Best Use of HPC in Life Sciences

Readers’ Choice:

ORNL researchers conducted simulations on Summit supercomputer to screen for potential drugs that target SARS-CoV-2, finding 77 potential small molecule targets.

Editors’ Choice:

St. Jude researchers used DDN storage technology and Nvidia A100 to develop a powerful computational method to discover regulatory noncoding variants in cancer by integrating whole-genome and transcriptome sequencing data from a single cancer sample.

Best Use of HPC in Physical Science

Readers’ Choice:

The European Centre for Medium-Range Weather Forecasts (ECMWF) worked with ORNL to run an unprecedented weather simulation on Summit: the entire atmosphere of the planet at a one-kilometer resolution for a four-month season.

Editors’ Choice:

An NCSA-led team used PSC’s Bridges system and TACC’s Stampede2 – both XSEDE-allocated resources – to obtain a correction factor that will allow for fast, accurate simulations of neutron star mergers for gravitational wave observatories.

Best Use of HPC in Response to Societal Plights
(COVID-19 focused research/ programs)

Readers’ Choice: 

From early in the pandemic, Folding@home marshalled millions of volunteer CPUs and GPUs around the world to conduct various simulations of COVID-19’s essential proteins in service of efforts to develop vaccines, therapeutics and a better understanding of the virus.

Editors’ Choice (tie): 

AMD and partners launched a COVID-19 HPC Fund that has, to date, contributed 12 petaflops of cloud and on-premises computing resources to 21 recipient organizations fighting COVID-19.

When the pandemic struck, RIKEN managed to launch Fugaku – now the most powerful publicly ranked supercomputer in the world – a full year ahead of schedule in order to make it available for COVID-19 research; the supercomputer is now hosting research ranging from SARS-CoV-2 protein analysis to simulation of viral droplets in trains.

Best Use of HPC in Energy

Readers’ Choice: 

A team at Argonne National Laboratory used an in-house fluid-thermal simulation code on Argonne’s Theta supercomputer to run the largest-ever combustion engine flow simulation.

Editors’ Choice:

Researchers from Georgia Tech and the Hanoi University of Science and Technology used SDSC’s Comet system and TACC’s Stampede2 system – both XSEDE-allocated resources – to identify four lead-free candidates for a more efficient, cheaper alternative to the silicon used in solar panels.

Best Use of HPC in Industry
(Automotive, Aerospace, Manufacturing, Chemical, etc.)

Readers’ Choice:

Aramco Research Center (Detroit), Argonne National Lab, and Convergent Science researchers used modeling on ACLF’s Theta supercomputer to resolve micron-scale manufacturing defects in fuel injector geometries for the first time. The broad goal is development of cleaner propulsion systems.

Editors’ Choice:

Covestro AG uses HPC cluster, including resources from Trubomole (quantum chem.), Lenovo (LICO), and Nvidia (DGX), to conduct virtual testing of materials and speed development of insulation foams and transparent polycarbonate plastics.

Best Use of HPC in Financial Services

Readers’ Choice:

Penguin Tundra OCP platform provides low latency big memory capabilities to MemX, a new independently owned, technology-driven stock exchange.

Editors’ Choice:

Moody’s Actuarial Risk as a Service relies on Microsoft Azure HPC technologies to deliver actuarial results 10x faster.

Best Use of High Performance Data Analytics & Artificial Intelligence

Readers’ Choice: 

University of Birmingham researchers, working on BlueBEAR (IBM/Power9 AI cluster), used data from Kepler Space Telescope to develop a data-driven method to improve understanding of stars based on sound waves that are trapped within their interiors.

Editors’ Choice:

Carnegie Mellon researchers working on Bridges and Bridges-AI used AI-quantum mechanical methods to accelerate and more accurately screen novel chemical compounds in search for small molecule antiviral candidates against COVID-19.

Best HPC Storage Product or Technology

Readers’ Choice: 

Intel DAOS with Lenovo ThinkSystem SR630 Servers

Editors’ Choice:

DDN EXAScaler

Best AI Product or Technology

Readers’ and Editors’ Choice:

Nvidia DGX A100

Best Use of HPC in the Cloud (Use Case)

Readers’ Choice:

Harvard Medical School used VirtualFlow on Google Cloud to conduct one of the largest virtual drug screenings ever performed to find compounds active against SARS-CoV-2, using 75 million CPU hours.

Editors’ Choice:

Researchers from the IceCube Neutrino Observatory with SDSC and the Open Science Grid leveraged 51,000 cloud GPUs simultaneously – the largest GPU “cloud burst” in history – to process data from neutrino sensors buried in the ice of the South Pole.

