In This Message

  • Upcoming Downtimes
    • Center-wide (April 16)
  • Center Announcements
    • Orion LFS setstripe wrapper
    • HPSS Decommissioning and Kronos Arrival
  • Meetings & Workshops ​ ​
    • AI Training Series: Enhancing PyTorch Performance on Frontier with the RCCL/OFI-plugin (Apr 17)
    • Performance Portability Training: Kokkos Training (April 25-26)
    • OpenMP Training Series: OpenMP Introduction (May 6)

Upcoming Downtimes

– Frontier, Orion, Andes, DTNs, FrontierSPI, Jupyter, and HPSS will be unavailable from 8:00 AM until 8:00 PM on Tuesday, April 16.

Center Announcements

Orion LFS setstripe wrapper

When manually setting the stripe for the OLCF center-wide Lustre scratch filesystem, Orion, it is important to keep in mind the filesystem’s layout. For example, striping across the resource’s capacity OST tier can negatively impact performance. To help ease use of the Lustre lfs setstripe tool, the OLCF provides a lfs wrapper. The wrapper can be loaded into your environment through the lfs-wrapper modulefile. The module is available for testing now and will be added to the default environment for all users on April 16 if no issues are reported. More information on the lfs wrapper can be found on: https://docs.olcf.ornl.gov/data/index.html#lfs-setstripe-wrapper

HPSS Decommissioning and Kronos Arrival

After decades in service and having served hundreds of users that have archived over 160 petabytes, the OLCF’s HPSS system is reaching end of its life and will be decommissioned early in 2025. In preparation for that process, we would like to share upcoming system changes and introduce the new nearline storage system, Kronos.  The HPSS system will become read-only once the Kronos system is in production and available to users. We are targeting early June 2024 for this change to take place. More information about using Kronos and recommendations on migrating your data from HPSS can be found at: https://docs.olcf.ornl.gov/systems/2024_olcf_system_changes.html.

Meetings & Workshops ​

AI Training Series: Enhancing PyTorch Performance on Frontier with the RCCL/OFI-plugin
April 17, 1:00 pm – 2:00 pm Eastern Time

In this seminar, an overview of using PyTorch on Frontier with the aws-ofi-rccl plugin will be provided, along with specific profiling examples run on Frontier. This seminar is intended for OLCF users that have an allocation on Frontier, but all are welcome to join and view the presentation. For more information or to register, see https://www.olcf.ornl.gov/calendar/pytorch-on-frontier/

Performance Portability Training: Kokkos Training
April 25-26, 12:00 p.m. EST

The Kokkos training session presented by the Kokkos team, is part of the Performance Portability training series. This series features training sessions on various performance portable programming solutions to help ease developer transitions between current and emerging high-performance computing (HPC) system.  For more information and registration: https://www.olcf.ornl.gov/calendar/kokkos-training-2024/

OpenMP Training Series: OpenMP Introduction
May 6th, 2024, 12:00 p.m. EST

The OpenMP training series presented by Michael Klemm of AMD the OpenMP ARB, and Christian Terboven of RWTH Aachen University, is part of the Performance Portability training series. The series features training sessions on various performance portable programming solutions to help ease developer transitions between current and emerging HPC systems.  For more information and registration: https://www.olcf.ornl.gov/calendar/openmp-training-series-openmp-introduction/