2024/04/27 更新

写真a

ホシノ テツヤ
星野 哲也
HOSHINO Tetsuya
所属
情報基盤センター 大規模計算支援環境研究部門 准教授
大学院担当
大学院情報学研究科
職名
准教授

研究分野 1

  1. 情報通信 / 高性能計算

経歴 2

  1. 名古屋大学   情報基盤センター   准教授

    2023年1月 - 現在

  2. 東京大学   情報基盤センター   助教

    2016年1月 - 2022年12月

 

論文 19

  1. Optimize Efficiency of Utilizing Systems by Dynamic Core Binding

    Kawai M., Ida A., Hanawa T., Hoshino T.

    ACM International Conference Proceeding Series     頁: 77 - 82   2024年

     詳細を見る

    掲載種別:研究論文(国際会議プロシーディングス)   出版者・発行元:ACM International Conference Proceeding Series  

    Load balancing at both the process and thread levels is imperative for minimizing application computation time in the context of MPI/OpenMP hybrid parallelization. This necessity arises from the constraint that, within a typical hybrid parallel environment, an identical number of cores is bound to each process. Dynamic Core Binding, however, adjusts the core binding based on the process’s workload, thereby realizing load balancing at the core level. In prior research, we have implemented the DCB library, which has two policies for computation time reduction or power reduction. In this paper, we show that the two policies provided by the DCB library can be used together to achieve both computation time reduction and power consumption reduction.

    DOI: 10.1145/3636480.3637221

    Scopus

    その他リンク: https://dblp.uni-trier.de/db/conf/hpcasia/hpcasia2024w.html#KawaiIHH24

  2. FMOプログラムABINIT-MPの整備状況2023

    望月 祐志, 中野 達也, 坂倉 耕太, 奥脇 弘次, 土居 英男, 加藤 季広, 滝沢 寛之, 成瀬 彰, 大島 聡史, 星野 哲也, 片桐 孝洋

      23 巻 ( 1 ) 頁: 4 - 8   2024年

     詳細を見る

    記述言語:日本語   出版者・発行元:日本コンピュータ化学会  

    <p>In August 2023, we released the latest version of our ABINIT-MP program, Open Version 2 Revision 8. In this version, the most commonly used FMO-MP2 calculations are even faster than in the previous Revision 4. It is now also possible to calculate excitation and ionization energies for regions of interest. Improved interaction analysis is also available. In addition, we have started GPU-oriented modifications. In this preliminary report, we present the current status of ABINIT-MP.</p>

    DOI: 10.2477/jccj.2024-0001

    CiNii Research

  3. Implementation of Radio Wave Propagation using RT Cores and Consideration of Programming Models

    Hashinoki, S; Ohshima, S; Katagiri, T; Nagai, T; Hoshino, T

    2023 IEEE INTERNATIONAL PARALLEL AND DISTRIBUTED PROCESSING SYMPOSIUM WORKSHOPS, IPDPSW     頁: 673 - 681   2023年

     詳細を見る

    掲載種別:研究論文(国際会議プロシーディングス)   出版者・発行元:2023 IEEE International Parallel and Distributed Processing Symposium Workshops, IPDPSW 2023  

    With the NVIDIA Turing architecture generation, several NVIDIA graphics processing units (GPUs) have introduced ray tracing acceleration hardware (RT cores). Ray tracing processing can be regarded as a simulation of wave and particle propagation, collision, and reflection. Therefore, it is expected to be applied to computational science and high-performance computing. However, few studies have been conducted using RT cores. The purpose of this research is to demonstrate the use of RT cores in the scientific and technical computing fields. We implemented a radio wave propagation loss calculation with the programmable ray tracing application framework OptiX and evaluated its performance. Furthermore, we investigated the challenges of reducing the description of framework-specific settings and the needs of hardware allocation. In the simple two spheres experiment, the RT core implementation showed the highest performance. Moreover, the acceleration was super linear scaling, between (10000, 5000) and (20000, 10000). In the experiment with a sphere and planes, the performance achieved by the RT cores was up to approximately 390 times higher than the parallel execution of the BVH search algorithm. We also proved that a large number of RT cores yielded higher performance. In the open data problem space experiment, we evaluated various GPUs and revealed that a larger number of RT cores is effective. These results show that RT cores are sufficiently effective for radio propagation calculations with an adequate number of ray projections. Through this research, we contributed to the RT core use in computational science by proposing an implementation method for ray tracing applications and revealing the effects of RT cores in radio wave propagation loss calculations.

