Publications

Journal Articles

2025

1. Directional Handover Analysis with Stochastic Petri Net and Poisson Point Process in Heterogeneous Networks

   Zhiyi Zhu, Junjun Zheng, Eiji Takimoto, Patrick Finnerty, and Chikara Ohta

   Mathematics, Jan 2025

   Abs DOI

   Handover is crucial for ensuring seamless connectivity in heterogeneous networks (HetNet) by enabling user equipment (UE) to switch its connection link between cells based on signal conditions. However, conventional analytical approaches ignored the distinctions between macro-cell to small-cell (M2S) and small-cell to macro-cell (S2M) scenarios during a handover decision-making process, which resulted in handover failures (HoF) or ping-pong handovers. Therefore, this paper proposes a novel framework, Do-SPN-PPP, that combines stochastic Petri net (SPN) and the Poisson point process (PPP) to quantitatively analyze M2S and S2M handover performance differences. The proposed framework also reveals and predicts how handover parameters affect UE residence time in a cell within the HetNet, and it exhibits a higher predictive accuracy compared with the traditional conventional analytical approach. In addition, the Monte Carlo simulation verified the Do-SPN-PPP framework, and the proposed framework exhibits a 96% reduction in computation time while maintaining a 95% confidence interval and 0.5% error tolerance compared with the simulation.

2024

1. Applications of cluster-based transfer learning in image and localization tasks

   Liuyi Yang, Patrick Finnerty, and Chikara Ohta

   Machine Learning with Applications, Dec 2024

   Abs DOI Code

   Transfer learning can address the issue of insufficient labels in machine learning. Using knowledge in a labeled domain (source domain) can assist in acquiring and learning knowledge in a domain (target domain) that lacks some or all labels. In this paper, we propose a new cluster-based semi-supervised transfer learning (CBSSTL) under a new assumption that samples in the target domain are unlabeled but contain cluster information. Furthermore, we propose a new transfer learning framework and a method for fine-tuning parameters. We tested and compared the proposed method with other unsupervised and semi-supervised transfer learning methods on well-known image datasets. The experimental results demonstrate the effectiveness of the proposed method. Additionally, we created a localization dataset for transfer learning. Finally, we tested and analyzed the proposed method on this dataset. Its particularly challenging nature makes it difficult for our method to work effectively.

2. Preamble Arbitration Rule and Interference Suppression-Based Polling Medium Access Control for In-Vehicle Ultra-Wideband Networks

   Makoto Okuhara, Nobuyuki Kurioka, Shigeki Mitoh, Patrick Finnerty, and Chikara Ohta

   IEEE Open Journal of Vehicular Technology, Oct 2024

   Abs DOI

   This paper introduces a preamble arbitration rule and interference suppression (PARIS) method for ultra-wideband (UWB) in-vehicle networks. Advancements in the automotive technology have led to increased reliance on wire harnesses, resulting in higher costs, electronic integration challenges, and adverse environmental effects. To address these problems, we explored the use of UWB wireless networks, which are characterized by low transmission power and superior signal penetration capabilities. A significant challenge associated with implementing UWB in automotive environments is the increased frame error rate (FER) caused by UWB interference. Our experiments indicate that vehicles equipped with identical UWB networks exhibit an FER of approximately 6% when positioned closely. This level of FER is problematic for automotive applications, where reliable communication is paramount. To mitigate this problem, we developed an PARIS communication algorithm that is robust against interference. As identified in this study, PARIS leverages two key characteristics of UWB. First, it prioritizes the timing of signal reception over radio signal power, enhancing interference suppression by activating the receiver at the optimal moment before the desired frame arrives, thereby minimizing data loss. Second, the algorithm exploits the hierarchical nature of preamble codes in simultaneously received frames, reducing data loss rate to the order of 10−5 by prioritizing frames from critical communication devices based on the preamble code hierarchy. Implementing the UWB-based PARIS method in wireless vehicle networks can reduce the weight of the wire harnesses by approximately 20%, offering a promising solution to the challenges posed by traditional wiring systems.

3. A CQI and Hysteretic-based Decision Algorithm to Prevent Handover Failures for Pedestrian Mobility in Mobile Communication HetNet

