Pleiades (supercomputer)

Pleiades (supercomputer). Pleiades is a petascale supercomputer housed at the NASA Advanced Supercomputing (NAS) Facility located at Ames Research Center in Silicon Valley. Operated by the NASA Advanced Supercomputing Division, it is one of the world's most powerful supercomputers dedicated to supporting the computational needs of the U.S. space agency's scientists and engineers. The system is instrumental for simulations in aerospace engineering, earth science, and astrophysics, enabling critical research for missions across NASA.

Overview

Pleiades serves as the primary high-performance computing resource for the NASA Advanced Supercomputing Division, providing immense computational capacity to thousands of researchers across the agency's nationwide centers, including JPL, GSFC, and LaRC. The system is integrated into the larger High-End Computing Capability (HECC) project, which aims to deliver a robust and scalable computing environment. Its operations are critical for projects with stringent deadlines, such as supporting Space Shuttle missions in the past and current endeavors like the Artemis program and the JWST.

Hardware and architecture

The supercomputer is a heterogeneous system, historically built by SGI utilizing Altix ICE and UV architectures, with later expansions and upgrades led by Hewlett Packard Enterprise using HPE Apollo and HPE SGI 8600 systems. It employs a multi-core processor design featuring tens of thousands of Intel Xeon and AMD Opteron CPUs, organized within a high-speed InfiniBand interconnect fabric for efficient parallel processing. The total system incorporates over 200,000 processor cores and is coupled with a massive multi-petabyte Lustre parallel file system for high-throughput data storage and access, managed by the NASA Advanced Supercomputing Division team.

Performance and capabilities

Pleiades has consistently ranked highly on the TOP500 list of the world's most powerful supercomputers, achieving a peak performance of over 7 petaflops following its major upgrades. The system's architecture is designed for both capability computing, running massive, tightly-coupled simulations, and capacity computing, handling numerous smaller jobs concurrently for the broad NASA user community. Its computational throughput is essential for complex modeling tasks, such as global climate simulations for the GISS ModelE and high-fidelity computational fluid dynamics for the SLS rocket.

Applications and research

The supercomputer supports a vast portfolio of scientific and engineering applications central to NASA's mission. In aerospace engineering, it is used for designing next-generation aircraft concepts like those from the Advanced Air Vehicles Program and simulating entry, descent, and landing for MSL and Perseverance rover missions. Astrophysics research includes modeling neutron star mergers and galaxy formation, while earth science applications run high-resolution models from the GEOS-5 system to study hurricanes and climate change. It also plays a key role in spacecraft design and mission planning for the Europa Clipper.

History and development

Pleiades was first deployed in 2008 as part of a sustained acquisition strategy to modernize NASA's computing infrastructure, with an initial performance of 487 teraflops. Major upgrades occurred in phases, often through collaborations with SGI and later Hewlett Packard Enterprise following its acquisition of SGI's assets, significantly boosting its performance to petascale levels. Its development has been closely tied to the evolving needs of projects like the Constellation program and the ISS. The system is named after the Pleiades star cluster, continuing NASA's tradition of naming its supercomputers after astronomical objects, following predecessors like Columbia.