In This Message

Center Announcements
  – ALCC Call for Proposals

Meetings & Workshops
  – OLCF User Conference Call is Today (Dec 8)
  – Perlmutter Training (Jan 5-7)
  – NVIDIA HPC SDK Training (Jan 12-13)
  – Vampir/Score-P Training (Jan 19-20)

Upcoming Scheduled Outages
  – Andes (Dec 1)
  – Summit, Andes, Spock, Data Transfer Nodes, Alpine (Dec 14)

Center Announcements

ALCC Call for Proposals
The DOE Office of Advanced Scientific Computing Research (ASCR) is now accepting application proposals for the 2022-23 allocation year of the ASCR Leadership Computing Challenge (ALCC) program. The program provides access to computational resources at ASCR’s three supercomputing facilities (OLCF, ALCF, and NERSC).

The proposal process consists of two steps, with a pre-proposal due on December 17, 2021 and the full proposal due on February 22, 2022. The allocation period for successful proposals will run from July 1, 2022-June 30, 2023.

For more information, see the announcement at https://ess.science.energy.gov/ascr-leadership-computing-challenge-alcc-call-open-for-2022-2023-allocation-year/.

Meetings & Workshops

OLCF User Conference Call is Today (Dec 8)
The final OLCF User Conference Call for the year will be held at noon (Eastern Time) today, December 8. The call will include a discussion on analysis tools at OLCF led by Benjamin Hernandez. For more information on that call, see https://www.olcf.ornl.gov/calendar/userconcall-dec2021/.

Perlmutter Training (Jan 5-7)
NERSC will host Perlmutter training on Jan 5 -7 provided jointly by HPE and NERSC staff. This is a continuation and extension of the June 2021 Perlmutter Introduction training, now focused more on using Perlmutter with hands-on exercises. ALCF and OLCF users are invited to this training, and NERSC training accounts will be provided based on availability and time of registration. Registration is available via this link.

NVIDIA HPC SDK Training (Jan 12-13)
On January 12-13, NVIDIA will present a 2-day seminar for NERSC, ALCF, and OLCF users about the various GPU programming models supported by NVIDIA’s HPC SDK compilers, including Standard Language Acceleration and Libraries, OpenACC, OpenMP offload, and CUDA. The basic GPU architecture and HPC SW developer considerations and the profiling tools will also be presented. Attendees will have the opportunity to do hands-on and homework exercises on Perlmutter Phase 1 GPUs (NERSC and ALCF users) and on Summit (OLCF users). The agenda and registration form are available via this link.

Vampir/Score-P Training (Jan 19-20)
OLCF will host Vampir and Score-P training on January 19-20 from 9:00 am-11:30 EST each day. Vampir is an established performance tool for visualizing and analyzing the behavior of highly parallel HPC applications. It is supported by the Score-P tool, which can be used to fully record the runtime behavior of HPC applications. The first day will feature lessons and hands-on demos using these tools. The second day will give users an opportunity to bring their own codes and instrument them for performance analysis with help from Score-P and Vampir developers.

For the detailed agenda and registration please see https://www.olcf.ornl.gov/calendar/vampir-score-p-training/.

Upcoming Scheduled Outages

  • The OLCF JupyterHub instance will be unavailable form 8:00 AM until noon on Monday, December 13.
  • Summit, Andes, Spock, the Data Transfer Nodes, and the Alpine filesystem will be unavailable from 8:00 AM until 11:59 PM on Tuesday, December 14.