https://www.fujitsu.com/global/about/resources/publications/technicalreview/ The Fujitsu Technical Review is a web-based publication launched in spring 2020 to deliver information on Fujitsu technology. It presents technical papers, technical explanations, round table discussions, and other articles focused on technology. May 15, 2020 00:00:00 +0900 2020-5-15T00:00:00+09:00 https://www.fujitsu.com/global/about/resources/publications/technicalreview/2020-02/article05.html?20210113 Jan,13 2021 00:00:00 +0900 As cutting-edge IT such as AI, IoT, and 5G become more prevalent in society, digital transformation (DX)—which transforms business by applying these technologies to utilize data in more sophisticated way—is making progress. If developed in wide-ranging industrial fields to realize a data-driven society, DX is expected to contribute to the resolution of various social issues and the achievement of Sustainable Development Goals (SDGs), in addition to the growth of individual companies. Against this backdrop, Fujitsu has established a policy of strengthening its DX business to promote its transformation from an IT company to a DX company. In this article, Corporate Executive Officer Izumi Nagahori, who supervises Fujitsu’s DX and data utilization businesses as the Head of the Digital Software & Solutions Business Group, is interviewed about his view on the value brought by Fujitsu’s idea of DX, Fujitsu’s concept of an ideal DX company and the vision for realizing that ideal. (Interview conducted in March 2020) https://www.fujitsu.com/global/about/resources/publications/technicalreview/2020-02/article05.html?20210113 2021-1-13T00:00:00+09:00 https://www.fujitsu.com/global/about/resources/publications/technicalreview/2020-03/index.html?20201111 Nov,11 2020 00:00:00 +0900 Supercomputer Fugaku is being jointly developed and improved by RIKEN and Fujitsu with the goal of creating a versatile and easy-to-use supercomputer that achieves the worldʼs top level of performance with low power consumption for the purpose of resolving various issues faced by modern society and priority issues in science and technology fields, as well as utilization in a broad range of fields including Big Data and AI. This special feature covers the A64FX CPU and the system implementation technologies that were developed for the Fugaku to achieve high-speed application execution performance with low power consumption, the system software for supporting stable operation and user-friendliness, the operation and system management software, and the application development environment. In addition, it also introduces Fugaku development from the perspective of applications in their fields of use. https://www.fujitsu.com/global/about/resources/publications/technicalreview/2020-03/index.html?20201111 2020-11-11T00:00:00+09:00 https://www.fujitsu.com/global/about/resources/publications/technicalreview/2020-02/index.html?20200515 May 15 2020 00:00:00 +0900 Today, as digitalization accelerates and the global network breaks down existing boundaries between society and business, making optimal use of data and achieving a digital transformation is becoming a major issue in all companies and organizations and in all corners of society. Initiatives such as Sustainable Development Goals (SDGs), Data Free Flow with Trust (DFFT), Society 5.0, and Connected Industries are being undertaken to bring about a quantum leap in prosperity and achieve sustainable growth. Achieving a “data-driven society” is the key to solving this issue and the research and development of data utilization technologies is the driving force behind this pursuit. This special feature introduces the research and development of data utilization technologies for achieving a data-driven society and Fujitsu Group initiatives for implementing these technologies in society. https://www.fujitsu.com/global/about/resources/publications/technicalreview/2020-02/?20200515 2020-05-15T00:00:00+09:00 https://www.fujitsu.com/global/about/resources/publications/technicalreview/2020-01/index.html?20200515 May 15 2020 00:00:00 +0900 We are now entering the third AI boom with the evolution of AI technologies such as deep learning. As research and development for practical applications of AI are progressing in various fields, questions as to the reliability of black-box AI are being raised. Fujitsu is conducting various R&D activities to achieve trustworthy AI to solve this problem. This special feature introduces cutting-edge technologies through papers selected for their focus on AI and the importance of explainability for AI trustworthiness . In addition, a round-table discussion by young Fujitsu engineers working at the forefront sheds light on some of the efforts being made to win the trust of customers in AI application projects. https://www.fujitsu.com/global/about/resources/publications/technicalreview/2020-01/index.html?20200515 2020-05-15T00:00:00+09:00 