When scaling your workload is a matter of saving lives

By Werner Vogels, CTO, Amazon.com

April 20, 2020

On March 16, 2020, at 9:26 PM, I received an urgent email from my friend DJ Patil, former White House Chief Data Scientist, Head of Technology for Devoted Health, a Senior Fellow at the Belfer Center at the Harvard Kennedy School, and Advisor to Venrock Partners. You don’t get that many titles after your name unless you’re pretty good at something. For DJ, that “something” is math and computer science.

DJ was writing to me from the California crisis command center. He explained that he was working with governors from across the country to model the potential impact of COVID-19 for scenario planning. He wanted to help them answer critical questions, like “How many hospital beds will we need?” and “Can we reduce the spread if we temporarily close places where people gather?” and “Should we issue a shelter-in-place order and for how long?” While nobody can predict the future, modeling the virus with all the factors they did know was their best shot at helping leaders make informed decisions, which would impact hundreds of thousands of lives.

DJ assembled a team of volunteers that consisted of some of the brightest minds from across Silicon Valley and the country. As though following a call to arms, these professionals came together, in a personal capacity, to fight COVID-19 the best way they knew how: with data.

The good news is that they had a model. And not just any model. DJ and his team have been working with one that had been primarily developed by the world renowned Johns Hopkins Bloomberg School of Public Health (JHSPH). This model, which is an open-source project, uses state or county population numbers along with transportation data to model the number of people who potentially would be exposed, infected, and/or hospitalized. The model also considers virus spread based on a variety of non-pharmaceutical interventions, including shutting down schools and parks and issuing quarantine orders.

However, the model was running on the JHSPH on-premises infrastructure and the model pipeline couldn’t scale to run a large number of scenarios simultaneously or to meet the needs of the country (and potentially the world). It was too slow. To get the required scale and speed, DJ and his team needed to run the model in the cloud—so they moved their on-premises code to AWS. This led to another challenge: The code hadn’t been written initially with the cloud in mind, so it couldn’t fully take advantage of the scale and optimizations possible with AWS. As a result, DJ’s team spent one week porting and running a single scenario for California, which still wasn’t fast enough.

Imagine how long it would take to scale the pipeline for 49 more states. It would require at least months of work. Adding multiple scenarios with different variables would delay it even more. DJ’s team didn’t have that kind of time. In the words of Dr. Anthony Fauci, the director of the National Institute of Allergy and Infectious Diseases, “You don’t make the timeline. The virus makes the timeline.”

DJ’s team needed to be able to onboard the model pipeline and run a full report in hours rather than weeks or months. That’s why DJ reached out to me. He wanted our help optimizing the model pipeline for the cloud. Presented with an opportunity to help my friend and support a project that could save lives, I immediately said “yes.”

When I sounded the alarm internally, something amazing happened. People from across the company volunteered to help because they knew that they had the expertise that the project needed. Nobody asked, “Who is responsible for this customer?” or “Who has the bandwidth to work on this project?” Everyone sprang into action immediately.

We wanted to get that model pipeline running like it had a jet engine. The first thing we had to do was create an architecture that would fuel the model every step of the way. We began by profiling the code and re-compiling key numerical libraries.

Next, we had to help optimize the model pipeline. That’s when our specialist team, including the high performance computing (HPC) group, stepped in. These professionals help organizations solve some of the biggest data-related problems and tame the largest workloads, like in genomics, computational chemistry, machine learning, and autonomous vehicle simulation. They worked with members of DJ’s team, JHSPH, and some state employees non-stop through two weekends to optimize the model pipeline through re-architecture and deployment on AWS.

We contributed to the open-source JHSPH model to enable a Continuous Integration & Deployment (CI/CD) pipeline for container deployments on Amazon Elastic Container Registry. Also, we orchestrated scalable deployment strategies on AWS Elastic Container Service through AWS Batch and integrated several other services, including Amazon S3 and Amazon EC2 Auto Scaling. A high level of coordination with AWS was critical to bringing all these technologies together.

Figure 1: Architecture for the Scalable COVID Scenario Pipeline using AWS Batch

What was the end result?

DJ’s team has reduced the time to on-board the model pipeline and generate a full report from one week for one scenario to under 12 hours for multiple scenarios. Now, the new model created by JHSPH is being rolled out on AWS to all 50 states and internationally to help with making decisions that directly impact the global spread of COVID-19.

But this isn’t where the story ends—it’s where it begins. While the spread of the virus is now global, we have yet to grasp its full impact on human society. As we learn more about the spreading conditions of the virus, iterating over existing models will be crucial. There will be many more sleepless nights for everyone working on this initiative as they continue to scale the model to more states and countries as well as analyze the effect of mitigation strategies.

If you have relevant skills in technology, communications, and/or operations, please consider joining the U.S. Digital Response team. It’s a volunteer-run, non-partisan effort to help federal, state, and local government with technology, data, design, operations, communications, project management, and other needs during the COVID-19 crisis. https://www.usdigitalresponse.org/

 

Subscribe to HPCwire's Weekly Update!

