日時: 2019年7月18日（木）、14:00 - 14:45
場所: R-CCS 6階講堂
・講演題目： Introducing Arm and Arm Research: enabling innovation across all compute continuum.
・講演者： Filippo Spiga, (SLSS, Arm Research (UK))
Arm vision is develop technology that invisibly enables opportunities for a globally connected population. Nowadays Arm IP is at the foundation of billions of devices across all compute continuum (from Cloud to Edge to mobile and IoT). The aim of this talk is to provide an overview of Arm, in particular how we operate and how we pursue innovation via our Research group in partnership with academic and industrial institutions worldwide. The talk will also highlight some the latest technology announcements and some key Arm Research activities in various areas, including High Performance Computing.
日時: 2019年7月1日（月）、13:00 - 13:55
場所: R-CCS 6階講堂
・講演題目： Molecular Mechanisms for Protein-Ligand Binding in a Living Cell
・講演者： 杉田 有治（粒子系生物物理研究チーム、チームリーダー）
The inside of cell is highly crowded with proteins, nucleic acids, ribosomes, metabolites, ions and water. In the macromolecular crowding environments, protein behaviors can be altered compared to those in dilute solution. The effect of macromolecular crowding was mainly explained via the excluded volume effect. However, recent in-cell NMR spectroscopy and atomistic molecular dynamics (MD) simulations in explicit solvent (R. Harada eta al. JACS (2012, 2013)) have shown the importance of weak protein-protein interactions on protein stability and dynamics. In the talk, we discuss the effect of macromolecular crowding on protein-ligand interactions. In the simulations of the all-atom model of Mycoplasma Genitalium (I. Yu et al. eLife (2016)), we observed that not only hydrophobic but also hydrophilic metabolites also stay on the surfaces of proteins longer than in the bulk solution. Non-specific and weak protein-metabolite interaction is likely important for the metabolite distributions. We also investigated kinase-inhibitor binding processes in dilute solution and crowded protein solution by all-atom MD simulations.
日時: 2019年7月1日（月）、13:55 - 14:50
場所: R-CCS 6階講堂
・講演題目： High-performance numerical library and numerical reproducibility
・講演者： 今村 俊幸（大規模並列数値計算技術研究チーム、チームリーダー）
As it is the second round on the PI progress report on R-CCS Café, two research topics are selected from the recent outstanding results and the future project.
First, the most outstanding result of my team activities was that Dr. Kudo won the best paper award in HPC Asia 2019, Guangzhou, China. The main contribution of his paper was dedicated to accelerating a very compact and scalable eigenvalue solver on several types of manycore processors. His main idea was to reconstruct carefully but boldly any part of the implementation by introducing a systematic code generator to achieve performance portability and future extensibility. What is more, another idea to incorporate the “BLAS+X” approach improved the TRD algorithm (Householder tridiagonalization) and extended the functionality of TRD beyond the batch operations.
The second topic is that new research pillar of “numerical reproducibility”. This is a new concept and approach to guarantee numerical precision on any software and hardware configurations. This project is based on several mathematical theories and hardware/software/algorithmic supports of higher precision arithmetic. In this project, we intend to guarantee input/output numerical reproducibility. Internally, we secure the numerical accuracy fully 64bit or necessary precision (bit fields) by a stochastic approach with CADNA and PROMISE developed by the LIPS6 group, France. Naturally, as the rounding error contaminates any floating calculations, we must extend the internal data format with a much wider one, for example, IEEE754 real128 or other individual floating point formats. We are going to overcome this difficulty by incorporating the arithmetic engine of the FPGA. Furthermore, we need to investigate software emulation of a wider precision floating point format and higher precision algorithms.
I would like to introduce these topics at the next R-CCS Café briefly.