    DOI: 10.1109/IPDPSW59300.2023.00115

    Web of Science

    Scopus

    その他リンク: https://dblp.uni-trier.de/db/conf/ipps/ipdps2023w.html#HashinokiOKNH23

  4. Auto-Tuning Mixed-precision Computation by Specifying Multiple Regions

    Ren X., Kawai M., Hoshino T., Katagiri T., Nagai T.

    Proceedings - 2023 11th International Symposium on Computing and Networking, CANDAR 2023     頁: 175 - 181   2023年

     詳細を見る

    出版者・発行元:Proceedings - 2023 11th International Symposium on Computing and Networking, CANDAR 2023  

    Mixed-precision computation is a promising method for substantially increasing the speed of numerical computations. However, using mixed-precision data is a double-edged sword. Although it can improve the computational performance, the reduction in precision brings more uncertainties and errors. It is necessary to determine which variables can be represented with a lower-precision format without affecting the accuracy of the results. Hence, much effort is spent on selecting appropriate variables while considering the execution time and numerical accuracy. Auto-Tuning (AT) is one of several technologies that can assist in eliminating this intensive work. In this study, we investigated an AT strategy for the 'Blocks' directive in the auto-Tuning language ppOpen-AT to tune multiple regions of a program and evaluated the effectiveness. A benchmark program of the nonhydrostatic icosahedral atmospheric model (NICAM), which is a global cloud resolving model, was considered as a study case. Experimental results indicated that when a single part of the program could perform well in the mixed-precision computation, a combination achieved a better performance. When used on the Flow Type I Subsystem (The Fujitsu PRIMEHPC FX1000), this method achieved almost 1.27× speedup compared with the NICAM benchmark program using all double-precision data.

    DOI: 10.1109/CANDAR60563.2023.00031

    Scopus

  5. Large-scale earthquake sequence simulations on 3D nonplanar faults using the boundary element method accelerated by lattice H-matrices

    So Ozawa, Akihiro Ida, Tetsuya Hoshino, Ryosuke Ando

    Geophysical Journal International     2022年10月

     詳細を見る

    掲載種別:研究論文(学術雑誌)   出版者・発行元:Oxford University Press (OUP)  

    Summary

    Large-scale earthquake sequence simulations using the boundary element method (BEM) incur extreme computational costs through multiplying a dense matrix with a slip rate vector. Hierarchical matrices (H-matrices) have often been used to accelerate this multiplication. However, the complexity of the structures of the H-matrices and the communication costs between processors limit their scalability, and they therefore cannot be used efficiently in distributed memory computer systems. Lattice H-matrices have recently been proposed as a tool to improve the parallel scalability of H-matrices. In this study, we developed a method for earthquake sequence simulations applicable to 3D nonplanar faults with lattice H-matrices. We present a simulation example and verify the mesh convergence of our method for a 3D nonplanar thrust fault using rectangular and triangular discretizations. We also performed performance and scalability analyses of our code. Our simulations, using over ${10^5}$ degrees of freedom, demonstrated a parallel acceleration beyond ${10^4}$ MPI processors and a &amp;gt; 10-fold acceleration over the best performance when the normal H-matrices are used. Using this code, we can perform unprecedented large-scale earthquake sequence simulations on geometrically complex faults with supercomputers. The software is made an open-source and freely available.

