We have been carrying out the FLAGSHIP 2020 Project to develop the Japanese next-generation flagship supercomputer, the Post-K, recently named “Fugaku”. We have designed an original many core processor based on Armv8 instruction sets with the Scalable Vector Extension (SVE), an A64FX processor, as well as a system including interconnect and a storage subsystem with the industry partner, Fujitsu. The “co-design” of the system and applications is a key to making it power efficient and high performance. We determined many architectural parameters by reflecting an analysis of a set of target applications provided by applications teams. In this paper, we present the pragmatic practice of our co-design effort for “Fugaku”. As a result, the system has been proven to be a very power-efficient system, and it is confirmed that the performance of some target applications using the whole system is more than 100 times the performance of the K computer.

In this paper, we introduce a new and highly scalable interconnect called Tofu interconnect D that will be used in the post-K machine. This machine will officially be operational around 2021. The letter D represents high “density” node and “dynamic” packet slicing for “dual-rail” transfer. Herein we describe the design and the evaluation results of TofuD. Due to the high-density packaging, the optical link ratio of TofuD has decreased to 25% from the 66% optical link ratio of Tofu2. TofuD applies a new technique called dynamic packet slicing to reduce latency and to improve fault resilience. The evaluation results show that the one-way 8-byte Put latency is 0.49 μs. This is 31% lower than the latency of Tofu2. The injection rate per node is 38.1 GB/s which is approximately 83% of the injection rate of Tofu2. The link efficiency is as high as approximately 93%.

Evaluating semantic similarity of text document pairs is an active research topic. Various models of document representation have been proposed. Each kind of representation model concentrates on a different kind of information from other kind of models. However, it is difficult for a single model to perform well in all scenarios because of the variety of textual documents. Leveraging these models to complement each other is possible to improve the performance. In this paper, we first make an analysis on the relations among document semantic similarity, human ratings and model performance. Based on the observations, we propose a rational solution of selecting different representation models and fusing the results of these models to compute document similarity for a given document collection. We leverage the performance and relations of different models to select proper models. Our fusion approach proposes a regression function with both nonlinear and linear factors and dynamic weights based on the similarities by various models. We report the effectiveness of our work based on a rated news document collection. The particular version of our general approach for this collection can integrate the information from both brief entity knowledge and detailed word content.

Vast amounts of earth science data have been stored and managed in various projects. Because earth science requires considerable experience and expertise, we generally generate metadata for earth science datasets. In such metadata, keyword information is important. Controlled keywords such as GCMD Science Keyword are used for metadata. However, selection of suitable keywords is not easy. We propose methods to suggest keywords from summary texts of datasets. We considered simple string matching and inferring by Labeled LDA, and apply them to the actual earth science metadata.

An efficient all-to-all communication algorithm for torus and mesh networks, A2AT, was proposed. A2AT schedules message sending sequence so that all links are fully used by exploiting function of concurrent message transfer in the node. By using A2AT, the hop count of messages equals the maximum number of messages sharing a link in their routes for all message transfers. A2AT can therefore maintain synchronization without the need for phasing operation such as an MPI barrier. When the VOQ which is an ideal configuration for A2AT was used, communication times for mesh/torus network obtained by A2AT were roughly 1.20 and 1.09 times higher, on average, than those of the ideal times. When the networks had the minimum number of virtual channels and a small buffer, assuming a practical network, A2AT was able to reduce communication times by 12.5% and 36.0% compared with those of the conventional algorithm. When two controllers are used, A2AT reduced 28.2% and 55.7% communication time with those by A2AND on 15×15×15 (=3,375 nodes) mesh and torus networks respectively (18.6% and 44.8% in average). A2AT also reduced 15.1% and 41.9% of communication time with those by A2AND on the same mesh and torus networks respectively (14.4% and 37.5% in average) when six controllers are used.

Chaotic dynamics in ear plethysmogram system was studied in both tracking experiment and a mathematical model. The experiments were conducted by using of a driving handle to track moving targets on a PC screen. The moving targets were driven by signals of different predictabilities with sin wave, chaotic and random waves. We found that chaos happened in ear plethysmogram and there were changes in the chaotic dynamics during tracking the signals with different predictable natures. The largest Lyapunov exponent of the plethysmograms was found to be significant higher during tracking chaotic signals. To understand how the chaos arises and to explain the changes in the Lyapunov exponent in ear plethysmograms in experiments, a mathematical model consisting of baroreflex feedback and autonomous interactions was proposed and studied numerically. The increase of the largest Lyapunov exponent in plethysmograms was explained successfully in the model in relation to the increased chaoticity, and hence the increased active temporal information processing in higher cerebral region.

