NYU Greene Supercomputer: Celebrating a Year of Breakthrough Research

By Keith Allison | November 16, 2021

the greene supercomputer racks

Making an Impact in COVID-19 Research, Political Science, Climate Change, and Beyond

In November 2020, NYU President Andrew Hamilton hosted an online event with University researchers and technologists celebrating the launch of the NYU Greene high performance computing cluster—NYU's newest supercomputer and one of the fastest supercomputers in higher education. Greene's impact on research was immediate. It was an essential resource for Tamar Schlick, Professor of Chemistry, Mathematics and Computer Science, who used Greene's computing power in her study of the RNA frameshifting element (FSE) of COVID-19. Understanding that aspect of the virus was an essential piece of understanding how to combat it.1

Professor Schlick's groundbreaking work was only the beginning. The speed and power of the Greene high performance computing (HPC) cluster enabled NYU to perform 10 million high performance computing jobs in 2021, up from the 7 million that were performed in 2020 using the University's previous supercomputer, Prince. Greene supports a variety of job types and sizes, including jobs requiring multiple CPU cores, multiple GPU cards, terabytes of memory, as well as single-core serial jobs. Greene has greatly enhanced research at NYU across multiple disciplines, including computational chemistry, genomics, political and behavioral science, artificial intelligence, climatology, VR, and more.

Additionally, the wait time to initiate a job also significantly decreased, from 2.87 hours in 2020 to just .81 hours in 2021 using Greene. In a nutshell: Greene has enabled researchers to complete bigger, more complex tasks in much less time.

Building a Supercomputer during COVID—for COVID Research

Building a supercomputer and bringing it online is a complicated process under the best of circumstances. Obviously, 2020 did not present the best of circumstances. Work on Greene began in 2019, including the construction of a new, more environmentally-sustainable data center. But the majority of physical tasks, everything from delivering the component pieces to putting it all together, fell after March 2020—when NYU HPC partner, Lenovo, and the entire global supply chain that manufactures and ships equipment had to close physical locations due to the pandemic. However, given the type of research that was planned for Greene, David Ackerman, NYU Associate Vice President of Research Technology, worked with Lenovo to navigate the complicated landscape and make progress in a safe way.

“It had to be up and running because many of the researchers using the system were going to be doing COVID research,” said Scott Tease, General Manager, HPC and AI at Lenovo. “NYU sent out a challenge to Lenovo and their other vendors asking for our support to get the system up during such difficult times—and we responded."2

The results of that effort, which spanned the globe during an extremely difficult time, were broadcast to the world on November 17, 2020, when President Hamilton officially launched Greene with a virtual ribbon cutting via Zoom which included a number of NYU researchers, members of the press, and the NYU Greene project team. With Greene operational, Professor Schlick was then able to conduct vital COVID-19 research that would not have been possible with the University's previous high performance computing infrastructure.  

In-progress and planned work from other NYU researchers was migrated from the former supercomputing cluster to Greene, and before long, everything was up and running faster than ever.

"The fact that Greene could process over 40% more research jobs with 70% less wait time is awesome," says Ackerman. "And the Center for Data Science has already funded state-of-the-art water cooled Nvidia GPUs that will only increase Greene's awesomeness!"

Among other projects, the new supercomputer enabled researchers Joshua Tucker, Richard Bonneau, and Jonathan Nagler at the Center for Social Media and Politics (CSMaP) to analyze terabytes of social interactions, such as “likes” and retweets from the Twitter Decahose. These analyses enabled them to answer a multitude of social science questions about how voters develop opinions in a democracy, especially in relation to COVID-19 political pronouncements and COVID-induced economic hardships. The Greene HPC cluster made it possible for them to analyze the network information (URL-sharing, retweeting, and following behavior) of politically-interested Twitter users to look at how ideology affects information exchange and relations to Twitter influencers.  

Sustainability, Year Two, and Beyond

As people have returned to campus during the fall 2021 semester, Greene's role in research has continued to expand. The High Speed Research Network (HSRN) was launched in tandem with Greene, providing an extremely fast, low-latency computer network dedicated entirely to research use. The new research computing data center has also resulted in a substantial evolution toward NYU's overall sustainability goals.

Some of the innovations that enable Greene and the new data center to increase computing power while decreasing the impact on the environment include:

  • a more efficient power usage effectiveness (PUE) rate of 1.35 or less (Greene's predecessor, Prince, scored a PUE of 2);
  • cutting-edge water-cooled compute nodes that remove 95% of the heat without any air conditioning, enabling portions of Greene to achieve a 1.08 PUE;
  • white equipment racks that reduce heat consumption by 15-20%. A “hot aisle,” where computers are arranged back-to-back, enables more efficient heat trapping and ventilation through the roof.

In addition, oceanographic researchers at NYU's Courant Institute are using Greene to study climate change,3 meaning that Greene can contribute to a more sustainable future in multiple ways. As Greene enters its second year, the possibilities for how it can contribute to NYU's research mission continue to expand, and researchers such as Professor Schlick are discovering new ways to process and interpret existing data that would not have been possible with the amount of computing power previously available to them.

“It’s of paramount importance that we continue to provide our researchers across the disciplinary spectrum with the resources they need to expand knowledge beyond what we may not even be able to conceive of as yet," said Andrew Hamilton, President of New York University. "But for the sake of our planet, we must also find ways to maintain this computing trajectory while simultaneously moving away from our reliance on fossil-fuel consumption. Greene is a wonderful example of how we can achieve both these goals.”4

Notes

  1. "Schlick Group Identifies RNA Frameshifting Element as Target for Battling COVID-19" (NYU Department of Chemistry) and "Artificial Intelligence Is Predicting Future COVID-19 Strains, Developing Treatments" (Observer)
  2. "The Show Must Go On: Installing a Supercomputer During Peak Pandemic" (Lenovo StoryHub)
  3. "Better, Faster, Greener - NYU Launches One of The Fastest Supercomputers in Higher Ed' (NYU News)
  4. See "Ocean Turbulence, Sea Levels, and Machine Learning" and "Climate Change, Machine Learning, and High Performance Computing" (The Download)