Bridges-2

Network outage planned for December 19th - 20th

December 19-20: Beginning on Tuesday, December 19 at 6AM Eastern time, the entire PSC machine room (all machines, VMs and filesystems) will be unreachable due to a major networking upgrade. We anticipate that this outage will last until 6PM Eastern time on Wednesday December 20.

Jobs running on individual systems should continue to run so long as they are not relying upon any network connections (license servers, input files from remote systems, etc.).

Thank you for your patience while we improve the network for the current and future needs of the PSC community.

Bridges-2, PSC’s flagship supercomputer, began production operations in March 2021. It is funded by a $10-million grant from the National Science Foundation.

Bridges-2 provides transformative capability for rapidly evolving computation- and data-intensive research, and creates opportunities for collaboration and convergence research. It supports both traditional and non-traditional research communities and applications. Bridges-2 integrates new technologies for converged, scalable HPC, machine learning and data; prioritizes researcher productivity and ease of use; and provides an extensible architecture for interoperation with complementary data-intensive projects, campus resources, and clouds.

Bridges-2 is available at no cost for research and education, and at cost-recovery rates for other purposes.

Core concepts and innovation

Bridges-2 was designed around the core concepts of converged HPC + AI + Data; a custom topology optimized for data-centric HPC, AI, and HPDA; heterogeneous node types for different aspects of workflows, including both CPUs and AI-targeted GPUS; 3 tiers of per-node RAM: 256GB, 512GB, and 4TB; an extremely flexible software environment; and community data collections and Big Data as a Service.

Innovative features built into Bridges-2 to facilitate these core concepts include:

  • AMD EPYC 7742 CPUs: 64-core2.25–3.4 GHz
  • AI scaling to 192 V100-32GB SXM2 GPUs
  • 100TB, 9M IOPs flash array accelerates deep learning training, genomics, and other applications
  • Mellanox HDR-200 InfiniBand doubles bandwidth & supports in-network MPI-Direct, RDMA, GPUDirect, SR-IOV, and data encryption
  • Cray ClusterStor E1000 Storage System
  • HPE DMF single namespace for data security and expandable archiving

Questions?

If you have questions about Bridges-2, please email help@psc.edu.

Bridges-2 Powers Snake-Virus Genome Analysis

New DNA Sequence Provides Clues to Virus’s Spread, Possible Targets for Future Therapies

Similar Genetic Elements Underlie Vocal Learning in Bats, Whales, and Seals

Non-Gene Regulatory DNA Identified via Artificial Intelligence also Associated with Autism in Humans

Human-AI Knowledge Sharing for Water Treatment Plants Operation

CMU Group Uses Bridges-2 to Create Virtual Plant to Learn from Veteran Engineers, and Next Train and Work with Rookies

OpenSoundscape Uses AI, Other Tools to Identify Wildlife Species from Audio Recordings in the Wild

Pitt Team Uses Bridges-2 to Build Automated, Open-Source Toolbox, Enabling Conservation Scientists to Survey Much Larger Regions with Limited Staffing

Neutron-Star Smash Up Sims Offer Window on Atomic Structure

Bridges-2 Simulations Show Specific Heat Effect that, with New Telescopes, Will Give Clues on Condensed Matter, Atomic Nuclei

CMU Group Outs Printer of Embarrassing Typographic Error

Infamous Typo in Shakespeare Folio Sheds Light on His Growing Status, Reduction in Status of Printing Profession in 17th-Century England

Node types

Regular Memory nodes

Regular Memory (RM) nodes provide extremely powerful, general-purpose computing, machine learning and data analytics, AI inferencing, and pre- and post-processing.

488 RM nodes have 256GB of RAM, and 16 have 512GB of RAM.

Details
All 504 RM nodes have:

  • NVMe SSD (3.84TB)
  • Mellanox ConnectX-6 HDR Infiniband 200Gb/s Adapter
  • Two AMD EPYC 7742 CPUS, each with:
    • 64 cores
    • 2.25-3.40GHz
    • 256MB L3
    • 8 memory channels

Extreme Memory nodes

Extreme Memory (EM) nodes provide 4TB of shared memory for genome sequence assembly, graph analytics, statistics, and other applications requiring a large amount of memory for which distributed-memory implementations are not available.

Details

Each of Bridges-2’s 4 EM nodes consists of:

  • 35.75MB LLC 6 memory channels
  • 4TB of RAM: DDR4-2933
  • NVMe SSD (7.68TB)
  • Mellanox ConnectX-6 HDR InfiniBand 200Gb/s Adapter
  • Four Intel Xeon Platinum 8260M “Cascade Lake” CPUs:
    • 24 cores
    • 2.40–3.90GHz

GPU nodes

GPU nodes provide exceptional performance and scalability for deep learning and accelerated computing. Bridges-2 launched with 24 GPU nodes, totaling 40,960 CUDA cores and 5,120 tensor cores. The GPU nodes from Bridges were later migrated to Bridges-2, adding the DGX-2 and nine more V100 GPU nodes to Bridges-2’s GPU  resources.

Details

The 24 original Bridges-2 GPU nodes contain:

  • 512GB of RAM: DDR4-2933
  • 7.68TB NVMe SSD
  • Two Mellanox ConnectX-6 HDR Infiniband 200Gb/s Adapter
  • Eight NVIDIA Tesla V100-32GB SXM2 GPUs
  • 1 Pf/s tensor
  • Two Intel Xeon Gold 6248 “Cascade Lake” CPUs:
    • 20 cores, 2.50–3.90GHz, 27.5MB LLC, 6 memory channels

The Bridges GPUs migrated to Bridges-2 include nine nodes with:

  •  192GB of RAM, DDR4-2666
  • 4 NVMe SSDs, 2TB each
  • 8 NVIDIA V10016GB GPUs
  • 2 Intel Xeon Gold 6148 CPUs:
    • 20 cores/CPU, 2.4 – 3.7 Ghz

The DGX-2 has:

  • 1.5TB RAM, DDR4-2666
  • 8 NVMe SSDs, 8.84TB each
  • 16 NVIDIA Volta V100-32GB GPUs
  • 2 Intel Xeon Platinum 8168 CPUs:
    • 24 cores/CPU, 2.7 – 3.7 GHz