3rd workshop on

hierarchical parallelism
for exascale computing

Friday,  18th November 2022

In cooperation with:

Held in conjunction with:

The International Conference for High Performance Computing, Networking, Storage and Analysis


The International Conference for High Performance Computing,
Networking, Storage and Analysis

HiPar22 welcomes HPC practitioners, from hardware and compiler experts to algorithms and software developers, to present and discuss new studies, approaches and cutting-edge ideas to utilize multi-level parallelism for extreme scale computing.

SUMMARY

Industrial competition, the explosion of AI and machine learning, and the computational needs of leading-edge scientific research are driving rapid changes and advancements in High Performance Computing (HPC). Computational nodes are becoming increasingly more powerful, featuring a large number of physical cores in multiple sockets and accelerators.
This increases the overall complexity, due to complex architectures (e.g., memory hierarchies and tens of compute elements), novel accelerator designs, and energy constraints. Hierarchical parallelism is an approach that is gaining momentum in HPC: it embraces the intrinsic complexity of current and future HPC systems, rather than avoiding it, by exploiting parallelism at all levels: compute, memory and network.

Hierarchical parallelism is the focus of this workshop. We aim at bringing together hardware, application and software practitioners proposing new strategies to fully exploit computational hierarchies, and examples to illustrate their benefits for extreme scale computing.

WORKSHOP PROGRAM

(Workshop date: Friday Nov. 18th --- 8.30am-12.00pm CST)

8.30AM – 8.40am (CST)

Welcome and Overview

8.40AM – 9.40am / INvited speaker – SLIDES

Paving the way to exascale for scientific applications

Speaker: Damien Lebrun-Grandie (Scientist, Oak Ridge National Labs, USA) 

9.40AM – 10.00am / PAPER 1 – SLIDES

OpenMP’s Asynchronous Offloading for All-pairs Shortest Path Graph Algorithms on GPUs

Authors:  Mathialakan Thavappiragasm (Oak Ridge National Lab., USA), Vivek Kale  (Brookhaven National Lab., USA)

10.00AM – 10.30am

break

10.30AM – 10.50AM / PAPER 2

Hierarchical Communication optimization for FFT

Authors: Mohit Kumar, Preeti Malakar (Indian Institute of Technology Kanpur)

10.50AM – 11.10AM / PAPER 3 – SLIDES

A High-Performance Design for Hierarchical Parallelism in the QMCPACK Monte Carlo code

Authors: Ye Luo (Argonne National Lab., USA), Peter Doak, Paul Kent (Oak Ridge National Lab., USA)

 

11.10AM – 11.30AM / PAPER 4 – SLIDES

Expressing Hierarchical Code Optimizations via MDH-Based Schedules

Authors: Ari Rasch, Richard Schulze, Sergei Gorlatch (University of Muenster, Germany)

11.30AM – 11.50AM / PAPER 5 – SLIDES

CAMP: a Synthetic Micro-Benchmark for Assessing Deep Memory Hierarchies

Authors: Wenqing Peng, Evgenij Belikov (EPCC, UK)

11.50AM – 12.00PM (CST)

CLOSING REMARKS

WORKSHOP DETAILS

HiPar22 is designed to showcase new studies, approaches, and cutting-edge ideas on hierarchical parallelism for extreme-scale computing. Our goal is to highlight not just success stories but also discuss drawbacks and challenges.

We welcome contributions from the HPC community addressing the use of emerging architectures — focusing particularly on those characterized by fewer but more powerful nodes as well as systems with hierarchical networks, where the hierarchy is not just characterized by performance metrics, but tiered communication semantics. Specifically, the emphasis is on the design, implementation, and application of programming models for multi-level parallelism, including abstractions for hierarchical memory access, heterogeneity, multi-threading, vectorization, and energy efficiency, as well as scalability and performance studies thereof.

Of particular interest are models addressing these concerns portably: providing ease of programming and maintaining performance in the presence of varied accelerators, hardware configurations, and execution models. Studies that explore the merits of specific approaches to addressing these concerns, such as generic programming or domain specific languages, are also in scope.