   Zhiyi Zhu, Eiji Takimoto, Patrick Finnerty, and Chikara Ohta

   IEEE Access, Jul 2024

   Abs DOI

   In current mobile communication networks, handover substantially influences the performance of heterogeneous networks (HetNet). The decision process and trigger timing of the handover decision algorithm (HDA) are integral aspects of handover and are indispensable for mobile user equipment (UE). However, the major A3RSRP (reference signal received power) HDA with the parameter of handover margin (HOM) used in co-frequency environments has poor performance with low-speed pedestrian scenarios in HetNet, including high handover failure (HoF) ratio, frequent handover (FHO), and high number of HoFs. In this paper, we focus on reducing the HoF ratio and propose a novel CQI and hysteretic (CH)-based HDA using downlink channel quality indicator (CQI) and hysteretic control (HC). This novel HDA enables the cell to check the downlink CQI of the UE during HDA, thereby correctly determining the trigger timing of the handover. Moreover, the cell can use the HC to dynamically adjust the HOM based on the CQI of the UE to adapt the handover to the HetNet environment. We evaluate and compare the performance of the CH and A3RSRP HDAs in three scenarios with different speeds, each with multiple HOM settings. The ns-3 simulation results indicate that the CH HDA shows a maximum improvement of 100% in HoF ratio and 5% and 100% in the number of handovers and HoFs, respectively, in the small-scale environment with a speed of 0.5 m/s and an initial HOM of 2.0 dB, while the CH HDA shows a minimum improvement of-4% in HoF ratio and 58% and 44% improvement in the number of handover and HoFs, respectively, in the large-scale environment with a speed of 0.5 m/s and an initial HOM of 3.0 dB, compared with the A3RSRP HDA. INDEX TERMS A3RSRP, channel quality indicator, handover, handover decision algorithm, handover failure, HetNet, hysteretic control, ns-3

4. Evolving APGAS Programs: Automatic and Transparent Resource Adjustments at Runtime

   Jonas Posner, Raoul Goebel, and Patrick Finnerty

   Lecture Notes in Computer Science, May 2024

   Abs DOI Code

   In the rapidly evolving field of High-Performance Computing (HPC), the need for resource elasticity is paramount, particularly in addressing the dynamic nature of irregular computational workloads. A key area of elasticity lies within programming models that typically offer limited support.
   Fully elastic programs are both malleable—capable of dynamically adjusting resources in response to external job scheduler requests—and evolving—autonomously deciding when and how to adjust resources, e.g., through automated decision-making. Previous elasticity approaches typically relied on iterative workloads and required complex code modifications.
   Asynchronous Many-Task (AMT) programming is emerging as a powerful alternative. In AMT, computations are split into fine-grained tasks, allowing transparent task relocation by the runtime system and unlocking significant potential for efficient elasticity.
   This work-in-progress proposes an extension to the existing AMT APGAS that recently incorporated malleability. Our extension adds evolving capabilities providing automatic and transparent resource adjustments to meet changing computational workloads at runtime. Our easy-to-use abstractions require only minimal code additions; adjustments such as process initialization and termination are managed automatically. Our extension is validated via a load-balancing library for irregular workloads.
   We propose two heuristics for automatic computational load detection: one that uses CPU loads provided by the operating system, and another that exploits detailed insights into task loads. We evaluate our approach using a novel synthetic benchmark that starts with a single task evolving into two irregular trees connected by a long sequential branch. Preliminary results are promising, indicating that both the CPU-based heuristic and the task-based heuristic showing similar efficiency.

5. Polling schedule algorithms for data aggregation with sensor phase control in in-vehicle UWB networks

   Hajime Migita, Yuki Nakagoshi, Patrick Finnerty, Chikara Ohta, and Makoto Okuhara

   IEICE Transactions on Communications, Mar 2024

   Abs DOI

   To enhance fuel efficiency and lower manufacturing and maintenance costs, in-vehicle wireless networks can facilitate the weight reduction of vehicle wire harnesses. In this paper, we utilize the Impulse Radio-Ultra Wideband (IR-UWB) of IEEE 802.15.4a/z for in-vehicle wireless networks because of its excellent signal penetration and robustness in multipath environments. Since clear channel assessment is optional in this standard, we employ polling control as a multiple access control to prevent interference within the system. Therein, the preamble overhead is large in IR-UWB of IEEE 802.15.4a/z. Hence, aggregating as much sensor data as possible within each frame is more efficient. In this paper, we assume that reading out data from sensors and sending data to actuators is periodical and that their respective phases can be adjusted. Therefore, this paper proposes an integer linear programming-based scheduling algorithm that minimizes the number of transmitted frames by adjusting the read and write phases. Furthermore, we provide a heuristic algorithm that computes a sub-optimal but acceptable solution in a shorter time. Experimental validation shows that the data aggregation of the proposed algorithms is robust against interference.

