Charm++ Workshop 2021 (Oct 18-19)

[matthiasdiener]

Call for Participation at the

19th Annual Workshop on Charm++ and its Applications

October 18-19, 2021
University of Illinois at Urbana-Champaign (Virtual)
https://charm.cs.illinois.edu/charmWorkshop/
Hosted by the Parallel Programming Laboratory

There will be several talks from Charm++ collaborators, external speakers, and members of the Parallel Programming Laboratory. Talks will cover various topics including load balancing, performance visualization, interoperability, fault tolerance, recent runtime research, and progress in Charm++ applications.

The keynotes for this year's workshop include:

Challenges of Programming models for The Supercomputer "Fugaku" and Beyond
Mitsuhisa Sato, RIKEN R-CCS
Scheduled for 9:10-10:10 AM CDT on Oct 18 (Mon)

A Tale of Two Cultures: Challenges in Developing Tools for Writing Parallel Software
Alex Aiken, Stanford University
Scheduled for 11:00-12:00 PM CDT on Oct 19 (Tue)

This year's workshop will be fully online via Zoom. Please register via the following link to receive an invite to the Zoom session.

Registration: https://charm.cs.illinois.edu/charmWorkshop/
Registration is free this year, but please register to help us coordinate.

We invite you to join us. Please forward this invitation to any colleagues who may be interested.

Charm++ is a C++ based parallel programming system based on an introspective adaptive runtime system, with many features suitable for addressing upcoming extreme scale as well as mid-scale challenges, and with multiple highly scalable parallel applications such as NAMD. Charm++ is developed by the Parallel Programming Laboratory (PPL) at the University of Illinois at Urbana-Champaign. The goal of PPL is to develop technology that improves the performance of parallel applications while also improving programmer productivity. The workshop is broadly focused on adaptivity in highly scalable parallel computing. It also takes stock of recent results in adaptive runtime techniques in Charm++ and the collaborative interdisciplinary research projects developed using it.

[matthiasdiener]

Keynotes

1. Challenges of Programming models for The Supercomputer "Fugaku" and Beyond

Mitsuhisa Sato, RIKEN Center for Computational Science (R-CCS)

Scheduled for 9:10-10:10 AM CDT on Oct 18 (Mon)

Abstract: The supercomputer "Fugaku" is an exascale system, which is operated since March 2021 in R-CCS, Japan. Fugaku is an ultra-scale "general-purpose" manycore-based system with 158,976 nodes, 7.6M cores in total. We developed a new manycore processor, A64FX, for Fugaku, which supports Arm instruction sets with Scalable Vector Extension (SVE) SIMD instruction sets and is equipped with HBM2 memory. While OpenMP-MPI hybrid programing model is a standard programming model, new programming models such as task-based parallel programming models and parallel object-oriented programming models are expected to be useful to exploit several level of parallelism in such large-scale manycore-based system and wide SIMD parallelism in each core. In this talk, the overview and performance of Fugaku will be presented, followed by challenges of programming models for Fugaku and beyond.

---

2. A Tale of Two Cultures: Challenges in Developing Tools for Writing Parallel Software

Alex Aiken, Stanford University

Scheduled for 11:00-12:00 PM CDT on Oct 19 (Tue)

Abstract: The scientific computing community has been developing parallel programming tools for decades, with an emphasis on performance. The software industry has only more recently focused on the development of tools for writing parallel software, with an emphasis on productivity. This talk will compare and contrast the two approaches, with an emphasis on identifying research challenges for the scientific community to achieve productivity comparable to widely-used industry tools.

[matthiasdiener]

Workshop Program: October 18 (Mon) (all times in Central Time)

