Welcome to the PAW-ATM Workshop.
Program
- 8:30 - 9:10 Anshu Dubey - Argonne National Laboratory
Keynote Speaker
An Application Perspective on Programming Models for the Future
(Abstract, Presentation)
- 9:10 - 9:20 Soren Rasmussen, Ethan D Gutmann, Brian Friesen, Damian Rouson,
Salvatore Filippone and Irene Moulitsas
Development and performance comparison of MPI and Fortran
Coarrays within an atmospheric research model
(Presentation)
- 9:20 - 9:35 Muhammed Abdullah Al Ahad, Christian Simmendinger,
Roman Iakymchuk, Erwin Laure and Stefano Markidis
Efficient Algorithms for Collective Operations with Notified
Communication in Shared Windows
(Presentation)
- 9:35 - 9:50 Jonas Posner, Lukas Reitz and Claudia Fohry
Comparison of the HPC and Big Data Java Libraries Spark, PCJ and APGAS
(Presentation)
- 9:50 - 10:00 James Bordner and Michael L. Norman
Computational Cosmology and Astrophysics on Adaptive Meshes
using Charm++
(Presentation)
- 10:00 - 10:30 Break
- 10:30 - 10:45 Paul H. Hargrove and Dan Bonachea
GASNet-EX Performance Improvements Due to Specialization for the
Cray Aries Network
(Presentation)
- 10:45 - 11:00 Louis Jenkins, Marcin Zalewski and Michael Ferguson
Chapel Aggregation Library (CAL)
(Presentation)
- 11:00 - 11:15 Maximilian H Bremer, John D Bachan and Cy P Chan
Semi-Static and Dynamic Load Balancing for Asynchronous Hurricane
Storm Surge Simulations
(Presentation)
- 11:15 - 11:30 Martin Biel and Mikael Johansson
Distributed L-shaped Algorithms in Julia
(Presentation)
- 11:30 - 12:00 Benjamin Robbins - Cray Inc.
Panel Chair
Panel Discussion
Panelists:
- Bradford L. Chamberlain - Cray Inc.
- Anshu Dubey - Argonne National Laboratory (ANL)
- Salvatore Filippone - Cranfield University, UK
- Kimberly Keeton - Hewlett Packard Enterprise (HPE)
Keynote Speaker
An Application Perspective on Programming Models for the Future
Salvatore Filippone and Irene Moulitsas
Development and performance comparison of MPI and Fortran
Coarrays within an atmospheric research model
Roman Iakymchuk, Erwin Laure and Stefano Markidis
Efficient Algorithms for Collective Operations with Notified
Communication in Shared Windows
Comparison of the HPC and Big Data Java Libraries Spark, PCJ and APGAS
Computational Cosmology and Astrophysics on Adaptive Meshes
using Charm++
GASNet-EX Performance Improvements Due to Specialization for the
Cray Aries Network
Chapel Aggregation Library (CAL)
Semi-Static and Dynamic Load Balancing for Asynchronous Hurricane
Storm Surge Simulations
Distributed L-shaped Algorithms in Julia
Panel Discussion
Panelists:
- Bradford L. Chamberlain - Cray Inc.
- Anshu Dubey - Argonne National Laboratory (ANL)
- Salvatore Filippone - Cranfield University, UK
- Kimberly Keeton - Hewlett Packard Enterprise (HPE)
Important dates
- Submission Deadline:
July 31, 2018 August 15, 2018
- Author Notification:
September 1, 2018 September 10, 2018
- Camera Ready: October 1, 2018
- Workshop Date: November 16, 2018 (8:30am - 12:00pm)
Summary
As high-performance computing hardware incorporates increasing levels of heterogeneity, hierarchical organization, and complexity, parallel programming techniques necessarily grow in complexity or in their ability to abstract away complexity. The concurrent development of multi- and many-core processors, deep memory hierarchies, and accelerators and the variety of ways to combine these makes the low-level language route unmanageable for domain experts tasked with developing applications. The technologies that a competent developer might be expected to master and combine include MPI plus CUDA, OpenMP, and OpenACC, most commonly denoted MPI + X. This approach inherently saddles the developer with low-level details that might better be handled by high-level abstractions.