6. On the Performance of Malleable APGAS Programs and Batch Job Schedulers

   Patrick Finnerty, Jonas Posner, Janek Bürger, Leo Takaoka, and Takuma Kanzaki

   SN Computer Science, Mar 2024

   Abs DOI Code

   Malleability—the ability for applications to dynamically adjust their resource allocations at runtime—presents great potential to enhance the efficiency and resource utilization of modern supercomputers. However, applications are rarely capable of growing and shrinking their number of nodes at runtime, and batch job schedulers provide only rudimentary support for such features. While numerous approaches have been proposed to enable application malleability, these typically focus on iterative computations and require complex code modifications. This amplifies the challenges for programmers, who already wrestle with the complexity of traditional MPI inter-node programming. Asynchronous Many-Task (AMT) programming presents a promising alternative. In AMT, computations are split into many fine-grained tasks, which are processed by workers. This makes transparent task relocation via the AMT runtime system possible, thus offering great potential for enabling efficient malleability. In this work, we propose an extension to an existing AMT system, namely APGAS for Java. We provide easy-to-use malleability programming abstractions, requiring only minor application code additions from programmers. Runtime adjustments, such as process initialization and termination, are automatically managed by our malleability extension. We validate our malleability extension by adapting a load balancing library handling multiple benchmarks. We show that both shrinking and growing operations cost low execution time overhead. In addition, we demonstrate compatibility with potential batch job schedulers by developing a prototype batch job scheduler that supports malleable jobs. Through extensive real-world job batches execution on up to 32 nodes, involving rigid, moldable, and malleable programs, we evaluate the impact of deploying malleable APGAS applications on supercomputers. Exploiting scheduling algorithms, such as FCFS, Backfilling, Easy-Backfilling, and one exploiting malleable jobs, the experimental results highlight a significant improvement regarding several metrics for malleable jobs. We show a 13.09% makespan reduction (the time needed to schedule and execute all jobs), a 19.86% increase in node utilization, and a 3.61% decrease in job turnaround time (the time a job takes from its submission to completion) when using 100% malleable job in combination with our prototype batch job scheduler compared to the best-performing scheduling algorithm with 100% rigid jobs.

7. BLE-based Outdoor Localization with Two-ray Ground-reflection Model using Optimization Algorithms

   Yu Sun, Patrick Finnerty, and Chikara Ohta

   IEEE Access, Mar 2024

   Abs DOI

   In outdoor positioning, the global positioning system (GPS) is currently the most commonly used method. Considering the power consumption required to use GPS, it may not be efficient for tracking in smaller areas, such as outdoor grazing areas in Japan, and using wireless sensor networks seems more feasible. There are several methods for BLE-based positioning. Because the angles of arrival (AoA) and time of flight (ToF) require additional equipment, the RSSI-based localization method is the most cost-efficient. Owing to the outdoor environment, the RSSI transmission model follows a two-ray ground-reflection model, this can lead to large errors in the trilateration positioning method. On the other hand, fingerprint positioning requires the creation and maintenance of a large database. This paper proposes an RSSI-based positioning method formulated as an optimization problem. We evaluated the performance of various algorithms in two application scenarios. Our simulation results show that a high localization accuracy can be obtained using this localization method. In contrast to other methods, this approach does not necessitate the use of supplementary equipment, as is the case with AoA and ToF, nor does it require the establishment and upkeep of an RSSI fingerprint database.

8. HAC: Hierarchical Agglomerative Clustering With Linear Programming for Wireless Sensor Networks

   Sidi Mohamed Mohi Dine, Patrick Finnerty, and Chikara Ohta

   IEEE Access, Jan 2024

   Abs DOI

   Wireless sensor networks are a collection of tiny sensor devices powered by energy-scarce batteries. Depending on the nature of the application, those tiny devices might be deployed in areas that make it difficult to recharge or change their battery. Therefore, it is important to optimize the energy consumption of these devices to prolong their lifetime for as long as possible. Clustering is regarded as the most efficient approach to extend the lifetime of wireless sensor networks. This study proposes a novel method integrating cluster head selection and cluster formation to obtain as long a network lifetime as is possible. We employ a linear programming model to resolve the issue of cluster head selection and a Hierarchical Agglomerative Clustering (HAC) method to form clusters in the network. We evaluated HAC on different sets of inputs where we varied the position of the sink from the corner of the network to the center. HAC outperformed other existing clustering techniques for the first node death (FND) and half node death (HND) lifetime definitions in the cases where the sink is at the corner of the network. In the cases where the sink is located at the center of the network, HAC outperformed the other existing techniques for all definitions of network lifetime.

2023

1. Automatically balancing relocatable distributed collections

   Patrick Finnerty, Tomio Kamada, and Chikara Ohta

   Concurrency and Computation: Practice and Experience, Apr 2023

   Abs DOI Code

   In previous work, we introduced a distributed collections library for the APGAS for Java programming model. This library makes it possible for programmers to develop complex distributed programs thanks to the many abstractions and computation patterns supported. In particular, programmers can fully and dynamically change the distribution of data entries through high-level abstractions. However, the problem of balancing the load between processes remains, especially in cases where multiple processes may be concurrently executing on a single host, or when the performance of the hosts used differs. To address this issue and to relieve the burden of programming a load balancing strategy for a specific application, we created a dynamic load balancer integrated into our library. This load balancer operates within a specific context in a manner which does not interfere with the program legibility. Internally, we implement a scheme inspired by the lifeline-based global load balancer scheme first introduced in X10. We evaluate the performance of our integrated load balancer on a small-scale Beowulf cluster.

