University of Tokyo Orders Fujitsu HPC

Fujitsu Limited has received an order for its new supercomputer from
the Supercomputing Division, Information Technology Center, The University of
Tokyo (SCD/ITC).

Fujitsu
supercomputer PRIMEHPC FX10 will be used for the University of Tokyo’s
supercomputer system, which will have a theoretical computational speed of 1.13
petaflops, making the system one of the few petaflops-class center systems
in the world.

The new
supercomputer system will start operations in April 2012. It will be used to
support SCD/ITC’s advanced research and educational activities in a variety of
fields of science and technology, and is expected to be used by corporations as
well.

Background
to the Deployment of the New Supercomputer
SCD/ITC was
established in 1965 as a large-scale computer center for shared use by
researchers in Japan, and has been operating under its current name since 1999.
The center has been contributing to research using supercomputers in
science and technology, and is used by more than 1,500 researchers inside and
outside of the university. To meet the needs of an increasing number of diverse
users and to handle large-scale applications, SCD/ITC selected PRIMEHPC FX10
for its computing performance, high data input/output speeds, and
data processing capabilities, as well as its compatibility
with the K computer.

The  is expected to be put to use in a variety of fields of science and
technology (including earth science, astrophysics, seismology, environmental
modeling, materials science, energy, biology, fluid dynamics, and solid
mechanics).

Overview of the New Supercomputer System:
The
calculation nodes of the  supercomputer system will use a PRIMEHPC FX10
configuration comprised of 4,800 nodes. Combining performance,
scalability and reliability with energy efficiency, PRIMEHPC FX10
further enhances Fujitsu’s technology used in the K computer, which achieved
top-ranked performance capability.

For the
supercomputer’s peripheral systems, 74 PRIMERGY servers, 234 ETERNUS storage
systems, Technical Computing Suite HPC middleware for peta-scale systems,
and FEFS, a high performance, highly-reliable distributed file system which
is a component of the Technical Computing Suite, will be deployed.

Specifications of the Supercomputer
System:

Compute Nodes: PRIMEHPC FX10

• Number of racks: 50
• Nodes: 4,800
• Theoretical Computing Speed: 1.13
  petaflops
• Memory Capacity: 150TB
• Interconnects: 6D Mesh/Torus Tofu
  Interconnects

Log-in Nodes Management Servers:

• 16 Primergy
  RX200S6 servers 58 Primergy RX300S6 servers

Local Files:

• 150 Eternus DX80S2 storage
  units, 1.1 petabytes: (RAID5)

Shared Files:

• 80 Eternus DX410S2
  storage units, 2.1 petabytes: (RAID6)

File System:

• FEFS

Kengo Nakajima, Ph.D., Director, Supercomputing Division, Information
Technology Center,
The University of Tokyo, commented: “Many
of our center’s users are Japan’s
top researchers, and they are also K computer users. By deploying PRIMEHPC
FX10, which is highly compatible with the K computer, we hope to contribute to
the further development of Japan’s computational science capabilities. We also
plan to use the new system in the HPC educational programs we have established
in our graduate school to foster the development of engineers who will lead
future advances in computational science. From the standpoint of the center’s
operation, the supercomputer’s superior performance per watt of 2GFLOPS/W is
especially important in light of the current electricity supply situation.”