    DOI: 10.1093/gji/ggac386

    arXiv

  6. Optimizations of H-matrix-vector Multiplication for Modern Multi-core Processors.

    Tetsuya Hoshino, Akihiro Ida, Toshihiro Hanawa

    CLUSTER     頁: 462 - 472   2022年

     詳細を見る

    掲載種別:研究論文(国際会議プロシーディングス)   出版者・発行元:IEEE  

    DOI: 10.1109/CLUSTER51413.2022.00056

    その他リンク: https://dblp.uni-trier.de/db/conf/cluster/cluster2022.html#HoshinoIH22

  7. Fortran標準規格do concurrentを用いたGPUオフローディング手法の評価

    星野 哲也, 塙 敏博

    情報処理学会研究報告(Web)   2022-HPC-183 巻   頁: 1 - 8   2022年

     詳細を見る

  8. A64FXにおける階層型行列演算の性能評価

    星野 哲也, 伊田 明弘, 塙 敏博

    情報処理学会研究報告(Web)   2021-HPC-180 巻   頁: 1 - 8   2021年

     詳細を見る

  9. Large-scale earthquake sequence simulations of 3D geometrically complex faults using the boundary element method accelerated by lattice H-matrices on distributed memory computer systems

    伊田 明弘, 星野 哲也

    arXiv preprint   - 巻   頁: 1 - 26   2021年

     詳細を見る

  10. Preliminary development of training environment for deep learning on supercomputer system 査読有り

    Y. Nomura, I. Sato, T. Hanawa, S. Hanaoka, T. Nakao, T. Takenaga, D. Sato, T. Hoshino, Y. Sekiya, S. Ohshima, N. Hayashi, O. Abe

    International Journal of Computer Assisted Radiology and Surgery   13 巻 ( Issue 1 supplement ) 頁: S105 - S106   2018年6月

     詳細を見る

    掲載種別:研究論文(国際会議プロシーディングス)  

    DOI: 10.1007/s11548-018-1766-y

  11. 有限要素法における係数行列生成部のマルチコア・メニィコア向け最適化

    中島研吾, 中島研吾, 星野哲也, 星野哲也, 成瀬彰, 塙敏博, 三木洋平

    情報処理学会研究報告(Web)   2018 巻 ( HPC-163 ) 頁: Vol.2018‐HPC‐163,No.28,1‐8 (WEB ONLY)   2018年2月

     詳細を見る

    記述言語:日本語  

    J-GLOBAL

  12. Load-Balancing-Aware Parallel Algorithms of H-Matrices with Adaptive Cross Approximation for GPUs. 査読有り

    Tetsuya Hoshino, Akihiro Ida, Toshihiro Hanawa, Kengo Nakajima

    IEEE International Conference on Cluster Computing, CLUSTER 2018, Belfast, UK, September 10-13, 2018     頁: 35 - 45   2018年

     詳細を見る

    掲載種別:研究論文(国際会議プロシーディングス)   出版者・発行元:IEEE Computer Society  

    DOI: 10.1109/CLUSTER.2018.00016

  13. Design of Parallel BEM Analysis Framework for SIMD Processors 査読有り

    星野 哲也

    International Conference on Computational Science   10860 巻   頁: 601 - 613   2018年

     詳細を見る

    記述言語:英語   掲載種別:研究論文(学術雑誌)  

    DOI: 10.1007/978-3-319-93698-7_46

    Scopus

  14. スーパーコンピュータ上でのDeep Learning学習環境の初期構築

    野村行弘, 佐藤一誠, 佐藤一誠, 佐藤一誠, 塙敏博, 花岡昇平, 中尾貴祐, 竹永智美, 佐藤大介, 星野哲也, 関谷勇司, 大島聡史, 林直人, 阿部修

    電子情報通信学会技術研究報告   117 巻 ( 281(MI2017 47-62) ) 頁: 1‐2   2017年10月

     詳細を見る

    記述言語:日本語  

    J-GLOBAL

  15. Pascal vs KNL: Performance Evaluation with ICCG Solve 査読有り

    Tetsuya Hoshino, Satoshi Ohshima, Toshihiro Hanawa, Kengo Nakaima, Akihiro Ida

    HPC in Asia Workshop Poster Session, ISC High Performance 2017     2017年6月

     詳細を見る

    記述言語:英語   掲載種別:研究論文(国際会議プロシーディングス)  

  16. OpenACCを用いたICCG法ソルバーのPascal GPUにおける性能評価

    星野哲也, 大島聡史, 塙敏博, 中島研吾, 伊田明宏

    情報処理学会研究報告(Web)   2017 巻 ( HPC-158 ) 頁: Vol.2017‐HPC‐158,No.18,1‐9 (WEB ONLY) - 9   2017年3月