日時: 2019年7月1日（月）、15:05 - 16:00
場所: R-CCS 6階講堂
・講演題目： Development of a Unified Continuum Mechanics Simulation Framework for Industrial Applications
・講演者： 坪倉 誠（複雑現象統一的解法研究チーム、チームリーダー）
A High-Performance Computing(HPC) framework for the fluid-structure interaction problems with complicated geometry have been developed, considering its application to industrial matters. To overcome the difficulty of mesh generation and treatment of moving boundary, hierarchically structured finite volume method was adopted as its data structure in which both the fluid motion and structure deformation are solved in unified Eulerian manner. To achieve higher computational efficiency of parallelization and scaling on the massively parallel environment, Building Cube Method (BCM) proposed by Nakahashi was adopted. In the method, numerical domain is first decomposed into cubic sub-domains based on the octree method. Then the same number of numerical grids is allocated to each cubic subdomain. The solid surface with complicated geometry is represented by the immersed boundary method (IBM). In the fluid-structure interaction problems with structure surface in motion, accurate representation of the immersed body is indispensable. Thus Lagrangian description for tracking the moving solid body surface is adopted in the Eulerian framework of solving fluid and structure motions. So far, the framework can handle maximum of tens of billions of numerical meshes using hundreds of thousands of CPU cores on the K-computer. Applications of the method to vehicle aerodynamics, aero-acoustics, combustion systems, and structure analysis are introduced.
Keywords: unified simulation, HPC, immersed boundary, hierarchically structured grid, industrial application
日時: 2019年6月24日（月）、15:30 - 16:10
場所: R-CCS 6階講堂
・講演題目： Examination of efficient calculation method of relativistic molecular orbital method
・講演者： 井上 頌基（量子系分子科学研究チーム、特別研究員）
Molecular orbital method is a powerful tool to theoretically clarify the properties of molecular systems. The calculation accuracy of the molecular orbital method depends on the number of basis functions expanding the molecular orbital, which is a one-particle wave function of electrons, the treating of electron correlation, and the treating of relativistic effects. Relativistic effects are important mainly for heavy and superheavy elements, sometimes at a qualitative level.
However, relativistic molecular orbital calculation has large computational cost than non-relativistic one. This is a obstacle to clarifying the chemistry of heavy element systems theoretically. The cost of relativistic calculations comes from the facts that the Fock matrix is larger than that of non-relativistic ones, and that integrals of high angular momentum must be calculated.
So, in this talk, I will introduce the research that has been performed as an approach to these problems.
日時: 2019年6月24日（月）、16:10 - 16:50
場所: R-CCS 6階講堂
・講演題目： An Introduction to Chemical Evolution of Galaxies
・講演者： 平居 悠（粒子系シミュレータ研究チーム、基礎科学特別研究員）
Galactic chemical evolution studies the enrichment histories of elements in the Universe. At the time of the Big Bang, the Universe consists of hydrogen, helium, and lithium. However, we are now living in the Universe with various kinds of elements. Most of the elements are synthesized in stars and distributed to space when a star ends its life.
These elements are inherited to the next generation stars. Elemental abundances of stars, therefore, reflect the nucleosynthetic histories in the galaxy. In the first part of this talk, I will show how the observed elemental abundances of stars in the Milky Way and satellite dwarf galaxies preserve information about their enrichment histories. In the Milky Way, the distribution of elements is different among old and new stars. The second part of this presentation focuses on the modeling of galaxy evolution. Hydrodynamic simulations of galaxies are presently a powerful tool to study galactic chemical evolution. It can self-consistently follow the formation of galaxies and the enrichment histories of elements. The last part discusses the enrichment of elements heavier than iron. The origin of elements synthesized by a rapid neutron-capture process such as europium, platinum, and gold is a long-standing problem in astronomy. Recently, binary neutron star mergers have been detected by the gravitational wave observations. I will show that they can be the source of such elements in galaxies. I will also present the future prospects of the studies of galactic chemical evolution.
