Cori (supercomputer)

Cori (supercomputer)
Name	Cori
Active	2016–2023
Location	National Energy Research Scientific Computing Center
Manufacturer	Cray
Purpose	Scientific research
Architecture	Cray XC40, Intel Xeon Phi
Power	3.9 MW
Speed	27.9 petaFLOPS (peak)
Cost	$70 million (approximate)

Cori (supercomputer). Cori was a Cray XC40 supercomputer operated by the National Energy Research Scientific Computing Center at Lawrence Berkeley National Laboratory. Named in honor of the biochemist Gerty Cori, it was a leading system for open scientific computing from its deployment in 2016 until its retirement in 2023. Cori was instrumental in advancing research across numerous fields, including climate science, high-energy physics, and materials science.

Overview

Cori represented a significant investment by the United States Department of Energy to provide cutting-edge computational resources for the national research community. The system was housed at the Lawrence Berkeley National Laboratory in California, a key facility managed by the University of California. As a flagship machine of the National Energy Research Scientific Computing Center, it supported thousands of researchers from institutions like the Massachusetts Institute of Technology, Stanford University, and the California Institute of Technology. Its primary mission was to enable simulations and data analysis for projects funded by the DOE Office of Science.

Hardware and architecture

The Cori system was a hybrid architecture built by Cray Inc., integrating two distinct node types within its Cray XC40 framework. One partition featured over 2,000 nodes powered by Intel Xeon "Haswell" processors for conventional workloads. The other, more innovative partition contained over 9,300 nodes, each equipped with a single Intel Xeon Phi "Knights Landing" many-core processor, which was a significant step in the Post-K computer era of many-core computing. The system utilized the high-performance Cray Aries interconnect and was supported by a massive Lustre (file system) parallel file system for data-intensive tasks. Its design emphasized energy efficiency, a critical consideration for facilities like the Texas Advanced Computing Center.

Software and applications

Cori's software environment was optimized for large-scale parallel applications common in Department of Energy research. It ran the Cray Linux Environment and supported programming models such as OpenMP, MPI, and OpenACC to harness its hybrid architecture. Key applications included the Community Earth System Model for climate prediction, Quantum ESPRESSO for materials modeling, and frameworks for analyzing data from the Large Hadron Collider at CERN. The system also hosted workflows for the Joint Genome Institute and supported codes from the INCITE program, facilitating research in astrophysics and combustion science.

Performance and rankings

At its peak, Cori achieved a theoretical performance of 27.9 petaFLOPS, with its Intel Xeon Phi partition delivering 14.0 petaFLOPS on the LINPACK benchmark. This performance earned it the #5 spot on the TOP500 list in November 2016, a notable achievement alongside contemporaries like the Sunway TaihuLight and the Titan (supercomputer). It also ranked highly on the HPCG benchmark, reflecting its balanced capability for real-world scientific applications. Throughout its operational life, it remained a top-tier resource in global rankings, competing with systems at Oak Ridge National Laboratory and the Argonne National Laboratory.

History and funding

The procurement and deployment of Cori were funded through the Department of Energy's CORAL initiative, a collaboration between Oak Ridge National Laboratory, Argonne National Laboratory, and Lawrence Livermore National Laboratory. The system was installed in phases, with the Intel Xeon partition becoming operational in 2015 and the Intel Xeon Phi partition added in 2016. It was officially named after Gerty Cori, the first American woman to win a Nobel Prize in Physiology or Medicine. After seven years of service, Cori was retired in 2023, making way for next-generation systems like Perlmutter (supercomputer) at the same facility.

Impact and scientific achievements

Cori enabled groundbreaking scientific discoveries across a vast portfolio of research. It was crucial for high-fidelity simulations of turbulence and combustion, advancing the design of more efficient engines. Researchers used it to analyze petabytes of data from the Dark Energy Spectroscopic Instrument to map the universe's expansion. In bioenergy, it facilitated genomic studies for the development of sustainable biofuels. Projects supported by the Advanced Scientific Computing Research program leveraged Cori to model complex phenomena in plasma physics and quantum chromodynamics, solidifying its legacy as a pivotal tool for 21st-century science.

Category:Supercomputers Category:Cray computers Category:Lawrence Berkeley National Laboratory