     詳細を見る

    記述言語:日本語   掲載種別:研究論文(学術雑誌)  

    J-GLOBAL

  17. A Directive-based Data Layout Abstraction for Performance Portability of OpenACC Applications 査読有り

    Tetsuya Hoshino, Naoya Maruyama, Satoshi Matsuoka

    PROCEEDINGS OF 2016 IEEE 18TH INTERNATIONAL CONFERENCE ON HIGH PERFORMANCE COMPUTING AND COMMUNICATIONS; IEEE 14TH INTERNATIONAL CONFERENCE ON SMART CITY; IEEE 2ND INTERNATIONAL CONFERENCE ON DATA SCIENCE AND SYSTEMS (HPCC/SMARTCITY/DSS)     頁: 1147 - 1154   2016年

     詳細を見る

    記述言語:英語   掲載種別:研究論文(国際会議プロシーディングス)   出版者・発行元:IEEE  

    Directive-based programming interfaces such as OpenACC and OpenMP are becoming more prevalent in application development targeting accelerators, in particular when porting existing CPU-only code. Unlike vendor-specific alternatives such as CUDA, they are designed to be portable across different accelerators, and therefore once necessary directives are added to an existing CPU-only code, it can be executed on different accelerator architectures depending on the availability of supporting compilers. However, it does not automatically mean that such code runs efficiently on different architectures, and in fact, architecture-specific coding such as choosing optimal data layouts is almost mandatory for optimal performance, imposing a significant burden if implemented manually. Towards realizing performance portability in accelerator programming, we propose a set of extended directives that allow the programmer to optimize data layouts for a given accelerator without modifying original program code. Unlike the manual approach, the code change is confined in the directives with the original code kept as it is. This paper evaluates the effectiveness of our proposed extensions in the OpenACC standard by extending UPACS and CCS-QCD OpenACC applications. A prototype source-to-source translator for the extensions achieves 123% and 120% of the baseline performance, respectively, which are comparable to manually tuned versions.

    DOI: 10.1109/HPCC-SmartCity-DSS.2016.34

    Web of Science

  18. An OpenACC extension for data layout transformation 査読有り

    Tetsuya Hoshino, Naoya Maruyama, Satoshi Matsuoka

    Proceedings of WACCPD 2014: 1st Workshop on Accelerator Programming Using Directives - Held in Conjunction with SC 2014: The International Conference for High Performance Computing, Networking, Storage and Analysis     頁: 12 - 18   2015年4月

     詳細を見る

    記述言語:英語   掲載種別:研究論文(国際会議プロシーディングス)   出版者・発行元:Institute of Electrical and Electronics Engineers Inc.  

    OpenACC is gaining momentum as an implicit and portable interface in porting legacy CPU-based applications to heterogeneous, highly parallel computational environment involving many-core accelerators such as GPUs and Intel Xeon Phi. OpenACC provides a set of loop directives similar to OpenMP for the parallelization and also to manage data movement, attaining functional portability across different heterogeneous devices
    however, the performance portability of OpenACC is said to be insufficient due to the characteristics of different target devices, especially those regarding memory layouts, as automated attempts by the compilers to adapt is currently difficult. We are currently working to propose a set of directives to allow compilers to have better semantic information for adaptation
    here, we particularly focus on data layout such as Structure of Arrays, advantageous data structure for GPUs, as opposed to Array of Structures, which exhibits good performance on CPUs. We propose a directive extension to OpenACC that allows the users to flexibility specify optimal layouts, even if the data structures are nested. Performance results show that we gain as much as 96 % in performance for CPUs and 165% for GPUs compared to programs without such directives, essentially attaining both functional and performance portability in OpenACC.

