Innovative Research Using the K Computer and Framework to Advance Computational Science and Technology
■ HPCI Plan and Strategic Program Field 5
The innovative High Performance Computing Infrastructure (HPCI) plan, which was established with the “K computer” at the core, was created to actualize an innovative computational environment capable of meeting diverse users needs, and aimed at promoting the exploi- tation of HPCI. As major policy objectives, building a framework to advance computational science and technology in major fields (HPCI Strategic Program), and producing revolutionary results through the exploitation of HPCI are cited.
In the HPCI Strategic Program, five potential fields have been identified as strategic fields that not only require K computer resources, but can also be expected to yield significant social and scholarly breakthroughs. “Field 5: The origin of matter and the Universe” is one of these. This field deals with fundamental physics, primarily concerning elementary particles, the atomic nuclei, and astrophysics. The insti- tution (strategic institution) that leads Field 5 research and development is the Joint Institute for Computational Fundamental Science (JICFuS), which is a joint research organization combining the University of Tsukuba’ s Center for Computational Sciences (CCS), the High Energy Accelerator Research Organization(KEK), and the National Astronomical Observatory of Japan(NAOJ).
■ Strategic Objective of Field 5
The history of the Universe is believed to have begun with the ultra high-temperature, ultra high-density state called the Big Bang, which occurred 13.7 billion years ago. Baryons such as protons and neutrons were made as the bound states of quarks and gluons with the tem- perature drop following the event. Subsequently, protons and neutrons bonded, and light atomic nuclei were generated.
On the other hand, it is also postulated that unidentified dark matter exists in much larger quantities than baryons in the Universe. Furthermore, it is currently thought that dark matter coalesced under the influence of gravity and created a structure, to which normal baryonic matter was drawn, eventually forming the galaxies and stars of the Universe, which then evolved into its currently observed form. In those galaxies, stars eventually perish due to gravitational collapse and supernova explosion, while new stars are also actively being born. Through this process, heavier atomic nuclei are generated. The history of the generation of matter has a close relationship to the history of the structural formation of the Universe.
The strategic objective of Field 5 is to understand, in a unified fashion, the context of the history of the Universe beginning with the Big Bang. This includes the origin and structure of matter and the Universe from elementary particles to nuclear synthesis, to stars and galaxy formation, by scientific computational methods that connect the multiple hierarchical steps. Four research and development subjects have been set toward the objective.
(1)Determination of the baryon-baryon interactions using lattice QCD at the physical point
(2)Elucidation of nuclear properties using ultra large-scale simulations of quantum many-body systems and its applications
(3)Clarification of the processes underlying supernova explosions and the formation of black holes
(4)Investigation on the formation of first-generation stars in the Universe out of density fluctuations of dark matter
Research of these subjects is carried out using the K computer, which has a peak performance of 10.5 Pflops (calculations enabled at about 10 quadrillion times per second). Results have been yielded steadily. The dark matter simulation performed on the K computer won the Gordon Bell Prize, the most prestigious of its type in the field of high performance computing, in November 2012.
In parallel with this research, the efficient use of computational resources in the field of computational fundamental science, personnel development, creation of a research network, and fostering cooperation across different fields have been pursued with the aim building a framework to advance computational science and technology within the field. As the core of these strategic institutions, the project office promotes the accomplishment of research and development subjects and structuring of the framework to advance computational science and technology.