Be the most informed person in the room! Stay ahead of the tech trends with industy updates delivered to you every week!

Intel Speeds NAMD by 1.8x: Saves Xeon Processor Users Millions of Compute Hours

August 12, 2020

Potentially saving datacenters millions of CPU node hours, Intel and the University of Illinois at Urbana–Champaign (UIUC) have collaborated to develop AVX-512 optimizations for the NAMD scalable molecular dynamics cod Read more…

By Rob Farber

Intel’s Optane/DAOS Solution Tops Latest IO500

August 11, 2020

Intel’s persistent memory technology, Optane, and its DAOS (Distributed Asynchronous Object Storage) stack continue to impress and gain market traction. Yesterday, Intel reported an Optane and DAOS-based system finishe Read more…

By John Russell

Summit Now Offers Virtual Tours

August 10, 2020

Summit, the second most powerful publicly ranked supercomputer in the world, now has a virtual tour. The tour, implemented by 3D platform Matterport, allows users to virtually “walk” around the massive supercomputer Read more…

By Oliver Peckham

Supercomputer Simulations Examine Changes in Chesapeake Bay

August 8, 2020

The Chesapeake Bay, the largest estuary in the continental United States, weaves its way south from Maryland, collecting waters from West Virginia, Delaware, DC, Pennsylvania and New York along the way. Like many major e Read more…

By Oliver Peckham

Student Success from ‘Scratch’: CHPC’s Proof is in the Pudding

August 7, 2020

Happy Sithole, who directs the South African Centre for High Performance Computing (SA-CHPC), called the 13th annual CHPC National conference to order on December 1, 2019, at the Birchwood Conference Centre in Kempton Pa Read more…

By Elizabeth Leake

AWS Solution Channel

University of Adelaide Provides Seamless Bioinformatics Training Using AWS

The University of Adelaide, established in South Australia in 1874, maintains a rich history of scientific innovation. For more than 140 years, the institution and its researchers have had an impact all over the world—making vital contributions to the invention of X-ray crystallography, insulin, penicillin, and the Olympic torch. Read more…

Intel® HPC + AI Pavilion

Supercomputing the Pandemic: Scientific Community Tackles COVID-19 from Multiple Perspectives

Since their inception, supercomputers have taken on the biggest, most complex, and most data-intensive computing challenges—from confirming Einstein’s theories about gravitational waves to predicting the impacts of climate change. Read more…

New GE Simulations on Summit to Advance Offshore Wind Power

August 6, 2020

The wind energy sector is a frequent user of high-power simulations, with researchers aiming to optimize wind flows and energy production from the massive turbines. Now, researchers at GE are preparing to undertake a lar Read more…

By Oliver Peckham

Intel’s Optane/DAOS Solution Tops Latest IO500

August 11, 2020

Intel’s persistent memory technology, Optane, and its DAOS (Distributed Asynchronous Object Storage) stack continue to impress and gain market traction. Yeste Read more…

By John Russell

Summit Now Offers Virtual Tours

August 10, 2020

Summit, the second most powerful publicly ranked supercomputer in the world, now has a virtual tour. The tour, implemented by 3D platform Matterport, allows use Read more…

By Oliver Peckham

Research: A Survey of Numerical Methods Utilizing Mixed Precision Arithmetic

August 5, 2020

Within the past years, hardware vendors have started designing low precision special function units in response to the demand of the machine learning community Read more…

By Hartwig Anzt and Jack Dongarra

Implement Photonic Tensor Cores for Machine Learning?

August 5, 2020

Researchers from George Washington University have reported an approach for building photonic tensor cores that leverages phase change photonic memory to implem Read more…

By John Russell

HPE Keeps Cray Brand Promise, Reveals HPE Cray Supercomputing Line

August 4, 2020

The HPC community, ever-affectionate toward Cray and its eponymous founder, can breathe a (virtual) sigh of relief. The Cray brand will live on, encompassing th Read more…

By Tiffany Trader

Machines, Connections, Data, and Especially People: OAC Acting Director Amy Friedlander Charts Office’s Blueprint for Innovation

August 3, 2020

The path to innovation in cyberinfrastructure (CI) will require continued focus on building HPC systems and secure connections between them, in addition to the Read more…

By Ken Chiacchia, Pittsburgh Supercomputing Center/XSEDE

Nvidia Said to Be Close on Arm Deal

August 3, 2020

GPU leader Nvidia Corp. is in talks to buy U.K. chip designer Arm from parent company Softbank, according to several reports over the weekend. If consummated Read more…

By George Leopold

Intel’s 7nm Slip Raises Questions About Ponte Vecchio GPU, Aurora Supercomputer

July 30, 2020

During its second-quarter earnings call, Intel announced a one-year delay of its 7nm process technology, which it says it will create an approximate six-month shift for its CPU product timing relative to prior expectations. The primary issue is a defect mode in the 7nm process that resulted in yield degradation... Read more…

