Corporate use briefing session
(1st session for FY2021)
Objectives of the briefing session
The Information Technology Center, The University of Tokyo offers a system that allows corporations to make use of its supercomputers (hereafter, “corporate use”) with the purpose of encouraging the spread of large-scale computing and contributing to society.
The first call for applications for FY2021 (Deadline: February 17) will commence soon.
This call for applications offers the following systems for corporate use.
- Integrated Supercomputer System for Simulation, Data and Learning
Wisteria/BDEC-01 Supercomputer System
- Large-scale Parallel Supercomputer System
Oakbridge-CX Supercomputer System
- Manycore-based Large-scale Supercomputer System
Oakforest-PACS Supercomputer System [ * March 31, 2022 Ended ]
- Integrated Supercomputer System for Data Analyses & Scientific Simulations
Reedbush H/L Supercomputer System [ * November 30, 2021 Ended ]
Wisteria/BDEC-01 is a new system scheduled to start operation on May 14, 2021, and will be available for corporate use on October 1.This system has two groups of computational nodes: a group of simulation nodes (Odyssey (Wisteria-O)) and a group of data and learning nodes (Aquarius (Wisteria-A)).Wisteria-O (Odyssey) is equipped with 7,680 FUJITSU Processor A64FX processors with a peak performance of 25.9 PFLOPS, and Wisteria-A (Aquarius) is equipped with 360 NVIDIA A100 Tensor cores, the latest GPU, with a peak performance of 7.2 PFLOPS. Wisteria-A (Aquarius) is equipped with 360 NVIDIA A100 Tensor cores for a peak performance of 7.2 PFLOPS.
The Oakbridge-CX system is a supercomputer system equipped with 1326 compute nodes consisting of Intel Xeon processors and Omni-Path architecture. The SSD is mounted on 128 nodes out of 1,368 nodes, and can handle processing that requires high input / output performance.
The Oakforest-PACS system is the latest manycore-based CPU, Intel Xeon Phi 7250 (Knights Landing, KNL) is equipped with 8,208 nodes, and total peak computational performance can reach approximately 25PFLOPS.
The Reedbush system is equipped with high-performance GPU, NVIDIA Tesla P100 (“Pascal”) in 184 out of a total of 604 nodes, making it compatible with the demands of new fields such as big data analysis and machine learning.
Under the corporate use system, in addition to the regular group use, various other options of use are also available, including trial use course (paying use/free of charge) and joint research. The cost of use is set at a reasonable rate based on electricity consumption. For the Reedbush system and Oakbridge-CX, a “fixed node” service is also provided, enabling users to customize nodes. Wisteria/BDEC-01 will also be introduced.
In addition, I introduce commercial softwareware HyperWorks into Reedbush system (in detail) and Oakbridge-CX system (in detail).You can use HyperWorks in a company by purchasing the special license for the Univ. of Tokyo super computer.Wisteria/BDEC-01 will also be introduced.
We will be holding a briefing session for corporate users as follows. After the briefing, individual consultations will be available.
Registration for briefing session
Registration formPlease submit the registration form at the latest by the day before the briefing session.
Registration to participate has closed.
Date and time of event
January 25, 2021 (Monday), 14:00 – 15:00
Online briefing using zoom.
Program for the Corporate Use Briefing Session
January 25 (Monday)
|14:00 - 14:15||Overview of the Center’s activities
|14:15 - 15:45||Overview of the corporate use service|
|14:45 - 15:00||Question and Answer|
|15:00 -||Individual consultation *To be postponed to a later date if there are too many applicants|
Research Support Team, Information Strategy Group, Information Systems Department,
The University of Tokyo