    DOI: 10.1109/WACCPD.2014.12

    Scopus

  19. CUDA vs OpenACC: Performance Case Studies with Kernel Benchmarks and a Memory-Bound CFD Application11 査読有り

    Tetsuya Hoshino, Naoya Maruyama, Satoshi Matsuoka, Ryoji Takaki

    PROCEEDINGS OF THE 2013 13TH IEEE/ACM INTERNATIONAL SYMPOSIUM ON CLUSTER, CLOUD AND GRID COMPUTING (CCGRID 2013)     頁: 136 - 143   2013年

     詳細を見る

    記述言語:英語   掲載種別:研究論文(国際会議プロシーディングス)   出版者・発行元:IEEE  

    OpenACC is a new accelerator programming interface that provides a set of OpenMP-like loop directives for the programming of accelerators in an implicit and portable way. It allows the programmer to express the offloading of data and computations to accelerators, such that the porting process for legacy CPU-based applications can be significantly simplified. This paper focuses on the performance aspects of OpenACC using two microbenchmarks and one real-world computational fluid dynamics application. Both evaluations show that in general OpenACC performance is approximately 50% lower than CUDA. However, for some applications it can reach up to 98% with careful manual optimizations. The results also indicate several limitations of the OpenACC specification that hamper full use of the GPU hardware resources, resulting in a significant performance gap when compared to a highly tuned CUDA code. The lack of a programming interface for the shared memory in particular results in as much as three times lower performance.

    DOI: 10.1109/CCGrid.2013.12

    Web of Science

▼全件表示

MISC 28

  1. CPU・GPU並列プログラミング入門(1)—Introduction to Parallel Programming on CPU and GPU(1)

    中島 研吾, 星野 哲也  

    シミュレーション = Journal of the Japan Society for Simulation Technology / 日本シミュレーション学会 編42 巻 ( 2 ) 頁: 103 - 109   2023年6月

     詳細を見る

    記述言語:日本語   出版者・発行元:小宮山印刷工業  

    CiNii Books

  2. 数値計算ライブラリの自動チューニングにおけるXAI適用の試み—An Adaptation of XAI to Auto-tuning for Numerical Calculation Library

    青木 将太, 片桐 孝洋, 大島 聡史, 永井 亨, 星野 哲也  

    計算工学講演会論文集 = Proceedings of the Conference on Computational Engineering and Science / 日本計算工学会 編28 巻   頁: 904 - 907   2023年5月

     詳細を見る

    記述言語:日本語   出版者・発行元:日本計算工学会  

  3. Fortran標準規格do concurrentを用いたGPUオフローディング手法の評価

    星野 哲也, 塙 敏博  

    情報処理学会研究報告(Web)2022-HPC-183 巻   頁: 1 - 8   2022年

  4. A64FXにおけるテンポラルブロッキングの実装と性能評価

    星野 哲也, 塙 敏博  

    研究報告ハイパフォーマンスコンピューティング(HPC)2021-HPC-178 巻 ( 17 ) 頁: 1 - 8   2021年3月

     詳細を見る

    担当区分:筆頭著者  

  5. 「計算・データ・学習」融合スーパーコンピュータシステム「Wisteria/BDEC-01」の概要

    中島研吾, 塙敏博, 下川辺隆史, 伊田明弘, 芝隼人, 三木洋平, 星野哲也, 有間英志, 河合直聡, 坂本龍一, 近藤正章, 岩下武史, 八代尚, 長尾大道, 松葉浩也, 荻田武史, 片桐孝洋, 古村孝志, 鶴岡弘, 市村強, 藤田航平  

    情報処理学会研究報告(Web)2021 巻 ( HPC-179 )   2021年

     詳細を見る

  6. 「計算・データ・学習」融合スーパーコンピュータシステムWisteria/BDEC-01の性能評価

    塙敏博, 中島研吾, 中島研吾, 下川辺隆史, 芝隼人, 三木洋平, 星野哲也, 河合直聡, 似鳥啓吾, 今村俊幸, 工藤周平, 中尾昌広  

    情報処理学会研究報告(Web)2021 巻 ( HPC-180 )   2021年

     詳細を見る

  7. A64FXにおける階層型行列演算の性能評価

    星野哲也, 伊田明弘, 伊田明弘, 塙敏博  

    情報処理学会研究報告(Web)2021 巻 ( HPC-180 ) 頁: 1 - 8   2021年

     詳細を見る

  8. Large-scale earthquake sequence simulations of 3D geometrically complex faults using the boundary element method accelerated by lattice H-matrices on distributed memory computer systems

