Buzzwords, such as “cloud” and “big data”, should not always be ridiculed even in the science community. It is true that they often lack precise definitions or academic novelties, but I don’t think researchers should underestimate the possibility that such buzzwords gather people’s attention and some of the people will produce realistic values.
In this first presentation from the new HPC Usability Team at AICS café, I will talk how this team will increase the number of users of supercomputers, which is one of the missions of this team, by making use of the power of buzzwords. Fortunately, some of the recently spotlighted techniques are closely related to supercomputing. For example, modern AI techniques heavily depend on huge computing power. IoT also has relations with supercomputing in some particular applications. If we succeed to show attractive examples of supercomputing in such emerging areas, more people will have interested in supercomputers. This presentation covers our strategy to realize such situation.

We have been developing the SCALE-LETKF system, utilizing a regional weather model known as the Scalable Computing for Advanced Library and Environment-Regional Model (SCALE-RM) and an ensemble data assimilation method known as the Local Ensemble Transform Kalman Filter (LETKF). The primary goal of the system is to make use of observational data obtained from an advanced weather radar, the Phased Array Weather Radar (PAWR), in numerical weather prediction (NWP). This new type of weather radar can observe heavy precipitation systems densely both in space and in time, providing an important data source for advanced weather monitoring; however, its effective use in high-resolution NWP is a frontier research topic. We investigate this rapid-update, high-resolution data assimilation problem using the SCALE-LETKF. We have also been applying the SCALE-LETKF system for an experimental near-real-time NWP system in Japan and surrounding area. We run 5-day weather forecast for this area every 6 hours in near real time using the K computer. This has been running stably and has provided important guidance on the performance of the model and the data assimilation settings. In addition, this talk will also discuss about the computational aspect of the SCALE-LETKF development and some ongoing issues and prospects of the system.

Research and development of quantum annealing, which is expected to perform combinatorial optimization processing at high speed and with high accuracy, has been actively advanced from three perspectives of hardware development, software development, and application search. To use a quantum annealing machine, we should perform the following steps. At first, by using the Ising model, we express the combinatorial optimization problem that we want to solve. Second, the Ising model obtained in the previous step is embedded in the qubit network of the quantum annealing machine. Next, we introduce a strong transverse magnetic field that expresses quantum fluctuation and gradually weaken it. We obtain a solution of the combinatorial optimization problem as a final state.

In this presentation, I will introduce the theoretical background of quantum annealing and previous studies. Also, studies on quantum annealing that I have done with collaborators will be shown. In addition, I will introduce the issues to overcome in order for quantum annealing to become a useful calculation technique.

Tropical cyclone genesis is one of the most challenging issues in tropical cyclone studies because of the complicated processes involved.
In the previous studies, the distinctive flow patterns found in tropical cyclone genesis are categorized.
However, a subjective analysis method was used in the categorize, and it is not easy to apply to statistical analysis.
Therefore, I newly developed an objective analysis method of low-level flow patterns.
By using this method, I have found that the intermediate scale environment mainly consists of shear line, confluence region of zonal winds, and easterly waves. Furthermore, I have revealed that the flow patterns are modulated by inter-seasonal oscillation. These results have largely contributed to advance the tropical cyclone genesis study by connecting largescale and mesoscale.

Plasma, the fourth state of matter where constituent atoms are ionized, is very common in the universe. The goal of plasma physics is to understand the behavior of plasmas around the sun or planets and to realize a fusion power plant by confining high-temperature plasmas in a controlled manner.
In plasma, we observe rich varieties of nonlinear phenomena. For example, unique structures, such as loop-like magnetic fields on the solar surface, or zonal flows in the planetary atmosphere, are self-generated through complex dynamics. Those structures are also known to exist in fusion plasmas.
In this talk, I briefly review plasma physics as a nonlinear science, and introduce our recent study of the magnetic reconnection process, a typical nonlinear phenomenon in plasmas.

In the upcoming post-moore era, for hardware resources distributed on many chips in a large system to efficiently be utilized under high network latency, data-flow computing is promising due to its decentralized synchronization feature. Although software implementation is expected to have high overhead for data-flow computing with many threads, we focus on offloading some parts of them to customized hardware on FPGAs. We introduce our big picture for data-flow HPC and our research progress including a data-flow compiler for FPGAs.

