-IOV Enabled InfiniBand Clusters Jie Zhang, Xiaoyi Lu, Dhabaleswar K. (DK) Panda (The Ohio State University)

Macaca: A Scalable and Energy-Efficient Platform for Coupling Cloud Computing with Distributed Embedded Computing Heng Zhang, Chunliang Hao, Yanjun Wu, Mingshu Li (Institute of Software, Chinese Academy of Science)

Benchmarking Streaming Computation Engines: Storm, Flink and Spark Streaming Sanket Chintapalli, Derek Dagit, Bobby Evans, Reza Farivar, Thomas Graves, Mark Holderbaugh, Zhuo Liu, Kyle Nusbaum, Kishorkumar Patil, Boyang Jerry Peng, Paul Poulosky (Yahoo! Inc)

04:45 PM to 04:55 PMWorkshop Closing Comments



What is the Big Deal About Approximate Computing?


Hank Hoffmann, University of Chicago


Approximate computing has recently received a great deal of attention from a range of researchers including circuit designers, hardware architects, and programming language designers. This talk discusses some of the recent trends in approximate computing and then argues that really approximation is something that application developers have been doing all along. So, perhaps the biggest insight in the current trend in approximation is that by exposing the things applications developers approximate to the rest of the computer system, there is the opportunity to do even more. We then investigate one of those things that is possible when the computer system can coordinate with an approximate application.

Specifically, we discuss JouleGuard: a framework that coordinates approximate applications with system resource usage to meet user-defined energy goals with control theoretic formal guarantees. We show results of using JouleGuard on three different platforms (a mobile, tablet, and server) with eight different approximate applications created from two different frameworks. We find that JouleGuard respects energy budgets, provides near optimal accuracy, adapts to phases in application workload, and provides better outcomes than application approximation or system resource adaptation alone. JouleGuard is general with respect to the applications and systems it controls, making it a suitable runtime for a number of approximate computing frameworks.