[hpc-announce] Opportunity for lightning talks at UrgentHPC SC20 workshop

Nick Brown n.brown at epcc.ed.ac.uk
Mon Sep 28 15:57:15 CDT 2020

Are you or your project undertaking work in the use of HPC to run urgent
workloads in responding to disasters? If so then there is the
opportunity to present a short, 5 minute, lightning talk at the Second
international workshop on HPC for Urgent Decision making (UrgentHPC).

This will run in conjunction with SC20: the International Conference for
High Performance Computing, Networking, Storage and Analysis, on Friday
the 13th of November between 14:30 and 18:30 (all times are Eastern Times.)

If this is something that you would be interested in doing then please
email us at urgenthpc-workshop at mlist.is.ed.ac.uk by Monday 5th of
October with a short one or two paragraph description of your proposed
lightening talk.

Workshop Scope
Responding to disasters such as COVID-19, wildfires, hurricanes, extreme
flooding, earthquakes, tsunamis, and accidents; technological advances
are creating exciting new opportunities that have the potential to move
HPC well beyond traditional computational workloads. Whilst HPC has a
long history of simulating disasters, what’s missing to support
emergency, urgent, decision making is fast, real-time acquisition of
data and the ability to guarantee time constraints.

Our ability to capture data continues to grow very significantly, and
combining high velocity data and live analytics with HPC models can aid
in urgently responding to real-world problems, ultimately saving lives
and reducing economic loss. It’s not just responding to disasters, but
also making urgent decisions addressing more general issues such as
human health emergencies and global diseases. The challenges here are
significant, but if HPC can be proven as a tool in responding to these
real-world issues, the impact for our community is huge.

Leveraging HPC for urgent decision making requires expertise in a wide
range of areas, from dealing with real-time data, to experience in
generating results within a specific time frame (real-time constraints),
and generating visualisations enabling front-line decision makers to
make correct choices first time, every time. It isn’t just technical
challenges, but also policy issues that also need to be considered such
as utilising our HPC machines in a more interactive manner to enable the
urgent exploration of numerous disaster responses.

This workshop will bring together stakeholders, researchers and
practitioners from across the HPC community to identify and tackle
issues involved in using HPC for urgent decision making. Success
stories, case-studies and challenges will be shared, with the goal of
further building up a community around leveraging HPC as an important
tool in urgently responding to disasters and societal challenges.

Workshop topics
* Use of HPC and big-data for tackling COVID-19
* Example use-cases and case-studies that use HPC for urgent decision making
* Techniques for integrating HPC workflows with real-time data
* Approaches to verify and validate unreliable real-time data, for
instance from sensors, IoT and satellites
* System design for data reduction and pre-processing at source, for
instance using edge computing and heterogeneous resources such as FPGAs
* The use of data formats and conversion techniques to support the
handling of data from numerous and diverse sources
* Algorithmic techniques to guarantee result generation in specific time
frames, such as result refinement which generates more accurate results
as time progresses
* Studies of leveraging HPC for workloads with real-time time constraints
* Changes to existing HPC technologies and policies that are required to
support using HPC interactively
* The ability for HPC codes to adapt their resource requirements
dynamically, for instance via elastic compute
* Visualisation and presentation techniques to support rapid and
accurate urgent decision making by the end user
* Reduction and feature extraction of results to highlight critical
issues of interest
* Complimenting results with provenance data for additional context and
* Data analysis techniques for making urgent decisions in response to
The University of Edinburgh is a charitable body, registered in Scotland, with registration number SC005336.

More information about the hpc-announce mailing list