This paper proposes a steady-state genetic algorithm for trajectory generation used in the imitation of a partner robot interacting with a human. Various types of genetic algorithms have been applied for the trajectory generation of robot manipulators. In this paper, we propose a trajectory generation method for the partner robot by a steady-state genetic algorithm based on the human motions pattern, and compare the proposed method with its related methods. Finally, we show experimental results of trajectory generation through interaction with a human.

Tracking experiments were conducted using the moving targets with different predictabilities. Human temporal information processing during the tracking was examined using physiological measurement. We used nonlinear methods of DFA, chaos theory and phase dynamics to analyzing the data. Our results on tracking errors in positions showed that the most unpredictable random target lead to most larger variance, whereas the target of most predictable sin wave and the one of chaotic motion with predictable in short term had lower variances. As for the tracking error with respect to instant phase, it found that target of sin wave gave arising a minimizing variance in compare with chaos and random. Further analysis by DFA showed that chaotic target resulted in significantly larger in scaling exponent for the instant phase differences between target and tracking, implying the most correlated process during the tracking. In addition, we found clearly changes in physiological status of subjects observed by ear plethysmogram during the tracking tasks. The largest Lyapunov exponent of the plethysmograms was found to be significant higher during tracking chaotic motion, showing a most active temporal information processing.

Imitation is a powerful tool for behavior learning and human communication. Basically, imitative learning is composed of model observation and model reproduction. This paper applies a spiking neural network and self-organizing map for model observation, and modular fuzzy neural networks and a steady-state genetic algorithm for model reproduction. The proposed method is applied for a partner robot interacting with a human. Experimental results show that the proposed method enables a robot to learn behaviors through imitation and can interact with a human efficiently.

We studied changes in fractal nature of facial skin thermogram in experiments conducted by two kinds of mental tasks. One of the tasks was associated with arithmetic and logistic mental activities. Another one was in relation with imagination and thinking activities. Performances using the mental tasks appeared to create different facial thermogram patterns that could be characterized quantitatively in the use of fractal dimensions. In our experiments, there was a lower fractal dimension found during arithmetic task in comparison with rest period before task, showing a smoother facial thermal imagery. Whereas in contrast, the task of thinking induced an increase of fractal dimension of the thermal imagery, leading to a more fluctuated facial thermal surface.

This paper discusses the design and implementation of a cluster file system, called PVFS-PM, on the SCore cluster system software. This is the first attempt to implement a cluster file system on the SCore system. It is based on the PVFS cluster file system but replaces TCP with the PMv2 communication library supported by SCore to provide a scalable, high-performance cluster file system. PVFS-PM improves the performance by factors of 1.07 and 1.93 for writing and reading, respectively, with 8 I/O nodes, compared with the original PVFS on TCP on a Gigabit Ethernet-connected SCore cluster.

The scalability of distributed-memory parallel computers makes them attractive candidates for solving large-scale problems. New languages, such as HPF, FortranD, and VPP Fortran, have been developed to enable existing software to be easily ported to such machines. Many distributed-memory parallel computers have been built, but none of them support the mechanisms required by such languages. We studied the mechanisms required by parallelizing compilers and proposed a new architecture to support them. Based on this proposed architecture, we developed a new distributed-memory parallel computer, the AP1000+, which is an enhanced version of the AP1000. Using scientific applications in VPP Fortran and C, such as NAS parallel benchmarks, we simulated the performance of the AP1000+.

The performance of message-passing applications depends on cpu speed, communication throughput and latency, and message handling overhead. In this paper we investigate the effect of varying these parameters and applying techniques to reduce message handling overhead on the execution efficiency of ten different applications. Using a message level simulator set up for the architecture of the AP1000, we showed that improving communication performance, especially message handling, improves total performance. If a cpu that is 32 times faster is provided, the total performance increases by less than ten times unless message handling overhead is reduced. Overlapping computation with message reception improves performance significantly. We also discuss how to improve the AP1000 architecture.

Low-latency communication is the key to achieving a high-performance parallel computer. In using state-of-the-art processors, we must take cache memory into account. This paper presents an architecture for low-latency message comunication and implementation, and performance evaluation.