HPC WORLD MLの皆様

高度情報科学技術研究機構 (RIST) の太田幸宏と申します。MLをお借
りして、オンラインセミナー「The 1st R-CCS/RIST Joint Seminar on
Advanced use of Supercomputer Fugaku and Arm computer systems」
(2025/4/23) の案内をさせていただきます。重複して受け取られた
場合はご容赦ください。

理化学研究所計算科学研究センター (RIKEN R-CCS）と高度情報科学
技術研究機構（RIST）は、共同でスーパーコンピュータ「富岳」および
Arm計算機システムの利用に関するセミナーを開催しています。
参加締め切りは2025/4/22です。皆様の参加をお待ちしています。

https://www.hpci-office.jp/en/events/seminars/seminar_FugakuAndArm_250423
========================================================================
# The 1st R-CCS/RIST Joint Seminar on Advanced use of Supercomputer Fugaku and Arm computer systems
[Date] 3:00 pm – 5:00 pm (JST), Wednesday, April 23rd, 2025

[Registration (Until April 22nd, 2025)]
https://www.hpci-office.jp/events/seminars/seminar_FugakuAndArm_250423

[Topics]
1. Using Singularity containers on Fugaku, basic usage and tips
Speakers: Motohiko Matsuda (RIKEN R-CCS)
Abstract: I would like to share some usage examples of
Singularity containers on Fugaku, which I already offered
in messages in Zendesk Community. Those examples supplement
the items covered in the tutorials held twice in 2023, and
they include: (1) installing MPI runtime in a container,
(2) installing Fujitsu compilers in a container, and (3)
a trick which might reduce loads in accessing files of Spack,
as well as other random topics. I hope this seminar will be
a chance to make communications active among users of Fugaku.

2. Executing a job on Fugaku with the best balance of computing performance and power consumption
Speakers: Kazunori Mikami (RIKEN R-CCS) and Fumichika Sueyasu (RIKEN R-CCS)
Abstract: We encourage Fugaku users to actively utilize
the power-saving feature, so called power knobs, of Fugaku
hardware for energy efficient operation. The advantage of
choosing the appropriate power-saving feature has been
discussed in various occasions. Some users, however, maybe
reluctant to adopt it mostly due to the risk of the performance
degradation, and we would like to address such concerns.
In this talk, we will provide an overview of Fugaku’s
power-saving feature, with the emphasis on the available power
modes that are set up in job management software. We will show
some measured results of the relationship between the performance
and the power consumption, as well as the effect of choosing
the data precision type, aiming to provide users with ideas
for achieving the most efficient job execution on Fugaku.
We will also introduce the Fugaku Points service, which provides
incentives to users who run power economical jobs on Fugaku.

[Contact]
hpc-support-admin[-at-]hpci-office.jp (Please replace [-at-] with @.)