2022

1. Topic-based distributed publish-process-subscribe system with metrics on geographic distance and permissible delay

   Ryota Matsumoto, Tomio Kamada, Patrick Finnerty, and Chikara Ohta

   IEICE Communications Express, Dec 2022

   Abs DOI

   As IoT devices spread, real-time data analysis and notification are increasingly important. To achieve low latency notifications in edge environments, we have proposed a distributed publish-process-subscribe system that allows multiple edge servers to handle a single topic. In our system, topic allocation to edge servers is formalized as a notification delay optimization problem. In this article, we introduce new metrics that consider the geographic distance of respective pub/sub pairs, or the permissible delay defined by each topic. Using these metrics succeeded in reducing the notification delay to neighboring subscribers and prioritizing topic allocation to meet the delay limit in simulation.

2. Pedestrian information collation through image-based person identification for pedestrian information sharing using V2V communications

   Naoko Enami, Saai Uemura, Koh Asano, Patrick Finnerty, Tomio Kamada, and Chikara Ohta

   IEICE Communications Express, Oct 2022

   Abs DOI

   The purpose of this research is to share pedestrian information among multiple vehicles by vehicle-to-vehicle communication and detect the positions of all pedestrians around the vehicle. In the past, we have proposed a method of collating shared pedestrian information through image-based person Re-Identification (ReID). In this paper, in order to improve the accuracy of pedestrian information collation, the state-of-the-art ReID network is adapted to our scenario by fine-tuning. In addition, a new pedestrian image dataset for sharing pedestrian information was constructed. In our experiments using our dataset, the proposed method improved the pedestrian information collation accuracy.

3. Supercharging the APGAS Programming Model with Relocatable Distributed Collections

   Patrick Finnerty, Yoshiki Kawanishi, Tomio Kamada, and Chikara Ohta

   Scientific Programming, Sep 2022

   Abs DOI Code

   In this article, we present our relocatable distributed collection library. Building on top of the AGPAS for Java library, we provide a number of useful intranode parallel patterns as well as the features necessary to support the distributed nature of the computation through clearly identified methods. In particular, the transfer of distributed collections’ entries between processes is supported via an integrated relocation system. This enables dynamic load-balancing capabilities, making it possible for programs to adapt to uneven or evolving cluster performance. The system we developed makes it possible to dynamically control the distribution and the data flow of distributed programs through high-level abstractions. Programmers using our library can, therefore, write complex distributed programs combining computation and communication phases through a consistent API. We evaluate the performance of our library against two programs taken from well-known Java benchmark suites, demonstrating superior programmability and obtaining better performance on one benchmark and reasonable overhead on the second. Finally, we demonstrate the ease and benefits of load balancing and a more complex application, which uses the various features of our library extensively.

4. A study on optimization of polling scheduling for in-vehicle UWB wireless networks

   Chikara Ohta, Taiyo Tanaka, Hajime Migita, Shuji Yamaguchi, Makoto Takenaka, Patrick Finnerty, and Tomio Kamada

   IEICE Communications Express, Jul 2022

   Abs DOI

   This paper proposes a polling-scheduling for UWB-based in-vehicle networks. The proposed scheduling aims to reduce the preamble overhead by aggregating periodic data readout of the in-vehicle sensors. Experimental results show that the proposed scheduling successfully suppresses data loss rate compared to the cyclic scheduling.

2021

1. Radio resource allocation based on adaptive and maximum reuse distance for LTE-V2X sidelink mode 3

   Daigo Yasuda, Patrick Finnerty, Tomio Kamada, and Chikara Ohta

   IEICE Communications Express, Oct 2021

   Abs DOI

   LTE-V2X is one of the promising wireless technologies for Vehicle to Everything (V2X), which is expected to enhance the safety of road traffic. In this paper, we propose a radio resource allocation scheme for LTE-V2X Sidelink Mode 3. The reliability of packet transmission is seriously affected by changes in vehicle density. To cope with this issue, our new scheme reuses radio resources efficiently by calculating the range of protection from mutual interference based on the vehicle density. Compared with existing schemes, the proposed scheme successfully maintains a lower error rate of packet transmission regardless of the vehicle density.

2. A self-adjusting task granularity mechanism for the Java lifeline-based global load balancer library on many-core clusters

   Patrick Finnerty, Tomio Kamada, and Chikara Ohta

   Concurrency and Computation: Practice and Experience, Feb 2021

   Abs DOI Code

   Global load balancer libraries should be easy to use and allow users to easily obtain good performance for their applications on a variety of distributed systems. In this article, we introduce a new tuning mechanism to our Java implementation of the lifeline-based global load balancer which automatically adjusts the task granularity to reach good performance based on some selected runtime metrics. We evaluate our system against four backtrack-search problems on both a many-core supercomputer environment and on a beowulf server, achieving ideal performance with our tuning mechanism on the supercomputer. We also identify the limits of our mechanism in handling situations with reduced imbalance.