Session: Keynote I (Chair: Laxmikant V. Kale)
09:00 - 09:10	Opening Remarks		Laxmikant V. Kale
09:10 - 10:10	Keynote	Challenges of Programming models for The Supercomputer "Fugaku" and Beyond	Mitsuhisa Sato
10:10 - 10:30	Break
Session: Astrophysical Simulations (Chair: TBA)
10:30 - 11:00	Talk	Recent developments in HPX and Octo-Tiger	Patrick Diehl
11:00 - 11:30	Talk	Enzo-E/Cello astrophysics and cosmology: algorithmic improvements for robustness and scalability	James Bordner, Michael L. Norman
11:30 - 12:00	Talk	Simulating neutron stars with discontinuous Galerkin methods and Charm++	Nils Deppe
12:00 - 13:00	Lunch Break
Session: Load Balancing (Chair: Pathikrit Ghosh)
13:00 - 13:30	Talk	Dynamically Load-balanced p-adaptive Discontinuous Galerkin Methods using Charm++	Weizhao Li, Aditya Pandare, Hong Luo and Jozsef Bakosi
13:30 - 14:00	Talk	Optimizing Distributed Load Balancing for Time-Varying Imbalance	Jonathan Lifflander, Nicole Slattengren, Philippe Pébaÿ, Phil Miller, Francesco Rizzi and Matthew Bettencourt
14:00 - 14:30	Talk	Vector Load Balancing in Charm++	Ronak Buch
14:30 - 14:50	Break
Session: Programming Languages (Chair: Sam White)
14:50 - 15:20	Talk	What's New with Chapel: Applications, Aggregators, and Accelerators	Brad Chamberlain
15:20 - 15:50	Talk	Advances in Charm-based Languages II	Justin Szaday
15:50 - 16:00	Break
Session: Plasma Physics & Astrophysics (Chair: Jaemin Choi)
16:00 - 16:30	Talk	Adaptive Plasma Physics Simulations: Dealing with Load Imbalance using Charm++	Esteban Meneses
16:30 - 17:00	Talk	A Universe of ChaNGa Applications	Thomas Quinn
17:00 - 17:30	Talk	ParaTreeT: A Fast, General Framework for Spatial Tree Traversal	Joseph Hutter

[matthiasdiener]

Workshop Program: October 19 (Tue) (all times in Central Time)

Session: Molecular Dynamics & Quantum Chemistry (Chair: Eric Bohm)
09:00 - 09:30	Talk	Addressing the Challenges of Heterogeneous Computing in NAMD	David Hardy
09:30 - 10:00	Talk	Towards Performance Portability in NAMD with oneAPI	Tareq Malas and Jaemin Choi
10:00 - 10:30	Talk	Scalable GW software for excited electrons using OpenAtom	Kayahan Saritas, Kavitha Chandrasekar
10:30 - 11:00	Break
Session: Keynote II (Chair: Laxmikant V. Kale)
11:00 - 12:00	Keynote	A Tale of Two Cultures: Challenges in Developing Tools for Writing Parallel Software	Alex Aiken
12:00 - 13:00	Lunch Break
Session: Adaptive MPI (Chair: Matthias Diener)
13:00 - 13:30	Talk	CharmMPI: From Research Code to Production Workhorse	Evan Ramos
13:30 - 14:00	Talk	Communication Optimizations for Adaptive MPI	Sam White
14:00 - 14:15	Break
Session: Performance Analysis (Chair: Ronak Buch)
14:15 - 14:45	Talk	Performance Evaluation of Python Parallel Programming Models: Charm4Py and mpi4py	Zane Fink
14:45 - 15:15	Talk	A synthetic tool for analysing adaptable workloads	Iker Martín-Álvarez, José Ignacio Aliaga, María Isabel Castillo, Sergio Iserte and Rafael Mayo
15:15 - 15:30	Break
Session: GPU Computing (Chair: Zane Fink)
15:30 - 16:00	Talk	Using concurrent GPU streams to support independent progress of asynchronous objects	Phil Miller, Jonathan Lifflander, Nicole Slattengren, Matthew Bettencourt and Roger Pawlowski
16:00 - 16:30	Talk	CharminG: A GPU-resident Runtime System	Jaemin Choi
16:30 - 16:45	Break
Session: Applications & Charm++ (Chair: TBA)
16:45 - 17:15	Talk	ExaM2M: Scalable and Adaptive Mesh-to-Mesh Transfer	Eric Mikida
17:15 - 17:30	Short Talk	CkLUN: Virtualizing Fortran File Operations in ADCIRC	Eric Bohm
17:30 - 17:45	Short Talk	New Charm++ Features & Discussions	PPL
17:45 - 18:00	Closing Remarks