    伊田 明弘, 星野 哲也  

    arXiv preprint- 巻   頁: 1 - 26   2021年

  9. An Optimization of H-matrix-vector Multiplication by Using Un-used Cores

    Tetsuya Hoshino, Toshihiro Hanawa, Akihiro Ida  

    HPC Asia 2020   2020年1月

  10. Numerical Linear Algebra Based on Lattice H-Matrices

    伊田 明弘, Ichitaro Yamazaki, Rio Yokota, Satoshi Ohshima, Tasuku Hiraishi, Takeshi Iwashita, Tetsuya Hoshino, Toshihiro Hanawa  

    HPC Asia   2020年1月

  11. メニーコアクラスタにおける階層型行列法の高速化に向けた性能評価

    星野哲也, 伊田明弘  

    計算工学講演会論文集(CD-ROM)24 巻   頁: ROMBUNNO.C‐07‐02   2019年6月

     詳細を見る

    記述言語:日本語   出版者・発行元:日本計算工学会  

    J-GLOBAL

  12. メニーコアプロセッサにおける高性能計算のための高レベル抽象化

    星野 哲也, Hoshino Tetsuya  

        2018年9月

     詳細を見る

    記述言語:英語  

  13. OpenCLを用いたFPGAによる階層型行列計算

    塙敏博, 伊田明弘, 星野哲也  

    情報処理学会研究報告(Web)2018 巻 ( HPC-163 ) 頁: Vol.2018‐HPC‐163,No.26,1‐8 (WEB ONLY)   2018年2月

     詳細を見る

    記述言語:日本語  

    J-GLOBAL

  14. 階層型行列計算のFPGAへの適用

    塙敏博, 伊田明弘, 星野哲也  

    情報処理学会研究報告(Web)2017 巻 ( HPC-161 ) 頁: Vol.2017‐HPC‐161,No.10,1‐10 (WEB ONLY)   2017年9月

     詳細を見る

    記述言語:日本語  

    J-GLOBAL

  15. 階層型行列法ライブラリHACApKを用いたアプリケーションのメニーコア向け最適化

    星野哲也, 伊田明弘, 塙敏博, 中島研吾  

    情報処理学会研究報告(Web)2017 巻 ( HPC-160 ) 頁: Vol.2017‐HPC‐160,No.15,1‐10 (WEB ONLY) - 10   2017年7月

     詳細を見る

    記述言語:日本語  

    J-GLOBAL

  16. GPU搭載スーパーコンピュータReedbush‐Hの性能評価

    塙敏博, 星野哲也, 中島研吾, 大島聡史, 伊田明弘  

    情報処理学会研究報告(Web)2017 巻 ( HPC-159 ) 頁: Vol.2017‐HPC‐159,No.9,1‐6 (WEB ONLY)   2017年4月

     詳細を見る

    記述言語:日本語  

    J-GLOBAL

  17. Xeon Phi+OmniPath環境におけるOpenMP,MPI性能最適化

    塙敏博, 星野哲也, 中島研吾, 大島聡史, 伊田明弘  

    情報処理学会研究報告(Web)2017 巻 ( HPC-158 ) 頁: Vol.2017‐HPC‐158,No.21,1‐8 (WEB ONLY)   2017年3月

     詳細を見る

    記述言語:日本語  

    J-GLOBAL

  18. ICCG法ソルバーのIntel Xeon Phi向け最適化

    中島研吾, 中島研吾, 大島聡史, 大島聡史, 塙敏博, 星野哲也, 伊田明弘, 伊田明弘  

    情報処理学会研究報告(Web)2016 巻 ( HPC-157 ) 頁: Vol.2016‐HPC‐157,No.16,1‐8 (WEB ONLY)   2016年12月

     詳細を見る

    記述言語:日本語  

    J-GLOBAL

  19. パイプライン型共役勾配法の性能評価

    塙敏博, 中島研吾, 中島研吾, 大島聡史, 大島聡史, 星野哲也, 伊田明弘, 伊田明弘  

    情報処理学会研究報告(Web)2016 巻 ( HPC-157 ) 頁: Vol.2016‐HPC‐157,No.6,1‐9 (WEB ONLY)   2016年12月