Conference Articles

2025

1. Efficient QUBO Formulation for Non-Orthogonal Multiple Access

   Patrick Finnerty, Hiroto Nakano, Chikara Ohta, and Fumio Ishizaki

   In 2025 IEEE International Conference on Consumer Electronics (ICCE), Jan 2025

   Abs DOI

   To support the rapid proliferation of IoT devices, non-uniform multiple access (NOMA) admission control appears as a promising technology. Indeed, through spectrum reuse, it allows more devices to communicate than the traditional orthogonal frequency-division multiple access scheme of 4G. Previous research has shown that the NOMA allocation problem can be formulated into a maximum independent set problem, which is NP-hard. Recent advances in quadratic unconstrained binary optimization (QUBO) solvers show promise in their capability to solve this kind of allocation problems efficiently. We explore various QUBO formulations to determine which is best-suited for the high-frequency requirements of the NOMA allocation problem.

2024

1. Evaluating the A3RSRP Handover to Determine Optimal Performance in HetNet using Stochastic Petri Net

   Zhiyi Zhu, Junjun Zheng, Eiji Takimoto, Patrick Finnerty, and Chikara Ohta

   In 2024 6th International Conference on System Reliability and Safety Engineering (SRSE), Oct 2024

   Abs DOI

   Handover is necessary for improving user equipment (UE) communication quality among different cells in mobile communication networks. However, the handover decision algorithm (HDA) parameters are often difficult to choose. The use of an inappropriate parameter, such as a handover margin (HOM), causes handover failure (HoF) or frequent handover (FHO), especially in heterogeneous networks (HetNet). Therefore, properly determining a handover parameter is crucial for avoiding the HoF and FHO, as well as improving network performance. In this paper, we propose a stochastic Petri Net (SPN) to model the behavior of a UE in a HetNet. We also evaluate and predict the impact on HoF and FHO by different values of HOM in the steady-state, where the FHO is considered as handover trigger probability (HoT). Finally, the HOM value for optimal handover performance is determined using the absolute difference between the HoF and HoT through numerical experiments. The optimal handover performance was obtained at a HOM value of 6.0 dB using the proposal SPN.

2. Poster: Project Wagomu: Elastic HPC Resource Management

   Jonas Posner, and Patrick Finnerty

   In ISC High Performance 2024, May 2024

   Abs

   Traditionally, hardware resource management in supercomputers has been static: jobs request a predetermined set of resources and retain them for the duration of their execution. Elastic resource management introduces dynamic control over resources, significantly improving the efficiency of supercomputers. This approach requires support for both the resource manager and the program to ensure seamless interaction. As a result, running jobs can adjust their resource allocation by either integrating or releasing resources initiated by either party. Project Wagomu envisions a novel, user-friendly, holistic approach that addresses both sides, including their interaction. On the program side, we develop resource elasticity techniques within a prototypical Asynchronous Many-Task (AMT) runtime system that operate automatically and transparently, i.e., no user code changes are required. AMT decomposes problems into small, independent execution units, or tasks, which are dynamically assigned to processors. Tasks can generate new tasks at runtime, accommodate irregular computation patterns, and access global data. AMT applications inherently adapt to resource variations initiated by the resource manager, autonomously initiate their own resource adjustments, and can be structured into distinct phases to adapt to fluctuating resource availability. On the resource management side, we develop novel elastic job scheduling algorithms. After evaluation through simulation, these algorithms will be integrated into a prototype job scheduler. This scheduler is designed to take into account various factors, including non-computational resources such as I/O bandwidth. It will enable elastic job scheduling, facilitate the initiation and management of resource adjustments, and competently handle resource adjustments initiated by jobs.

3. Poster: Multi-Agent Transformer for Dynamic Allocation of Distributed Applications on the Edge

   Ryota Matsumoto, Patrick Finnerty, Chikara Ohta, and Tomio Kamada

   In 2024 IEEE Vehicular Networking Conference (VNC), May 2024

   Abs DOI

   Edge computing has to potential to deliver vital lowlatency services for ITS. However, due to the limited resources available on the edge, carefully allocating such services and adapting to changes in client distribution is necessary. In this poster, we show that a MAT reinforcement learning approach can perform better than heuristics.

4. Enhancing Supercomputer Performance with Malleable Job Scheduling Strategies

   Jonas Posner, Fabian Hupfeld, and Patrick Finnerty

   In Euro-Par 2023: Parallel Processing Workshops, Apr 2024

   Abs DOI

   In recent years, supercomputers have experienced significant advancements in performance and have grown in size, now comprising several thousands nodes. To unlock the full potential of these machines, efficient resource management and job scheduling—assigning parallel programs to nodes—are crucial. Traditional job scheduling approaches employ rigid jobs that use the same set of resources throughout their lifetime, resulting in significant resource under-utilization.
   By employing malleable jobs that are capable of changing their number of resources during execution, the performance of supercomputers has potential to increase. However, designing algorithms for scheduling malleable jobs is challenging since it requires complex strategies to determine when and how to reassign resources among jobs while maintaining fairness.
   In this work, we extend a recently proposed malleable job scheduling algorithm by introducing new strategies. Specifically, we propose three priority orders to determine which malleable job to consider for resource reassignments and the number of nodes when starting a job. Additionally, we propose three reassignment approaches to handle the delay between scheduling decisions and the actual transfer of resources between jobs. This results in nine algorithm variants.
   We then evaluate the impact of deploying malleable jobs scheduled by our nine algorithm variants. For that, we simulate the scheduling of job sets containing varying proportions of rigid and malleable jobs on a hypothetical supercomputer. The results demonstrate significant improvements across several metrics. For instance, with 20% of malleable jobs, the overall completion time is reduced by 11% while maintaining high node utilization and fairness.

