PMAM 2023

Workshop Proceedings

You can find the workshop proceedings from here.

Workshop Program

A pdf version of the workshop program is available here.

Special Issue of CC-PE

The information of the special issue of CC-PE can be found here.

Keynote by Tao B. Schardl

You can find the keynote slides from here.

Title

OpenCilk: Architecting a Task-Parallel Software Infrastructure for Modularity, Extensibility, and Performance

Abstract

OpenCilk is a new open-source software infrastructure to support Cilk task-parallel programming, especially for researchers and teachers. OpenCilk provides a full-featured implementation of the Cilk and an integrated infrastructure that includes three main components: a compiler, based on the award-winning Tapir/LLVM compiler; a streamlined and efficient work-stealing runtime scheduler; and a framework for productivity-tool development, based on compiler instrumentation for fork-join parallel computations. OpenCilk’s architecture allows for substantial code reuse and easy exploration of design choices in task-parallel language abstraction, compilation strategy, runtime mechanism, and productivity-tool development.

OpenCilk is modular — modifying one component for the most part does not necessitate modifications to other components — and easy to extend — its construction naturally encourages code reuse. For example, we extended OpenCilk to support multiple additional parallel runtime systems, including Cilk Plus, OpenMP, and oneTBB. Supporting each new parallel runtime required fewer than 2000 new lines of code. At the same time, OpenCilk produces high-performing executables that achieve high work-efficiency and parallel scalability.

Biography

Tao B. Schardl is a Research Scientist in the MIT Computer Science and Artificial Intelligence Laboratory (CSAIL) and Chief Architect of OpenCilk. His research combines algorithms and systems to develop technologies that support principled, scientific approaches to writing fast code. He has previously worked on parallel programming models, theories of performance, compilers, runtime systems, diagnostic tools, parallel algorithms, and the future of computer performance. He received the US Department of the Air Force Artificial Intelligence Accelerator Scientific Excellence Award in 2022 for his work on OpenCilk. His work on the Tapir/LLVM compiler earned the best paper award at the ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming in 2017. His work on computer performance in the post-Moore's Law era was published in Science and has been spotlighted in two Turing-award lectures. Dr. Schardl earned his S.B., M.Eng, and Ph.D. in Computer Science and Engineering from MIT.

Introduction

Rapid advancements in multicore and chip-level multi-threading technologies open new challenges and make multicore systems a part of the computing landscape. From high-end servers to mobile phones, multicores and manycores are steadily entering every single aspect of the information technology.

However, most programmers are trained in sequential programming, yet most existing parallel programming models are prone to errors such as data race and deadlock. To fully utilize multicore and manycore hardware, parallel programming models that allow easy transition of sequential programs to parallel programs with good performance and enable development of error-free codes are urgently needed.

This workshop is dedicated primarily to gather researchers and practitioners addressing the main challenges and share experiences in the emerging multicore and manycore software engineering and distributed programming paradigm. This workshop aims to provide a discussion forum for people interested in programming environments, models, tools and applications specifically designed for parallel multicore and manycore hardware environments.

Important Dates

Paper submission deadline :~~November 30, 2022~~ final extension to December 21th, 2022 (Anywhere On Earth)
Notification of acceptance : January 21, 2023
Camera-ready papers due : February 1, 2023
Workshop day: February 26, 2023

Objectives, scope and topics of the workshop

The program committee cordially invites any novel research ideas in (but not limited to) the following topics:

programming models and systems for multicore, manycore, and clusters of multicore/manycore
parallel machine learning and AI algorithms
parallel resource management on large-scale datacenters
multicore and manycore software engineering
parallel and distributed algorithms on GPU, FPGA, and multicore clusters
parallel libraries and frameworks
performance analysis, efficiency and effectiveness
massively parallel processing on multicore/manycore systems and clusters
automated parallelization and compilation techniques
debugging and performance autotuning tools and techniques for multicore/manycore applications
parallel algorithms, applications and benchmarks on multicore/manycore systems
runtime power/energy management on multicore/manycore systems and clusters
fault tolerance and resilience

Names and contacts of key organizers

Steering Committee

Pavan Balaji
Facebook, USA
pavanbalaji.work at gmail dot com

Minyi Guo
Shanghai Jiaotong University, China
myguo at sjtu dot edu dot cn

Zhiyi Huang
University of Otago, New Zealand
hzy at cs dot otago dot ac dot nz

Organization co-chairs

Quan Chen
Shanghai Jiaotong University, China
chen-quan at sjtu dot edu dot cn

Zhiyi Huang
University of Otago, New Zealand
hzy at cs dot otago dot ac dot nz

Min Si
Facebook, USA
minsi.atwork at gmail dot com

Programme Committee

Pedro Alonso Jordá, Universidad Politécnica de Valenci, Spain
Richard Barrett, Sandia National Laboratories, USA
Chen Ding, University of Rochester, USA
Zidong Du, ICT, Chinese Academy of Sciences, China
Andrew Ensor, Auckland University of Technology, New Zealand
Khaled Hamidouche, AMD Research, USA
Toshihiro Hanawa, University of Tokyo, Japan
Palden Lama, University of Texas at San Antonio, USA
Laurent Lefevre, INRIA, University of Lyon, France
Dong Li, Oak Ridge National Laboratory, USA
Shigang Li, ETH, Zurich, Switzerland
Chen Liu, Clarkson University, USA
Xiaoyi Lu, University of California Merced, USA
Kavitha Madhu, Argonne National Laboratory, USA
Oliver Sinnen, University of Auckland, New Zealand
Peter Strazdins, Australian National University, Australia
Liqiang Wang, University of Central Florida, USA
Hailong Yang, Beihang University, China
Esma Yildirim, Queensborough Community College of CUNY, USA
Amelie Chi Zhou, Shenzhen University, China

Manuscript submission

Papers reporting original and unpublished research results and experience are solicited. All paper submissions will be handled electronically via Hotcrp: https://pmam23.hotcrp.com.

Papers must not exceed 10 pages in standard SIGPLAN proceedings format (\documentclass[sigplan,screen]{acmart}). The Latex template is available at here.

Authors must submit their paper through Hotcrp: https://pmam23.hotcrp.com. If you have problems accessing the system, e-mail your submission to: pmam2023 at lists dot otago dot ac dot nz

Proceedings

All accepted papers will be published in the PMAM 2023 proceedings by the ACM Digital Library, and will be included in the Elsevier databases Scopus and Compendex (EI indexed).

Selected best papers of PMAM will be considered for publication in a special issue of CC-PE.

Registration

Information about registration at PPoPP 2023 main website.

Contact Us

For further information regarding the workshop and paper submission, please send your request or enquiry to:
pmam2023 at lists dot otago dot ac dot nz