https://www.fujitsu.com/global/about/resources/publications/technicalreview/topics/article003.html?20201008 October 8 2020 00:00:00 +0900 In recent years, there has been an increase in use of video images in various fields such as detection of suspicious people on the street and analysis of operator behaviors on production lines. With the advent of fifth-generation mobile communications system (5G) services featuring high-speed, large-capacity communications, which enable the use of ultra-high-definition 4K video and high-volume video captured by cameras, there are also high expectations that such images can be analyzed by AI to recognize small behaviors overlooked by the human eye. AI analysis of video data requires a massive number of calculations. As the number of cameras used changes, the number of calculations also changes. The cloud is suitable for these types of calculations, as it is highly scalable and users only need to pay for what they use. Conventionally, however, it has been difficult to use the cloud for this purpose, because transmission of large video data uses up the network bandwidth. To solve the above problem, Fujitsu has independently developed a technology that achieves high compression of video data specialized for AI analysis. The application of the new technology enables video analysis with the cloud. This article describes the problems with conventional video compression technology and provides details of our video data high-compression technology for AI analysis, along with its evaluation results. https://www.fujitsu.com/global/about/resources/publications/technicalreview/topics/article003.html?20201008 2020-10-08T00:00:00+09:00 https://www.fujitsu.com/global/about/resources/publications/technicalreview/topics/article004.html?20201009 October 9 2020 00:00:00 +0900 Traditionally, magnetic tape storage has been used mainly for backup purposes. Owing to its large capacity and low cost, as well as recent increases in transfer speed and the spread of the linear tape file system (LTFS), which is a file system for tape that enables the handling of magnetic tape data in file units, magnetic tape storage is now expected to find increasing use for archival purposes. However, while magnetic tape storage allows fast sequential accessing of continuous areas on tape, random accessing of discontinuous locations was slow and had to be improved for magnetic tape storage to become usable for archiving large amounts of data. In response, Fujitsu Laboratories expanded the functionality of LTFS and developed methods for the virtual integration of multiple tape cartridges, data management according to tape characteristics, and access order control to boost the random access performance of magnetic tape. https://www.fujitsu.com/global/about/resources/publications/technicalreview/topics/article004.html?20201009 2020-10-9T00:00:00+09:00 https://www.fujitsu.com/global/about/resources/publications/technicalreview/topics/article005.html?20200821 August 21 2020 00:00:00 +0900 The K computer and its successor supercomputer “Fugaku” are world-class supercomputers that are massively parallel computers made up of 88,192 and 158,976 interconnected nodes respectively. This 100k-node scalability is made possible by an interconnect technology developed by Fujitsu that features high dimensionality. The partitioning and virtual torus functions of the technology prevent interference in the communications between multiple parallel programs and support optimization of communication patterns within each parallel program to ensure stable communication performance, and allow partitions to remain in use even when they contain failed nodes for high availability. This article describes the interconnect technology with high dimensionality used in the K computer and the supercomputer Fugaku. https://www.fujitsu.com/global/about/resources/publications/technicalreview/topics/article005.html?20200821 2020-08-21T00:00:00+09:00 https://www.fujitsu.com/global/about/resources/publications/technicalreview/topics/article002.html?20200812 August 12 2020 00:00:00 +0900 This article describes an approach to solving social issues by connecting sites facing problems with advanced technologies. In recent years, attention has been focused on solving social issues with advanced technologies such as AI, but as of yet there have been few successful examples. The main reason for this is that due to the complexity and uncertainty of projects aimed at solving social issues, it is difficult to identify the essential problems to be solved. To address this problem, we have adopted the concept of systems science, which provides a framework for the problem-solving process for a variety of subjects with different complexities and uncertainties. Based on this, we present the difficulty of solving social issues using advanced technologies such as AI and have proposed a new social system design approach to connect sites with technology. https://www.fujitsu.com/global/about/resources/publications/technicalreview/topics/article002.html?20200812 