5. Malleable APGAS Programs and Their Support in Batch Job Schedulers

   Patrick Finnerty, Leo Takaoka, Takuma Kanzaki, and Jonas Posner

   In Euro-Par 2023: Parallel Processing Workshops, Apr 2024

   Abs DOI

   Malleability—the ability for applications to dynamically adjust their resource allocations at runtime—presents great potential to enhance the efficiency and resource utilization of modern supercomputers. However, applications are rarely capable of growing and shrinking their number of nodes at runtime, and batch job schedulers provide only rudimentary support for these features. While numerous approaches have been proposed for enabling application malleability, these typically focus on iterative computations and require complex code modifications. This amplifies the challenges for programmers, who already wrestle with the complexity of traditional MPI inter-node programming. Asynchronous Many-Task (AMT) programming presents a promising alternative. Computations are split into many fine-grained tasks, which are processed by workers. This way, AMT enables transparent task relocation via the runtime system, thus offering great potential for efficient malleability. In this paper, we propose an extension to an existing AMT system, namely APGAS for Java, that provides easy-to-use malleability. More specifically, programmers enable application malleability with only minimal code additions, thanks to the simple abstractions we provide. Runtime adjustments, such as process initialization and termination, are automatically managed. We demonstrate the ease of integration between our extension and future batch job schedulers through the implementation of a simplistic malleable batch job scheduler. Additionally, we validate our extension through the adaption of a load balancing library handling multiple benchmarks. Finally, we show that even a simplistic scheduling strategy for malleable applications improves resource utilization, job throughput, and overall job response time.

2023

1. Interference Suppression Arbitration Control for In-vehicle Ultra Wideband Wireless Networks

   Makoto Okuhara, Nobuyuki Kurioka, Patrick Finnerty, and Chikara Ohta

   In 2023 IEEE 9th World Forum on Internet of Things (WF-IoT), Oct 2023

   Abs DOI

   This paper proposes an interference suppression arbitration control (ISAC) method for Ultra Wideband(UWB)-based vehicular networks. The method suppresses the effects of interference by a preceding frame by activating the reception state at the optimum timing before receiving the desired frame, reducing the data loss ratio. In addition, it utilizes the preamble code’s reception advantage property to prioritize critical communication equipment. The experimental results show that the method reduces the data loss rate to a 10^−5 magnitude. If this UWB-based ISAC method can be put to practical use in wireless vehicle wiring harnesses, it can be expected to reduce the weight of the wiring harness by approximately 20 %.

2. Distributed Cell Set: A Library for Space-Dependent Communication/Computation Overlap on Manycore Cluster

   Yoshiki Kawanishi, Patrick Finnerty, Tomio Kamada, and Chikara Ohta

   In Proceedings of the 14th International Workshop on Programming Models and Applications for Multicores and Manycores, Feb 2023

   Abs DOI

   The increase in the number of cores available in modern processors makes it important for implementations to maximize their use within a node by overlapping communication and computation. However, when the dependencies between communication and computation are complex and evolve over the course of the execution, their implementation becomes tedious and may lead to bugs. In this paper, we focus on spatial simulation where elements move over time due to interactions with other surrounding elements. We propose a Distributed Cell Set Library that manages the distribution of the simulation space using unit areas (cell units). We make it possible to manage the granularity with which cell communication may overlap with computation, and for multithreaded computation. In addition, we make it possible to describe the relationships between inter-node communication and the pending computation in cell units. For evaluation, we introduce the implementation of a two-dimensional molecular dynamics simulation using our library. We show that using our computation and communication overlapping method, the delays to reach global synchronization that are necessary in the original implementation can be avoided.

2022

1. Optimization of Polling-based MAC Schedule Considering Data Aggregation for In-vehicle UWB Wireless Networks

   Hajime Migita, Taiyo Tanaka, Shuji Yamaguchi, Makoto Takenaka, Patrick Finnerty, Tomio Kamada, and Chikara Ohta

   In 2022 IEEE 8th World Forum on Internet of Things (WF-IoT), Yokohama, Japan, Oct 2022

   Abs DOI

   This paper proposes an optimization method to obtain a polling schedule for UWB-based in-vehicle networks. The proposed method aims to reduce the preamble overhead by aggregating periodic data readout of the in-vehicle sensors. For this sake and we formulate scheduling as an integer linear programming problem. We then describe a periodic data phase optimization method that maximizes channel usage efficiency. In order to reduce the computation time and we also propose a method to compute the polling schedule by splitting the integer linear programming problem into smaller problems. Experimental results show that the proposed scheduling successfully suppresses the data loss rate compared with the round-robin scheduling.