The workshop is not limited to the traditional HPC software community. As one example, another key topic is the use of hierarchical parallelism in dealing with the challenges arising in machine learning, due to the growing importance of this field, the large scale of systems tackled in that area, and the increasing interest at SC.

Submissions are encouraged in, but not limited to the following areas:

  • Programming heterogeneous nodes;
  • Hierarchical work scheduling and execution;
  • Leading edge programming models, for example fully distributed task-based models and hybrid MPI+X, with X representing shared memory parallelism via threads, vectorization, tasking or parallel loop constructs;
  • Hardware, software and algorithmic advances for efficient use of memory hierarchies, multi-threading and vectorization;
  • Novel approaches leveraging asynchronous execution to maximize efficiency;
  • Efficient use of nested parallelism, for example CUDA dynamic parallelism, for large scale simulations;
  • Implementations of algorithms that are natural fits for nested work (for example approaches that use recursion);
  • Challenges and successes in managing computing hierarchies;
  • Examples demonstrating effective use of the combination of inter-node and intra-node parallelism;
  • Challenges and successes of porting of existing applications to many-core and heterogeneous platforms;
  • Recent developments in compiler optimizations for emerging architectures;
  • Applications from emerging AI fields, for example deep learning and extreme-scale data analytics.

Papers submissions are solicited in the following categories:

(a) Research papers:

Intended to describe original work and ideas that have not appeared in another conference or journal, and are not currently under review for any other conference or journal.
Regular papers must be at least (4) and not exceed (10) letter size pages (U.S. letter – 8.5″x11″).
Accepted regular papers will be published in the workshop proceedings in cooperation with IEEE Computer Society.

(b) Highlight talks/ideas:

Intended for material that is not mature enough for a paper, to present novel, interesting ideas
or preliminary results that will be formally submitted elsewhere.
Submissions must not exceed (2) letter size pages (U.S. letter – 8.5″x11″).
These will NOT be included in the proceedings but will be part of the HiPar program. 

Please note:

Papers must follow the IEEE format: https://www.ieee.org/conferences/publishing/templates.html.
The page limits above only apply to the core text, content-related appendices, and figures.
References and reproducibility appendix do not count against the page limit.

When deciding between submissions with comparable evaluations, priority will be given to those with higher quality of presentation and whose focus relates more directly to the workshop themes.

Papers must be submitted at https://submissions.supercomputing.org

HiPar22 follows the SC22 reproducibility and transparency initiative:
https://sc22.supercomputing.org/submit/reproducibility-initiative

HiPar22 requires all submission to include an Artifact Description (AD) Appendix.
The Artifact Evaluation (AE) remains optional.

We also encourage authors to follow the transparency initiative for two reasons:
(a) it helps the authors themselves with the actual writing and structuring of the paper to express the research process;
(b) it helps readers understand the thinking process used by the authors to plan, obtain and explain their results.

IMPORTANT DATES

  • Submission Deadline (extended):
    September 2nd, 2022 (AoE)
  • Author Notification:
    September 12, 2022
  • Camera Ready:
    October 7, 2022
  • Final Program:
    October 7, 2022
  • Workshop Date:
    Friday, November 18th 2022

ORGANIZATION

WORKSHOP CHAIR
Francesco Rizzi
NexGen Analytics
Organizing Committee
Lee Howes

Meta

Ulrike Meier Yang

Lawrence Livermore National Lab

Program Committee CHAIRS
Christian Trott

Sandia National Labs

Filippo Spiga

NVIDIA

Program Committee
Gabriele Fatigati

Cineca

Tom Deakin

Univ. of Bristol

Marta Garcia Gasulla

BSC

Anja Gerbes

ZIH Dresden

Mark Hoemmen

NVIDIA

Zhiqi Tao

Intel

Matthew Bettencourt

NVIDIA

Daniel Arndt

Oak Ridge National Labs

Nur Aiman Fadel

CSCS

Phil Miller

NexGen Analytics

James Lin

Shanghai J.Tong Univ.

Nicholas Malaya

AMD

Aram Markosyan

Meta

Flavio Vella

Univ. of Trento

Cezary Skrzyński

NexGen Analytics