Higher-level parallel programming models offer rich sets of abstractions that feel natural in the intended applications. Such approaches include programming languages (Fortran, UPC, Chapel, Julia), systems for large-scale data processing and analytics (Spark, Tensorflow, Dask), and frameworks and libraries that extend existing languages (Charm++, UPC++, Coarray C++, HPX, Legion, Global Arrays). While there are tremendous differences between these approaches, all strive to support better programmer abstractions for concerns such as data parallelism, task parallelism, dynamic load balancing, and data placement across the memory hierarchy.
This workshop will bring together applications experts who will present concrete practical examples of using such alternatives to MPI in order to illustrate the benefits of high-level approaches to scalable programming. The workshop expands upon the two similar workshops, PAW16 and PAW17, by broadening the theme beyond partitioned global address space languages. We invite you to take part in the Parallel Applications Workshop, Alternatives To MPI, and to join this vibrant and diverse community of researchers and developers.
Scope and Aims
The scope of the PAW-ATM workshop is to provide a forum for exhibiting case studies of higher-level programming models as MPI alternatives in the context of applications as a means of better understanding applications of MPI alternatives. We encourage the submission of papers and talks detailing such applications, including characterizations of scalability and performance, of expressiveness and programmability, as well as any downsides or areas for improvement in existing higher-level programming models. In addition to informing other application programmers about the potential that is available through MPI alternatives, the workshop is designed to communicate these experiences to compiler vendors, library developers, and system architects in order to achieve broader support for high-level approaches to scalable programming.
We also specifically encourage submissions covering big data analytics, deep learning, and other novel and emerging application areas, beyond well established HPC domains.
Topics
Topics include, but are not limited to:
- Novel application development using parallel programming languages.
- Examples that demonstrate performance, compiler optimization, error checking, and reduced software complexity.
- Applications from big data analytics, bioinformatics, and other novel areas.
- Performance evaluation of applications developed using MPI alternatives.
- Algorithmic models enabled by high-level parallel abstractions.
- Experience with the use of new compiler and runtime environments.
- Libraries using or supporting MPI alternatives.
- Benefits of hardware abstraction and data locality on algorithm implementation.
Submissions
Submissions are solicited in two categories:
Full-length papers will be published in the workshop proceedings in cooperation with IEEE TCHPC (pending proposal acceptance). Submitted papers must be original work that has not appeared in and is not under consideration for another conference or a journal. Papers shall not exceed ten (10) pages including text, appendices, and figures. Appendix pages related to the reproducibility initiative not included.
Extended abstracts will be evaluated separately and will not be included in the published proceedings; they are intended for timely communications of novel work that is going to be formally submitted elsewhere at a later stage, and/or of already published work that is nonetheless deemed appropriate for dissemination in this venue. Extended abstracts shall not exceed four (4) pages.
Preferential treatment will be given to full-length paper submissions. Accepted full-length papers will be given longer presentation slots at the workshop than the abstract-only option.
Submissions shall be submitted through Linklings. Submissions must use 10pt fonts in the IEEE format. The page limit (8 pages minimum for publication) includes figures, tables, and your appendices, but does not include references, for which there is no page limit. Reproducibility initiative dependencies (Artifact Description or Computational Results Analysis) are also not included in the page limit.PAW-ATM follows the reproducibility initiative of SC18, please refer to the FAQ provided for additional information. A template format for the artifact description appendix is provided.
Organization
Workshop Chair
Karla Morris - Sandia National Laboratories
Organizing Committee
- Bradford L. Chamberlain - Cray Inc.
- Salvatore Filippone - Cranfield University
- Costin Iancu - Lawrence Berkeley National Laboratory
Program Committee Chair
- Bill Long - Cray Inc.
Program Committee
- Bradford L. Chamberlain - Cray Inc.
- Valentin Churavy - Massachusetts Institute of Technology
- Salvatore Filippone - Cranfield University, UK
- Alex Gittens - Rensselaer Polytechnic Institute
- Costin Iancu - Lawrence Berkeley National Laboratory
- Hartmut Kaiser - Louisiana State University
- Laxmikant Kale - University of Illinois
- Seung-Hwan Lim - Oak Ridge National Laboratory
- Bill Long - Cray Inc.
- Karla Morris - Sandia National Laboratories
- Mitsuhisa Sato - RIKEN Advanced Institute for Computational Science
- Sean Treichler - NVIDIA
- Jeremiah J. Wilke - Sandia National Laboratories
Advisory Committee
- Katherine A. Yelick - Lawrence Berkeley National Laboratory
- Damian W. I. Rouson - Sourcery Institute
Contact
In case of questions please email us at: paw-atm18@cranfield.ac.uk