2020-08-12T00:00:00+09:00 https://www.fujitsu.com/global/about/resources/publications/technicalreview/topics/article001.html?20200805 August 5 2020 00:00:00 +0900 The optical network accommodating massive amounts of data traffic has increased its capacity dramatically by ten times over the last five years. From here on, however, it will be necessary to achieve even larger capacities in the 400 Gbps—1 Tbps per wavelength class beyond today’s mainstream optical transmission capacity of 100 Gbps to accommodate ever-increasing volumes of traffic. To achieve this, it will be necessary to overcome waveform distortion (nonlinear distortion) in the optical signal caused by nonlinear optical effects in the optical fiber transmission, which is a remaining limiting factor that hinders increases in capacity. A number of relatively simple proposals for mitigating such nonlinear distortion have been made at academic societies, but the circuit scale of these proposals has been extremely large preventing them from being realistic solutions. We have overcome this problem by proposing a nonlinear distortion compensation method that greatly reduces the computational complexity of nonlinear distortion compensation with respect to multilevel and polarization multiplexing in current optical transmission. Using this technology, we have achieved a practical optical transceiver equipped with nonlinear distortion compensation as a world’s first. https://www.fujitsu.com/global/about/resources/publications/technicalreview/topics/article001.html?20200805 2020-08-05T00:00:00+09:00 https://www.fujitsu.com/global/about/resources/publications/technicalreview/2020-02/article02.html?20200515 May 15 2020 00:00:00 +0900 Various approaches to data utilization are currently being taken by industry, government, and academia. The Japanese government has positioned data utilization as a key measure for the realization of Society 5.0, and is promoting various proposals and national projects, such as Data Free Flow with Trust (DFFT). For its part, Academia is also conducting research ranging from basic science to applied research for the realization of a data-driven society. In the private sector, an alliance for the promotion of standardization and the formulation of technical standards for data distribution among business operators has been established. Against this backdrop, Fujitsu is working on a cross-sector data exchange platform and information banks. This article provides an overview of the current status of activities related to data utilization in industry, government, and academia, and it introduces Fujitsu’s own activities in this area. https://www.fujitsu.com/global/about/resources/publications/technicalreview/2020-02/article02.html?20200515 2020-05-15T00:00:00+09:00 https://www.fujitsu.com/global/about/resources/publications/technicalreview/2020-02/article03.html?20200515 May 15 2020 00:00:00 +0900 This article describes research and development of Fujitsu AI technologies toward digital transformation (DX) of the manufacturing industry. Manufacturing sites backed by high productivity and excellent quality control are now faced with key issues such as a decline in the working-age population and difficulties in worker training. However, to provide high added value and boost market competitiveness, there is a need for revolutionizing not only manufacturing sites but also the entire manufacturing process. In response to these challenges, Fujitsu Laboratories of Europe (FLE) is working on solutions that has the potential to revolutionize both the manufacturing industry as well as digitalizing manufacturing sites through AI technologies. At FLE, we have developed new anomaly detection technology for detecting problems on manufacturing lines and detecting manufacturing defects in a wide variety of materials and components with high accuracy. We are also in the process of developing a deep-learning application technology called an AI simulator to also expand the application range of AI technologies to the design process by means of massively accelerated simulations of physical phenomena. https://www.fujitsu.com/global/about/resources/publications/technicalreview/2020-02/article03.html?20200515 2020-05-15T00:00:00+09:00 https://www.fujitsu.com/global/about/resources/publications/technicalreview/2020-02/article04.html?20200515 May 15 2020 00:00:00 +0900 Advances in IoT and 5G technologies have led to growing interest in “digital twins” that digitalize the real world in real time so that this data can be transformed into value by storing it on the cloud where it can be used without needing to know the devices it came from. Key requirements for building a digital twin are the ability to process large amounts of stream data at high speed and to make additions or changes to the processing without interrupting services. In response, Fujitsu Laboratories has developed the Dynamically-Reconfigurable