By Tiffany Trader

Supercomputer Modeling Tests How COVID-19 Spreads in Grocery Stores

April 8, 2020

In the COVID-19 era, many people are treating simple activities like getting gas or groceries with caution as they try to heed social distancing mandates and protect their own health. Still, significant uncertainty surrounds the relative risk of different activities, and conflicting information is prevalent. A team of Finnish researchers set out to address some of these uncertainties by... Read more…

By Oliver Peckham

Supercomputer-Powered Research Uncovers Signs of ‘Bradykinin Storm’ That May Explain COVID-19 Symptoms

July 28, 2020

Doctors and medical researchers have struggled to pinpoint – let alone explain – the deluge of symptoms induced by COVID-19 infections in patients, and what Read more…

By Oliver Peckham

Nvidia Said to Be Close on Arm Deal

August 3, 2020

GPU leader Nvidia Corp. is in talks to buy U.K. chip designer Arm from parent company Softbank, according to several reports over the weekend. If consummated Read more…

By George Leopold

Intel’s 7nm Slip Raises Questions About Ponte Vecchio GPU, Aurora Supercomputer

July 30, 2020

During its second-quarter earnings call, Intel announced a one-year delay of its 7nm process technology, which it says it will create an approximate six-month shift for its CPU product timing relative to prior expectations. The primary issue is a defect mode in the 7nm process that resulted in yield degradation... Read more…

By Tiffany Trader

Supercomputer Simulations Reveal the Fate of the Neanderthals

May 25, 2020

For hundreds of thousands of years, neanderthals roamed the planet, eventually (almost 50,000 years ago) giving way to homo sapiens, which quickly became the do Read more…

By Oliver Peckham

10nm, 7nm, 5nm…. Should the Chip Nanometer Metric Be Replaced?

June 1, 2020

The biggest cool factor in server chips is the nanometer. AMD beating Intel to a CPU built on a 7nm process node* – with 5nm and 3nm on the way – has been i Read more…

By Doug Black

HPE Keeps Cray Brand Promise, Reveals HPE Cray Supercomputing Line

August 4, 2020

The HPC community, ever-affectionate toward Cray and its eponymous founder, can breathe a (virtual) sigh of relief. The Cray brand will live on, encompassing th Read more…

By Tiffany Trader

Neocortex Will Be First-of-Its-Kind 800,000-Core AI Supercomputer

June 9, 2020

Pittsburgh Supercomputing Center (PSC - a joint research organization of Carnegie Mellon University and the University of Pittsburgh) has won a $5 million award Read more…

By Tiffany Trader

Leading Solution Providers

Contributors

Nvidia’s Ampere A100 GPU: Up to 2.5X the HPC, 20X the AI

May 14, 2020

Nvidia's first Ampere-based graphics card, the A100 GPU, packs a whopping 54 billion transistors on 826mm2 of silicon, making it the world's largest seven-nanom Read more…

By Tiffany Trader

Australian Researchers Break All-Time Internet Speed Record

May 26, 2020

If you’ve been stuck at home for the last few months, you’ve probably become more attuned to the quality (or lack thereof) of your internet connection. Even Read more…

By Oliver Peckham

15 Slides on Programming Aurora and Exascale Systems

May 7, 2020

Sometime in 2021, Aurora, the first planned U.S. exascale system, is scheduled to be fired up at Argonne National Laboratory. Cray (now HPE) and Intel are the k Read more…

By John Russell

‘Billion Molecules Against COVID-19’ Challenge to Launch with Massive Supercomputing Support

April 22, 2020

Around the world, supercomputing centers have spun up and opened their doors for COVID-19 research in what may be the most unified supercomputing effort in hist Read more…

By Oliver Peckham

Joliot-Curie Supercomputer Used to Build First Full, High-Fidelity Aircraft Engine Simulation

July 14, 2020

When industrial designers plan the design of a new element of a vehicle’s propulsion or exterior, they typically use fluid dynamics to optimize airflow and in Read more…

By Oliver Peckham

John Martinis Reportedly Leaves Google Quantum Effort

April 21, 2020

John Martinis, who led Google’s quantum computing effort since establishing its quantum hardware group in 2014, has left Google after being moved into an advi Read more…

By John Russell

$100B Plan Submitted for Massive Remake and Expansion of NSF

May 27, 2020

Legislation to reshape, expand - and rename - the National Science Foundation has been submitted in both the U.S. House and Senate. The proposal, which seems to Read more…

By John Russell

Google Cloud Debuts 16-GPU Ampere A100 Instances

July 7, 2020

On the heels of the Nvidia’s Ampere A100 GPU launch in May, Google Cloud is announcing alpha availability of the A100 “Accelerator Optimized” VM A2 instance family on Google Compute Engine. The instances are powered by the HGX A100 16-GPU platform, which combines two HGX A100 8-GPU baseboards using... Read more…

By Tiffany Trader

  • arrow
  • Click Here for More Headlines
  • arrow
Do NOT follow this link or you will be banned from the site!
Share This