2. Integrating a global load balancer to an APGAS distributed collections library

   Patrick Finnerty, Tomio Kamada, and Chikara Ohta

   In Proceedings of the Thirteenth International Workshop on Programming Models and Applications for Multicores and Manycores, Seoul and Republic of Korea, Apr 2022

   Abs DOI PDF

   In this paper and we introduce the global load balancer integrated into our distributed collections library for the APGAS for Java programming model. Inspired by the lifeline-based Global Load Balancer scheme and our load balancer makes it possible for programmers to perform actions on the objects recorded into a distributed collection while allowing the library to relocate entries between hosts if load imbalances occur. The programming model we adopt introduces minimal impact on the legibility of programs and with the regions of the program where entries of a distributed collection may be relocated clearly identified. Internally and we implement a hybrid scheme which balances the load both between the threads on a host and between the hosts participating in the computation. This allows for multiple concurrent computations on multiple collections with individual termination detection. We evaluate the performance of our integrated Global Load Balancer and its ability to handle various situations on a many-core supercomputer.

2020

1. Self-adjusting task granularity for Global load balancer library on clusters of many-core processors

   Patrick Finnerty, Tomio Kamada, and Chikara Ohta

   In Proceedings of the Eleventh International Workshop on Programming Models and Applications for Multicores and Manycores, San Diego, California, Feb 2020

   Abs DOI Code

   Achieving load balance is a challenge for irregular applications. Balancing runtimes and libraries aim at relieving the programmer from this difficult task by proposing a layer of abstraction between the computation at hand and the hardware used to actually perform it. With the rise of many-core architectures, load balancers for distributed computation are now expected to handle unbalance both between and within compute nodes. We bear a particular interest in backtrack search algorithms where branches of the exploration tree can be processed in parallel. Profile-based load balancers cannot be applied to them as the computation need only be performed once. It is therefore vital to promptly detect and address load unbalances from the start.A popular solution consists in using the fork/join model. Another approach on which we focus in this article is implemented in the load balancing library of X10. Both of these schemes are sensitive to task granularity. Fined-grained tasks may incur significant overhead, especially with the increased number of parallel workers of many-core architectures, while coarse-grained tasks may cause starvation. Determining a suitable grain size for a particular application is difficult.In this article we present our Java implementation of the X10 load balancing scheme. We extend it with a tuning mechanism which dynamically adjusts the task granularity, increasing the library’s versatility and usability. We demonstrate the capabilities of our tuning mechanism to suit many-core clusters using four applications: Unbalanced Tree Search, Pentomino, the Traveling Salesman Problem, and N-Queens. We achieve near identical performance to the ideal fixed grain executions on the first three but suffer from a more significant performance gap on the N-Queens problem. We advance hypotheses as to why.

Miscellaneous

2025

1. A study on improving handover performance based on handover direction in a HetNet

   Zhiyi Zhu, Shoma Suzuki, Eiji Takimoto, Patrick Finnerty, and Chikara Ohta

   Mar 2025

   2025年電⼦情報通信学会総合大会, 無線通信システムA (B-5A-45)
2. Hyper-ubiquitous Architecture: Portable Performance Models for Heterogeneous Hardware

   Patrick Finnerty, Arthur-Mustafa Farwati, and Chikara Ohta

   Mar 2025

   2025年電⼦情報通信学会総合大会, ネットワークシステム (B-6-14)

2024

1. Portable Performance Models

   Patrick Finnerty, and Chikara Ohta

   Dec 2024

   電子情報通信学会 関西支部若手研究発表会
2. Indoor Localization using Router-to-Router RSSI and Transfer Learning for Dynamic Environments

   Liuyi Yang, Patrick Finnerty, and Chikara Ohta

   Dec 2024

   電子情報通信学会 関西支部若手研究発表会
3. 電子健康情報のための階層型IDベース暗号とブロックチェーンを用いたアクセス制御に関する一考察

   Takuto Miyata, Patrick Finnerty, and Chikara Ohta

   Jun 2024

   IEICE Tech. Rep., vol. 124, no. 64, NS2024-30, pp. 11-16, June 2024.
4. Ethereum を用いた属性ベースアクセス制御のポリシー検索効率改善

   黒木 翔太, フィネルティ パトリック, and 太田 能

   Mar 2024

   2024年電子情報通信学会総合大会, ネットワークシステム (B-6-90)
5. 偏在するモバイルクライアントを想定した地域データ共有エッジアプリケーションの開発環境

   鎌田 十三郎, 山下 優真, 藤澤 和広, 釆野 翔太, フィネルティ パトリック, and 太田 能

   Mar 2024

   2024年電子情報通信学会総合大会, ネットワークシステム (B-6-30)
6. 死角にいる歩行者の情報共有を目的とした歩行者マップに関する一検討

   福田 亘剛, 榎並 直子, 鎌田 十三郎, フィネルティ パトリック, and 太田 能

   Mar 2024

   2024年電子情報通信学会総合大会, ITS (A-13-05)

