Virtual ICM Seminar with Ben Whitney, "MGARD: Hierarchical Compression of Scientific Data," to Be Held Feb. 18

Feb. 11, 2021 — The Interdisciplinary Centre for Mathematical and Computational Modelling (ICM) at the University of Warsaw invites enthusiasts of HPC and all people interested in challenging topics in Computer and Computational Science to the ICM Seminar in Computer and Computational Science that will be held on February 18, 2021 (7:00 p.m. CEST). The event is free.

Ben Whitney will present the lecture entitled: “MGARD: Hierarchical Compression of Scientific Data.” His speech will be an introduction to MGARD (MultiGrid Adaptive Reduction of Data) which is one of a new generation of scientific data compressors aiming to win the trust of domain scientists with rigorous mathematical underpinnings and guaranteed error bounds.

To register, visit https://supercomputingfrontiers.eu/2021/seminars/

Ben Whitney is a postdoc in the Computer Science and Mathematics Division at Oak Ridge National Laboratory. He received his PhD in Applied Mathematics from Brown University under the supervision of Mark Ainsworth. After graduating in 2018, he took a postdoc position at Brown and then came to ORNL in 2019. He works in compression methods for scientific data and scientific software development.

Abstract:

Lossy compression will be an essential tool in the efforts to manage and understand the imposing volumes of data that will be produced by exascale simulations and experiments. While ad hoc methods like spatial subsampling are easy to implement and understand, the absence of guaranteed bounds on the errors they incur hinders their widespread adoption by the scientific community. MGARD (MultiGrid Adaptive Reduction of Data) is one of a new generation of scientific data compressors aiming to win the trust of domain scientists with rigorous mathematical underpinnings and guaranteed error bounds. This tutorial will be an introduction to MGARD, starting with the hierarchical decomposition algorithm comprising the core of the method and its ties to the theories of finite element methods, multigrid solvers, and wavelet analysis. We will then discuss quantization techniques enabling bounds on the compression error as measured both in the original data and in quantities of interest computed from the data. We will conclude with a variety of numerical examples and a demonstration of the use of MGARD’s publically available implementation.

Virtual ICM Seminars in Computer and Computational Science are a continuation of the Supercomputing Frontiers Europe conference, which took place virtually in March 2020 and will be back with the new HPC trends to explore in the summer 2021.

Dates to Remember:

Supercomputing Frontiers Europe 2021 Conference: 19-23 July 2021
Call for Papers deadline: 16 May 2021

Worldwide Open Science online meetings in HPC, Artificial Intelligence, Quantum Computing, BigData, IoT, computer and data networks are a place to meet and discuss with such personalities as David Bader (New Jersey Institute of Technology; advisor to the White House), Stephen Wolfram (Founder & CEO, Wolfram Research), Alan Edelman (MIT), Aneta Afelt (ICM, Espace-DEV, IRD Montpellier France), Simon Mutch (University of Melbourne) or Scott Aaronson (University of Texas at Austin). For the listing of all ICM seminars please check this link with recordings.

Since March 2020, over 2400 people have participated in virtual meetings organized by ICM (SCFE + 14 seminars) from almost all the time zones. The supercomputingfrontiers.eu website with access to the SCFE and seminars’ recordings, has been visited by more than 15k visitors from 122 countries. The organizer of meetings with outstanding scientists is the Interdisciplinary Centre for Mathematical and Computational Modelling – ICM University of Warsaw.

About the Interdisciplinary Centre for Mathematical and Computational Modelling (ICM), University of Warsaw (UW)

The Interdisciplinary Centre for Mathematical and Computational Modelling (ICM) University of Warsaw is involved in interdisciplinary scientific research based on mathematical modeling, computer simulations and modeling, multi-scale and large-scale calculations. The popular meteo.pl weather portal has been established by using ICM computed weather forecasts. ICM researchers study problems related to civil aviation (collaborating with ICAO), modeling of social processes and most recently working on ICM Epidemiological Model for the COVID-19 epidemic in Poland. ICM took part in securing access for Polish scientists to the entire body of scientific literature, including over 8,000 journal titles, by maintaining the Virtual Library of Science. ICM networking team has participated in a number of cutting edge networking solutions, both for high throughput and low latencyrequirements. Recently, ICM engineers have established a production 100Gbps connection over 12,375 4 miles CAE-1 (Collaboration Asia Europe-1) line between Warsaw and Singapore.