日時: 2019年6月11日（火）、14:05 - 14:40
場所: R-CCS 6階講堂
・講演題目： A brief history of NEST on supercomputers
・講演者： Susanne Kunkel (PostDoc, The Norwegian University of Life Sciences)
NEST is a simulator for large-scale spiking neuronal networks with a history dating back to the late 1990s. The definition of what is considered a large-scale network has changed since as ever more powerful HPC facilities have become available. Contemporary supercomputers even provide the resources to represent brain-scale spiking neuronal networks. However, in order to enable neuronal simulators to exploit such resources, simulation code had to undergo fundamental design changes. Today, NEST works efficiently for a broad variety of models and on various platforms, from laptops to supercomputers. In my talk I will give an introduction to the NEST simulator and review the technological advances of the last years that have made NEST such an extremely scalable simulation tool.
日時: 2019年6月11日（火）、14:40 - 15:05
場所: R-CCS 6階講堂
・講演題目： Meeting the performance challenges of spiking network simulations on general purpose computers
・講演者： Jari Pronold (PhD student, The Jülich Research Centre, Germany)
Today’s extremely scalable simulation technology for spiking neuronal networks enables the representation of models of more than a billion of neurons and their connections using the entire K computer. However, the runtimes of the largest possible simulations carried out so far were too long to allow for observations of the network dynamics over long periods of time, and also small to medium-scale simulations typically run in far more than real-time. The performance challenges for spiking neuronal network simulators such as NEST on general purpose computers arise from the inherent sparse but broad connectivity between neurons and from the unpredictable neuronal spiking activity. In distributed simulations of spiking networks, this requires frequent communication of spike data, and on each compute node routing of the incoming spikes to the local targets. This entails irregular memory access and hence constitutes a major performance bottleneck, which is a problem that I will address in my talk. I will present recent developments in simulation technology
that aim at meeting such performance challenges.
日時: 2019年6月7日（金）、13:00 - 13:55
場所: R-CCS 6階講堂
・講演題目： Working on the stable operation of the K computer
・講演者： 宇野 篤也（運用技術部門 システム運転技術ユニット ユニットリーダー）
The system operation and development unit has operated the K computer for more than 7 years. One of our important missions is operating the K computer stably. During the operation of the K computer, we have encountered some problems related to the file system, the job scheduling, the power consumption or etc. In this talk, I introduce our approaches to these problems for the stable operation of the K computer.
日時: 2019年6月7日（金）、13:55 - 14:50
場所: R-CCS 6階講堂
・講演題目： Systemization of performance optimization technique
・講演者： 南 一生（運用技術部門 チューニング技術ユニット ユニットリーダー）
Modern supercomputers are highly parallel machine combining inter-nodes process parallelism and inter-core thread parallelism. And the memory hierarchy including the cache in the node is also complicated. On the other hand, applications that run on supercomputers cannot fully utilize the performance of hardware unless high parallelization and individual node tuning are performed according to that hardware.Therefore, the two points“programming conscious of parallelism” and “programming conscious of execution performance”are essential techniques for users, researchers, and programmers who use the present supercomputers equipped with tens of thousands of processors and containing various enhancements and new functions. Here, we call the technique as the performance-optimizing techniques to application programs. Performance optimization of application is not always done by the application developer. It is difficult to interpret applications developed by others, evaluate their performance, discover problems, and solve problems. To systemize of the techniques of performance optimization will provide useful information for engineers and researchers who want to optimize the execution performance of applications. In this talk, I will talk about systemization of the techniques of performance optimization for single CPUs and high parallelism. Specifically, the following content is included.
-Classification of applications from the viewpoint of single CPU performance.
-Explain of busy time.
-Relationship of busy time and performance.
-Relationship of busy time and classification of applications.
-Relationship of busy time and performance tuning.
-Maximum performance estimation when busy time depends only on bandwidth.
-Accumulation of tuning techniques each application classification.
-Classification of problem regarding high parallelism.
-Accumulation of tuning techniques each problem classification.