Best HPC Cloud Platform

Readers’ Choice: 

Amazon Web Services

Editors’ Choice:  

Microsoft Azure Cloud

Best HPC Server Product or Technology

Readers’ Choice:

NVIDIA A100 GPU

Editors’ Choice:

AMD 2^nd Gen EPYC “Rome” processors

Best HPC Programming Tool or Technology

Readers’ Choice:

Singularity

Editors’ Choice:

NVIDIA CUDA Toolkit

Best HPC Interconnect Product or Technology

Readers’ Choice and Editors’ Choice:

NVIDIA Mellanox HDR 200G InfiniBand

Best HPC Collaboration
(Academia / Government / Industry)

Readers’ Choice:

NHS organizations, UK public health agencies, Wellcome Sanger Institute (using DDN storage) and more than a dozen academic partners formed the COVID-19 Genomics UK (COG-UK) Consortium, which runs on Cloud Infrastructure for Microbial Bioinformatics (CLIMB) – on systems implemented by Birmingham and Cardiff University in partnership with Dell and Lenovo – to conduct real-time sequencing of tens of thousands of samples from positive COVID-19 cases in the UK.

Editors’ Choice:

COVID-19 HPC Consortium – created by the DOE and IBM with the White House OSTP, NSF, XSEDE and dozens of partners – connects researchers to an unprecedented international array of supercomputing resources spanning academia, government and industry in response to the COVID-19 pandemic.

Top Energy-Efficient HPC Achievement

Readers’ Choice:

LRZ introduced the “Wintermute” module to its Data Center Data Base (DCDB) facility monitoring system, allowing users to easily and efficiently deploy online operational data analytics for large-scale HPC installations and automatically identify optimization opportunities.

Editors’ Choice:

Sandia National Laboratories installed a novel thermosyphon cooling system for its datacenters, saving 554,000 gallons of water and 195 megawatt-hours of electricity over just six months.

Top HPC-Enabled Scientific Achievement

Readers’ Choice:

A research team led by Australian National University used the SuperMUC-NG system at LRZ to run the largest-ever magneto-hydrodynamic simulation, illuminating the role of turbulence in the birthing of stars.

Editors’ Choice:

A team at ORNL used the Summit supercomputer to compare genes of cells from COVID-19-infected patients, uncovering signs of a “bradykinin storm” phenomenon that may help to explain the wide range of symptoms associated with COVID-19.

 Top Supercomputing Achievement

Readers’ Choice: 

Nvidia Selene supercomputer debuts as seventh-fastest system with new GPU architecture; the 27 petaflops system was stood up in less than a month.

Editors’ Choice: 

Supercomputer Fugaku – a joint project of Riken, Fujitsu and Arm – is first-Arm based #1 system, sweeping Top500, HPCG, HPL-AI and Graph500.

Top 5 New Products or Technologies to Watch

Readers’ Choice:
NVIDIA A100 “Ampere” GPU

AMD 2^nd Gen EPYC “Rome” processors

NVIDIA HPC SDK

MemVerge Big Memory

BeeGFS

 Editors’ Choice:
NVIDIA A100 “Ampere” GPU

AMD 2nd Gen EPYC “Rome” processors

Fujitsu A64FX Arm processors

Cerebras CS-1 system

DAOS with Intel Optane Persistent Memory

Top 5 Vendors to Watch

Readers’ Choice:
NVIDIA

AMD

Intel

Arm

Lenovo

Editors’ Choice:
NVIDIA

AMD

Intel

HPE

Google Cloud

Workforce Diversity Leadership Award

Readers’ Choice (tie): 

The National Center for Women & Information Technology (NC-WIT) is a national nonprofit organization focused on women’s participation in computing across the entire ecosystem.

STEM-Trek is a nonprofit organization that supports travel, mentoring and professional development for science, technology, engineering, and mathematics (STEM) scholars from demographic groups that are typically underrepresented in their use of advanced cyberinfrastructure to accelerate scientific discovery.

Editors’ Choice:

Berkeley Lab’s IDEA program has a mission to promote, advance, integrate and operationalize inclusion, diversity, equity and accountability (IDEA) approaches and best practices within all levels of the organization.

Outstanding Leadership in HPC

Readers’ Choice: 

Mateo Valero, Founder and Director of the Barcelona Supercomputing Centre, which hosts the MareNostrum supercomputer – HPC pioneer since 1990 and the driving force behind the renaissance of European HPC independence.

Editors’ Choice: 

Yutaka Ishikawa, Lead for Japan’s Flagship 2020 project at Riken & Satoshi Matsuoka, Director, Riken for bringing the Fugaku project to fruition, standing up the world’s number one supercomputer, and supporting COVID-19 research a full year earlier than planned