Asynchronous Consistent EveNt processing Architecture (Dracena) as a reactive platform that can be used to develop and deploy services speedily and flexibly by creating cloud-based digital twins of people, things, and contexts from the real world. The platform features the ability to process large amounts of stream data in real time and to make additions or changes to this processing without shutting down the system. This article describes the architecture of Dracena, the technologies developed for it, and the applications where it is used. https://www.fujitsu.com/global/about/resources/publications/technicalreview/2020-02/article04.html?20200515 2020-05-15T00:00:00+09:00 https://www.fujitsu.com/global/about/resources/publications/technicalreview/2020-01/article02.html?20200527 May 27 2020 00:00:00 +0900 Congestion at facilities such as airports, shopping centers, and event venues occurs at choke points and popular stores where people congregate. Pedestrian simulation has attracted attention in recent years as a way to predict congestion and to visualize the effectiveness of congestion mitigation measures. While simulation is typically used to predict the future or visualize possibilities, a close analysis of the calculation process involved can also uncover the reasons why a phenomenon is caused. This is also the case with pedestrian simulation, only with the additional difficulty that identifying the causes of congestion takes a lot of effort by a person with specific expertise. To overcome this problem, Fujitsu Laboratories has developed a technique that uses micro-dynamics analysis to automatically analyze pedestrian simulation logs. When applied to the pedestrian simulation of an airport, the technique was able to identify the causes that made reasonable sense. Moreover, eliminating the causes succeeded in reducing congestion. This paper describes pedestrian simulation and the micro-dynamics analysis technique and presents the results of an evaluation experiment. https://www.fujitsu.com/global/about/resources/publications/technicalreview/2020-01/article02.html?20200527 2020-05-27T00:00:00+09:00 https://www.fujitsu.com/global/about/resources/publications/technicalreview/2020-01/article03.html?20200515 May 15 2020 00:00:00 +0900 AI technology which has been increasingly used in recent years has a disadvantage; it is difficult to explain the reasons behind its outputs in understandable ways to humans. To address this problem, research on explainable AI (XAI) is underway. However, it is necessary to enable XAI to interact with its users to support human decision making. As a prototype of interactive XAI, Fujitsu Laboratories has developed AI to recommend relocation destinations with the cooperation of the local government of Itoshima City, Fukuoka Prefecture in Japan. This is a recommender system that supports users in urban areas with decision-making when selecting suitable candidate relocation destinations. In response to the preferences input by users, the system provides not only lists of the candidate relocation destinations but also information for the users to learn about the destinations in detail. This article describes an interactive AI that realizes collaboration with humans. It also presents the results of a subject experiment on those who wish to relocate to Itoshima City and the future issues and direction. https://www.fujitsu.com/global/about/resources/publications/technicalreview/2020-01/article03.html?20200318 2020-05-15T00:00:00+09:00 https://www.fujitsu.com/global/about/resources/publications/technicalreview/2020-01/article04.html?20200515 May 15, 2020 00:00:00 +0900 There are increasing competitive and regulatory incentives to deploy AI mindfully within financial services. An important aspect towards that end is to explain AI decisions to various stakeholders. State-of-the-art explainable AI systems mostly serve AI engineers and offer little value to other stakeholders. Towards addressing this gap, we built the first dataset that is representative of loan-applicant friendly explanations. We also designed a novel generative adversarial network (GAN) that can accommodate smaller datasets to generate user-friendly purpose-driven explanations. In this article, we demonstrate how our system can generate explanations serving multiple purposes, including those that help educate loan applicants and those that help them take appropriate action towards a future approval. https://www.fujitsu.com/global/about/resources/publications/technicalreview/2020-01/article04.html?20200515 2020-05-15T00:00:00+09:00 https://www.fujitsu.com/global/about/resources/publications/technicalreview/2020-01/article05.html?20200515 May 15, 2020 00:00:00 +0900 Now that we have entered the third AI boom with the evolution of machine learning technologies such as deep learning, research and development for practical applications of AI are progressing in various fields, but at the same time questions as to