Advancing the Frontiers of Math and Science at the Flatiron Institute

BrandPost By Tony Rea
Dec 04, 2019
AnalyticsBig DataHadoop

An interdisciplinary scientific hub is driving progress in computational sciences with a high performance computing system that can quickly process huge amounts of data to accelerate time to insight.

elusiveorigi credit joseph insley argonne leadership computing facility
Credit: Joseph Insley/Argonne Leadership Computing Facility

Flatiron Institute is a large, interdisciplinary scientific hub within the Simons Foundation that is dedicated to the advancement of scientific research through computational methods, including data analysis, modeling and simulation. The institute addresses the need for new computational approaches to analyze the vast experimental and observational datasets being generated and collected in biology, chemistry and physics.

Flatiron, which began operations in 2016, provides a permanent home for professional scientists who are driven by the problem of developing, deploying and supporting state-of-the-art computational methods for modeling and analysis. The institute supports research centers devoted to computational astrophysics, computational biology, computational mathematics and computational quantum physics, each of which is led by a distinguished scientist.

Scientists at Flatiron have the opportunity to dedicate themselves to long-term research that is not driven by standard funding cycles. The institute currently serves the needs of 160 scientists, and has plans to grow to 250 scientists over time.

The Scientific Computing Core (SCC) is the technical backbone of the Flatiron Institute. The SCC develops, deploys and maintains computational infrastructure — from supercomputers to desktop PCs — dedicated solely to the use of Flatiron researchers. The SCC also creates and disseminates software tools for the wider computational science community, with the goal of ensuring that technology is not the limiting factor of scientific progress.

The SCC’s main computing resource is a high performance computing cluster housed in the institute’s

data center in New York City’s Flatiron District. This system is based on Dell EMC PowerEdge™ servers

that collectively have more than 10,000 Intel® Xeon® Scalable Processor cores, all connected via Dell H-Series networking based on Intel® Omni-Path Architecture (Intel® OPA) and managed with Bright Cluster Manager.

On the storage side, this steadily growing HPC system uses Dell DSS 7500 servers to hold thousands of disks and nearly 20 petabytes of storage, which is accessed via a Ceph file system. The mix of files in the system’s archive ranges from millions of tiny files to thousands of 100-gigabyte files.

In addition to the resources in the New York City data center, Flatiron gives researchers access to two off-site supercomputers. One of these resources is located at Brookhaven National Laboratory on Long Island, New York, and the other is at the San Diego Supercomputer Center at the University of California, San Diego. Collectively, the institute’s distributed infrastructure gives researchers access to 30,000 processing cores.

The computationally driven research conducted at the Flatiron Institute spans the gamut of astrophysics, biology, mathematics and quantum physics. Researchers use the Scientific Computing Core systems to analyze neuronal activity, predict the dynamic behavior of materials and molecules, analyze astronomical events and piece together the mysteries of the origins of life — to highlight just a few examples of the diverse research taking place every day at the Flatiron Institute.

A common thread through this diverse range of research is the need for HPC systems that can quickly process huge amounts of data to accelerate time to insight. And that’s what the Flatiron Institute is all about.

To learn more

For the full story, read the Dell EMC case study “Pure science: Flatiron Institute delivers the computational power to fuel leading-edge scientific research.” And to explore Dell EMC PowerEdge servers with Intel Xeon processors in HPC applications, visit Dell EMC solutions for HPC and AI.