     詳細を見る

    記述言語:日本語  

    J-GLOBAL

  20. データ解析・シミュレーション融合スーパーコンピュータシステムReedbush‐Uの性能評価

    塙敏博, 中島研吾, 大島聡史, 伊田明弘, 星野哲也, 田浦健次朗  

    情報処理学会研究報告(Web)2016 巻 ( HPC-156 ) 頁: Vol.2016‐HPC‐156,No.10,1‐10 (WEB ONLY) - 10   2016年9月

     詳細を見る

    記述言語:日本語  

    J-GLOBAL

  21. データレイアウト最適化指示文によるOpenACCアプリケーションの高速化

    星野 哲也  

    研究報告ハイパフォーマンスコンピューティング(HPC)2016-HPC-155 巻   頁: 1 - 8   2016年

  22. 圧縮性流体プログラムのOpenACCによる高速化

    星野 哲也  

    研究報告ハイパフォーマンスコンピューティング(HPC)2016-HPC-153 巻   頁: 1 - 10   2016年

  23. OpenACCディレクティブ拡張によるデータレイアウト最適化

    星野哲也, 丸山直也, 松岡聡  

    研究報告ハイパフォーマンスコンピューティング(HPC)2014 巻 ( 45 ) 頁: 1 - 8   2014年7月

     詳細を見る

    記述言語:日本語   出版者・発行元:一般社団法人情報処理学会  

    近年増加傾向にある GPU 等のアクセラレータを搭載した計算環境への既存プログラムの移植方法として,CUDA・OpenCL に代表されるローレベルなプログラミングモデルを用いる方法に対し,ディレクティブベースの OpenACC のようなハイレベルなプログラミングモデルを用いる方法が注目されている.このようなディレクティブベースのプログラミングモデルの利点として,元のプログラムを維持したまま移植を行えるために,デバイス間の機能的な可搬性が高いことがあげられる.しかし現状の OpenACC などの High-level なプログラミングモデルは,スカラプロセッサとメニーコアアクセラレータの得意とするデータレイアウトの相違に対応することが出来ず,異なる性質を持ったデバイス間の性能可搬性に問題がある.そこで本研究では,データレイアウトを抽象化し,異なるデバイス間での性能可搬性を向上させるための OpenACC の拡張ディレクティブを試作し,姫野ベンチマークのデータレイアウトをトランスレーターにより変更し,マルチコア CPU,Intex Xeon Phi,K20X GPU のそれぞれで評価を行った.その結果,オリジナルと同一のデータレイアウトと比較して,Intel Xeon Phi では 27%,K20X GPU では 24%の性能向上が得られることを確認した.

    CiNii Books

  24. CPU-GPUそれぞれに最適なデータレイアウトを選択可能にするOpenACCディレクティブ拡張

    星野哲也, 丸山直也, 松岡聡  

    研究報告ハイパフォーマンスコンピューティング(HPC)2014 巻 ( 5 ) 頁: 1 - 5   2014年2月

     詳細を見る

    記述言語:日本語   出版者・発行元:一般社団法人情報処理学会  

    近年増加傾向にある GPU 等のアクセラレータを搭載した計算環境への既存プログラムの移植方法として,CUDA・OpenCL に代表される Low-level なプログラミングモデルを用いる方法に対し,ディレクティブベースの OpenACC のような High-level なプログラミングモデルを用いる方法が考えられる.このようなディレクティブベースのプログラミングモデルの利点として,元のプログラムを壊さずに移植を行えるために,デバイス間の可搬性が高いことがあげられる.しかし現状の OpenACC などのプログラミングモデルは,スカラプロセッサとメニーコアアクセラレータの得意とするデータレイアウトの相違等に対応することが出来ず,異なる性質を持ったデバイス間の性能可搬性に問題がある.そこで本研究では,データレイアウトを抽象化し,異なるデバイス間での性能可搬性を向上させるための OpenACC の拡張ディレクティブを試作し,評価を行った.