whether AI can be trusted are increasingly arising. For example, research on explainable AI (XAI), which clarifies the way in which AI draws conclusions, is actively being carried out around the world. Further, at the G20 Ministerial Meeting on Trade and Digital Economy held prior to the G20 Osaka Summit, “AI Principles” summarizing what AI should be were drawn up. For its part, Fujitsu formulated the “Fujitsu Group AI Commitment”. Against this backdrop, Fujitsu’s AI engineers continuously ask themselves what they can do to make the AI they develop trustworthy for customers, partners, and even end users. As part of this quest, a round-table discussion with four young AI engineers who are active at the forefront of AI application projects was organized to clarify what trustworthy AI is through their efforts. https://www.fujitsu.com/global/about/resources/publications/technicalreview/2020-01/article05.html?20200515 2020-05-15T00:00:00+09:00 https://www.fujitsu.com/global/about/resources/publications/technicalreview/2020-03/article10.html?20201111 Nov,11 2020 00:00:00 +0900 Recent years have seen dramatic progress in the parallelization and scaling up of high performance computing (HPC) systems consisting of supercomputers and clusters of ordinary computers. On the other hand, ensuring that such a large-scale, high-performance computing environment operates in a stable manner for the benefit of users has become a major issue. Operations management software plays an important role in managing and operating a system. The development of operations management software for the supercomputer Fugaku (hereafter, Fugaku) was carried out with a view to achieving stable operation based on continuity with K computer know-how and on the solutions found to operation problems after eight years of operating the K computer and using its software technology. A user-friendly system was also a key pursuit of this development initiative. https://www.fujitsu.com/global/about/resources/publications/technicalreview/2020-03/article10.html?20201111 2020-11-11T00:00:00+09:00 https://www.fujitsu.com/global/about/resources/publications/technicalreview/2020-03/article09.html?20201111 Nov,11 2020 00:00:00 +0900 The development of AI technology is indispensable for the realization of Society 5.0, which has been announced as part of Japan's science and technology policy. As such development proceeds, the computational resources required for AI learning continue to increase. Through the development of the K computer and the supercomputer Fugaku (hereafter, Fugaku), Fujitsu has been providing high performance computing (HPC) systems with abundant computing resources. Now, in order to utilize the abundant computational resources of HPC systems for AI learning, we are working with RIKEN to develop an AI infrastructure on Fugaku. https://www.fujitsu.com/global/about/resources/publications/technicalreview/2020-03/article09.html?20201111 2020-11-11T00:00:00+09:00 https://www.fujitsu.com/global/about/resources/publications/technicalreview/2020-03/article08.html?20201111 Nov,11 2020 00:00:00 +0900 In this article, I would like to take a historical look at activities on the application side during supercomputer Fugaku’s climb to the top of supercomputer rankings as seen from a vantage point in early July 2020. I would like to describe, in particular, the change in mindset that took place on the application side along the path toward Fugaku’s top ranking in four benchmarks in 2020 that all began with the Application Working Group and Computer Architecture/Compiler/System Software Working Group in 2011. Somewhere along this path, the application side joined forces with the architecture/system side based on the keyword “co-design.” During this journey, there were times when our opinions clashed and when efforts to solve problems came to a halt, but there were also times when we took bold steps in a unified manner to surmount difficult obstacles. I will leave the description of specific technical debates to other articles, but here, I would like to describe the flow of Fugaku development from the application side based heavily on a “science-driven” approach along with some of my personal opinions. Actually, we are only halfway along this path. We look forward to the many scientific achievements and solutions to social problems that Fugaku is expected to bring about. https://www.fujitsu.com/global/about/resources/publications/technicalreview/2020-03/article08.html?20201111 2020-11-11T00:00:00+09:00 https://www.fujitsu.com/global/about/resources/publications/technicalreview/2020-03/article07.html?20201111 Nov,11 2020 00:00:00 +0900 The objectives of developing the supercomputer Fugaku (hereafter, Fugaku) are to achieve a system with up to 100 times greater application performance than that of the K computer and to ensure general versatility to support the operation of a broad range of applications. Accomplishing this required an integrated development