Venus is sometimes called Earth's twin, because the planetary radius and the surface gravity of both are almost the same. On the other hand, the rotation period of Venus is very long as 243 Earth days, in other words the solid part is rotating slowly. However, the Venusian atmosphere rotates around the planetary-rotation axis in about 4 Earth days, about 60 times faster than its solid part. This fast zonal wind is called 4-day circulation or superrotation. To explore and reveal the mechanism of the atmospheric superrotation of Venus, numerical simulations with global atmospheric models have been performed. In this presentation, we review the simulation researches of the Venusian atmospheric circulation and show recent results of our high-resolution simulation. In particular, we focus on the planetary-scale streak structure suggested by the latest observation by the Venus Climate Orbiter/AKATSUKI, and discuss its mechanism.

Molecular wave packet dynamics triggered by optical and non-adiabatic transitions as well as intersystem crossing caused by spin-orbit couplings has a potential for creating a new molecular state. Photochemical process driven by non-resonant intense short pulse lasers involves a creation of various non-stationary non-adiabatic electron wave packets, which opens new paths for hidden chemical reaction and bond alternation. Toward a novel molecular state, we explore the nature of electron dynamics in chemical reactions and excited state energy transfer in terms of a quantum coherence, decoherence and entanglement created by wave packet branching.

【14:00-15:00】Keita Teranishi (Sandia National Laboratories, California)
Title: Toward Resilient Asynchronous Many Task Programming
Abstract: As semiconductor technology reaches its physical limit, the performance improvement of high performance computing systems no longer follows the predictions by Moore’s law. One of the viable approaches to address this stagnation is to relax the reliability of computing systems, and leave the application users to manage it. To enable this idea, it is essential for programming model to embrace a resilience capability. Today, the major resilience framework is coordinated checkpoint and restart (C/R), which involves global coordination of processes and threads for accommodating consistent global application state. However, this global recovery model entails inherent scalability issues and disproportionate use of resources to respond to local failures. These issues are better handled through asynchronous many task (AMT) programming model that is intended for deriving good scalability from unprecedented parallelism and complexity of node architecture of future HPC systems. A runtime system with AMT enables abstractions for encapsulating streams of program execution (tasks) and organizing the application data as objects rather than a sequence of data mapped to the system address space. In the AMT model, resilience is achieved through task re-execution and replication facilitated with versioning and replication of data objects. However, extensive research on task-based resilience is still required to determine the roadmap of resilience in the context of the programming environment. We will discuss our ongoing activities on the resilience of high performance AMT programming model and the challenges for scalable HPC application resilience.

【15:00-15:30】Break

【15:30－16:00】Miwako Tsuji（Programming Environment Research Team）
Title: Fault tolerance features in an XMP-YML scientific workflow programming model
Abstract: Supercomputers in the exa-scale era would consist of a huge number of nodes arranged in a multi-level hierarchy. There are many important challenges to exploit such systems such as scalability, programmability, reliability etc... In this talk, we focus on the scalability, programmability and fault tolerance features of a multi SPMD programming model. We have developed a development and execution environment based on workflow and PGAS (Partitioned Global Address Space) . We have extended the environment by incorporating fault resilience scheduling policy into the workflow scheduler.

【16:00-16:30】Balazs Gerofi (System Software ResearchTeam)
Title: IHK/McKernel: A Lightweight Multi-kernel based Operating System for Extreme Scale Supercomputing
Abstract: RIKEN Advanced Institute for Computation Science leads the development of Japan's next generation flagship supercomputer, the successor of the K Computer. Part of this effort is to design and develop a system software stack that suits the needs of future extreme scale computing. In this talk, we focus on operating system research and discuss IHK/McKernel, our multi-kernel based operating system framework. IHK/McKernel runs Linux with a light-weight kernel side-by-side on compute nodes with the primary motivation of providing scalable, consistent performance for large scale HPC simulations, but at the same time to retain a fully Linux compatible execution environment. We present an overview of the system architecture, provide preliminary results on up to two thousand Intel Xeon Phi nodes and outline future research directions.

Data assimilation, the process by which atmospheric data is combined with atmospheric models, is essential for skillful weather forecasts. Inserting data into models allows us to characterize more accurately the weather state at the start of the forecast, thereby extending the time in which the forecast is useful. However, data assimilation is a very computationally expensive process, and often costs as much as the actual weather forecasts. One way to reduce the cost is to lower the precision of the data assimilation computations. Lower precision computations use fewer bits to produce the answer to a calculation, and are therefore computationally cheaper. Lowering precision also introduces errors, but these errors may be acceptable, given that our models and observations are imperfect.
I will present some results from my PhD research on the subject of precision in data assimilation. I will show how lowering the precision of the data assimilation algorithm affects the quality of the output. I will demonstrate that the lowest precision that you can use is related to the overall quality of the model and the observations – the better the model, the more important precision becomes. Additionally, I will show that, by reusing the computational resources that we save when lowering precision, we can actually improve the data assimilation product and – ultimately – the skill of weather forecasts.