Success Stories

Short History

NEC has been part of the HPC business for decades and is looking back upon some major installations in that field and is continuing to do so in the future.

As an example the famous Earth Simulator from the Japan Agency for Marine-Earth Science and Technology must be named. Based on the SX-6 architecture it has shown a LINPACK benchmark of 35.86 TFlops. This performance made it the fastest supercomputer in the world from June 2002 - November 2004 (based on the TOP500 list). 

In 2011 the TSUBAME 2.0 for the GSIC Center at the Tokio Institute of Technology saw another top 5 supercomputer in cooperation with HPE. In November 2010 the 4th rank was achieved with 1.192 PFLOPS. This performance could be extended to 2.843 PFlops with an update in November 2013.

Present and Future

TSUBAME 2.5 was still ranking at the top 100 (90th place) as of 11/2019 but shortly after fell out of the top 500. Compared to the current number one Supercomputer Fugakua (11/2021) it shows you how rapidly the overall average performance in the HPC sector has risen and will continue to do so in the future. The first supercomputer with over 1 TFlop (ASCI Red) was in 1997, the first over 1 PFlop (IBM Roadrunner) in 2008 and the first up to over 1 EFlop is planned for 2022 (planned for: 2022 Aurora with 2 EFlop for the Argonne Laboratory, 2022 Frontier 1.5 EFlops for the Oak Ridge Laboratory and 2023 El Capitan 2 EFlops for the Lawrence Livermore Laboratory).

There are many other successful NEC HPC installations like at the Japan Agency for Marine-Earth Science and Technology (Earth Simulator), National Institute for Fusion Science (NIFS Plasma Simulator), RWTH Aachen (CLAIX), University of Mainz (Mogon II), Institute for Molecular Science in Japan (Molecular Simulator), German Aerospace Center DLR (CARA) or Center for Computational Sciences at University of Tsukuba (Cygnus) the most recent installations should be shown in more detail on the following pages.

Research & Academic

Universities traditionally are at the frontline of supercomputing. They have always had a need for compute power for research and education.

According to the TOP500 list (as of November 2021) Research (108 systems representing 21.6%) and Academic (76 systems representing 15.2%) provide a total of 36.8% of all systems on the list. From the current list 8 out of 10 systems are from research or academic.

The first three EFlops supercomputer (as far as there is any information available) will also all be built for laboratories. So as it goes for supercomputing the most powerful machines are built for the academic and research arena.

However many of the industrial companies do not run or do not publish their LINPACK benchmark results so the list should not be seen as absolute. With the Selene system (#6 from NVIDIA) and Voyager-EUS2 (#10 from Microsoft Azure) beeing the first 2 in the top 10 for quite some time.