environment that supports everything from the development to the execution of applications that extract the maximum performance from the hardware, but there were significant challenges in achieving this goal. First, it required development optimized for the CPU used by Fugaku and support for various applications. Moreover, it required support for new hardware and increased memory usage due to the increase in the total number of processes in the message passing interface (MPI) communication library. In addition, it required improvements of the practical usability of application development support tools. This article describes such initiatives for the application development environment. https://www.fujitsu.com/global/about/resources/publications/technicalreview/2020-03/article07.html?20201111 2020-11-11T00:00:00+09:00 https://www.fujitsu.com/global/about/resources/publications/technicalreview/2020-03/article06.html?20201111 Nov,11 2020 00:00:00 +0900 RIKEN and Fujitsu are working on the development of the supercomputer Fugaku (hereafter, Fugaku) as a successor to the K computer, with general public service slated to commence in FY 2021. While Fugaku will inherit many of the software assets of the K computer, we are working on various improvements such as computational performance, resource utilization efficiency, and usability. https://www.fujitsu.com/global/about/resources/publications/technicalreview/2020-03/article06.html?20201111 2020-11-11T00:00:00+09:00 https://www.fujitsu.com/global/about/resources/publications/technicalreview/2020-03/article05.html?20201111 Nov,11 2020 00:00:00 +0900 RIKEN and Fujitsu are jointly developing the supercomputer Fugaku as the successor to the K computer with a view to starting public use in FY2021. While inheriting the software assets of the K computer, the plan for Fugaku is to make improvements, upgrades, and functional enhancements in various areas such as computational performance, efficient use of resources, and ease of use. As part of these changes, functions in the file system have been greatly enhanced with a focus on usability in addition to improving performance and capacity beyond that of the K computer. Additionally, as reducing power consumption and using power efficiently are issues common to all ultra-large-scale computer systems, power management functions have been newly designed and developed as part of the upgrading of operations management software in Fugaku. https://www.fujitsu.com/global/about/resources/publications/technicalreview/2020-03/article05.html?20201111 2020-11-11T00:00:00+09:00 https://www.fujitsu.com/global/about/resources/publications/technicalreview/2020-03/article04.html?20201111 Nov,11 2020 00:00:00 +0900 The supercomputer Fugaku (hereafter, Fugaku) achieved the world’s top position in four categories of computing performance according to supercomputer rankings announced in June 2020.The prototype of Fugaku also won the world’s first place in the performance per power consumption ranking announced in November 2019. Realization of such high performance and power saving owes not only to the performance of the newly developed CPU but also the system implementation technologies such as the four times larger number of CPUs per rack mounted as compared with predecessor K computer. https://www.fujitsu.com/global/about/resources/publications/technicalreview/2020-03/article04.html?20201111 2020-11-11T00:00:00+09:00 https://www.fujitsu.com/global/about/resources/publications/technicalreview/2020-03/article03.html?20201111 Nov,11 2020 00:00:00 +0900 The A64FX was developed as a processor for the supercomputer Fugaku. For the semiconductors, the 7- nm CMOS process technology of TSMC has been used. For higher density, the Tofu interconnect D and the PCI Express controllers have been implemented in the CPU chip, and the high-bandwidth 3D stacked memory is integrated in the package. https://www.fujitsu.com/global/about/resources/publications/technicalreview/2020-03/article03.html?20201111 2020-11-11T00:00:00+09:00 https://www.fujitsu.com/global/about/resources/publications/technicalreview/2020-03/article01.html?20201111 Nov,11 2020 00:00:00 +0900 For the purpose of leading the resolution of various social and scientific issues surrounding Japan and contributing to the promotion of science and technology, strengthening industry, and building a safe and secure nation, RIKEN (The Institute of Physical and Chemical Research), and Fujitsu have worked on the research and development of the supercomputer Fugaku (hereafter, Fugaku) since 2014. The installation of the hardware was completed in May 2020 and, in June, the world’s top performance was achieved in terms of the results of four benchmarks: Linpack, HPCG, Graph500, and HPL-AI. At present, we are continuing to make adjustments ahead of public use in FY2021. https://www.fujitsu.com/global/about/resources/publications/technicalreview/2020-03/article01.html?20201111 2020-11-11T00:00:00+09:00