2023

1. 非直交多元接続無線システムにおけるイジングマシンを利用したアドミッションコントロール最適化の一考察

   中野 碩人, 太田 能, フィネルティ パトリック, and 石崎 文雄

   Dec 2023

   IEICE Tech. Rep., vol. 123, no. 307, NS2023-157, pp. 181-186, Dec. 2023.
2. Society5.0/Beyond 5G 次代の「優しい社会」に向けた MEC に夜個人データ流通および分散基盤技術

   菅谷 みどり, Patrick Finnerty, 松本 和人, 横谷 哲也, 金井 謙治, 中里 秀則, and 太田 能

   Oct 2023

   IEICE Tech. Rep., vol. 123, no. 198, NS2023-107, pp. 177-180, Oct. 2023.
3. 低遅延エッジアプリケーション開発のためのコンフィギュアラブルテストベッド

   釆野 翔太, フィネルティ パトリック, 鎌田 十三郎, and 太田 能

   Sep 2023

   電⼦情報通信学会 2023 ソサイエティ大会
4. 分散Publish-Process-Subscribeシステムにおけるアニーリングマシンを用いたトピック割り当て最適化

   南方 亮真, フィネルティ パトリック, 太田 能, and 石崎 文雄

   Jun 2023

   電子情報通信学会技術研究報告 ネットワークシステム研究会, 高松, vol. 123, no. 68, NS2023-22
5. OpenFlowにおける伝送遅延と負荷分散を考慮したコントローラ割当手法

   周 松毅, 平石 翔, フィネルティ パトリック, 太田 能, and 石崎 文雄

   Jun 2023

   電子情報通信学会技術研究報告 ネットワークシステム研究会, 高松, vol. 123, no. 68, NS2023-20
6. 車載ハーネスの軽量化を実現する有線/無線連携通信技術の研究開発

   右田 創, 竹中 誠, フィネルティ パトリック, 太田 能, 奥原 誠, 栗岡 伸行, 清水 聡, and 臼井 誠

   May 2023

   IEICE Tech. Rep., vol. 123, no. 51, IN2023-6, pp. 29-33, May 2023.

2022

1. A study on cluster optimization in wireless sensor networks based on linear programming

   Sidi Mohamed Mohi Dine, Patrick Finnerty, Tomio Kamada, and Chikara Ohta

   May 2022

   第66回 システム制御情報学会研究発表講演会
2. 車載UWB通信のためのポーリングスケジューリング最適化と実装評価

   太田 能, 田中 太陽, 右田 創, 山口 修史, 竹中 誠, フィネルティ パトリック, and 鎌田 十三郎

   May 2022

   第66回 システム制御情報学会研究発表講演会
3. 車載 UWB 通信のためのポーリングスケジューリング最適化

   右田 創, 田中 太陽,  創右田, 山口 修史, 竹中 誠, フィネルティ パトリック, 鎌田 十三郎, and 太田 能

   Mar 2022

   2022年電子情報通信学会通信総合大会
4. エッジ環境向け分散 publish-process-subscribe システムにおける近接通知の低遅延化に関する一検討

   松本 凌太, 鎌田 十三郎, フィネルティ パトリック, and 太田 能

   Mar 2022

   2022年電子情報通信学会通信総合大会
5. 車載 UWB 通信のためのポーリングスケジューリング実装評価

   田中 太陽, 右田 創, 山口 修史, 竹中 誠, フィネルティ パトリック, 鎌田 十三郎, and 太田 能

   Mar 2022

   2022年電子情報通信学会通信総合大会

2021

1. NOMAパケットスケジューリングにおける量子アニーリング適用に関する一検討

   中田 渚生, フィネルティ パトリック, 鎌田 十三郎, 太田 能, and 石崎 文雄

   Sep 2021

   電気情報通信学会 2021年ソサイエティ大会予稿集, B-15-4 通信講演論文
2. LTE-V2X Sidelink のための無線リソース割当に関する一検討

   安田 大悟, フィネルティ パトリック, 鎌田 十三郎, and 太田 能

   Sep 2021

   電気情報通信学会 2021年ソサイエティ大会予稿集, B-15-3 通信講演論文
3. エッジコンピューティング環境向け分散データストアのための継続的クエリの実現手法

   町田 凌, 鎌田 十三郎, フィネルティ パトリック, and 太田 能

   Sep 2021

   電気情報通信学会 2021年ソサイエティ大会予稿集, B-7-10 通信講演論文
4. Experience in testing MPI-Java parallel and distributed programs with JUnit

   Patrick Finnerty, Yoshiki Kawanishi, Tomio Kamada, and Chikara Ohta

   Jul 2021

   SWoPP 2021 第135回Pro研究発表会

   Code

Theses

2022

1. High-Productivity Abstractions and their Efficient Runtime for Dynamically Load-Balanced Distributed Programs on Multi/Many-core Clusters

   Patrick Finnerty

   Jul 2022

   HTML PDF