    CiNii Books

  25. ディレクティブベースプログラミング言語OpenACCの性能評価

    星野哲也, 丸山直也, 松岡聡  

    ハイパフォーマンスコンピューティングと計算科学シンポジウム論文集2013 巻   頁: 91 - 91   2013年1月

     詳細を見る

    記述言語:日本語  

  26. 大規模流体アプリケーションのCUDA・OpenACCへの移植性の評価

    星野 哲也, 丸山 直也, 松岡 聡  

    研究報告ハイパフォーマンスコンピューティング(HPC)2012 巻 ( 42 ) 頁: 1 - 9   2012年7月

     詳細を見る

    記述言語:日本語  

    地震や気象予測,航空機や高層ビル設計といったシミュレーションに利用される数値流体力学アプリケーションは,近年一般的になりつつある GPU を用いたスーパーコンピュータにおいて,目覚ましい成果を上げている.しかし,GPU を用いたプログラミングは,高い性能を得ること難しいと言われており,レガシープログラムの GPU 環境への移植が問題となっている.本稿では,実際に利用されている大規模流体アプリケーションである UPACS を手動により CUDA 化し,性能と移植コストの面から評価を行った.また,プログラムの移植性を解決すると期待されている,OpenACC の予備評価を行った.これら評価の結果を示し,今後解決すべき課題について述べる.Computational fluid dynamics (CFD) applications used for an earthquake and meteorological simulation are one of the most important application executed with high-speed supercomputers. Especially, GPU-based supercomputers have been showing remarkable performance of CFD applications. However, GPU-programing is still difficult to obtain high performance, which prevents legacy applications from being ported to GPU environment. We apply classical optimizations to a real-world CFD application UPACS and evaluate it&#039;s performance and porting costs, and we also evaluate OpenACC expected to provide portability across CPUs and GPUs. We demonstrate these results of evaluation and mention performance problems should be resolved in the future.

    CiNii Books

  27. 大規模流体アプリケーションのGPUによる高速化手法の評価

    星野哲也, 丸山直也, 松岡聡  

    先進的計算基盤システムシンポジウム論文集2012 巻   頁: 73 - 74   2012年5月

     詳細を見る

    記述言語:日本語  

  28. OpenACCプログラミング

    丸山 直也, 星野 哲也  

    映像情報メディア学会誌 : 映像情報メディア = The journal of the Institute of Image Information and Television Engineers66 巻 ( 10 ) 頁: 817 - 822   2012年

     詳細を見る

    記述言語:英語   出版者・発行元:一般社団法人映像情報メディア学会  

    DOI: 10.3169/itej.66.817

    Scopus

    CiNii Books

▼全件表示

科研費 4

  1. 実アプリケーションの時空間ブロッキングによる高速化に関する研究

    研究課題/研究課題番号:22K17898  2022年4月 - 2024年3月

    日本学術振興会  科学研究費助成事業  若手研究

    星野 哲也

      詳細を見る

    担当区分:研究代表者 

    配分額:1430000円 ( 直接経費:1100000円 、 間接経費:330000円 )

    スーパーコンピュータに搭載される最新世代のCPUは大きな共有キャッシュを有し、これを効率的に利用する最適化手法として知られる時空間ブロッキングは、科学・工学分野のシミュレーションで頻出するステンシル計算を高速化するための手法である。しかし時空間ブロッキングは煩雑なプログラミングを要求するため、実アプリケーションへの応用は進んでいない。本研究課題では、比較的簡単なコード変換によって実現可能ながら、大容量の共有キャッシュメモリを用いることで高効率実行が可能なoverlapped方式の時空間ブロッキング手法に着目し、様々なCPUにおける性能モデリング及び実アプリケーションでの有効性の検証を行う。
    本研究は、特に微分方程式を解く際に頻出する計算パターンのステンシル計算を高速化する手法として知られる、時空間ブロッキングを対象として性能モデリングを構築することが目的である。時空間ブロッキングには冗長計算が必要なoverlapped方式と冗長計算を削減したtrapezoid方式などがあるが、本年度は3次元の拡散方程式プログラムに対してoverlapped(x,y次元)+trapezoid(z次元)のハイブリッドによる時空間ブロッキングの実装を行なった。実装されたプログラムをXeon CPU, A64FXなどで評価を行なっている。Xeon CPUではL1/L2キャッシュの性能から、概ね想定した通りの性能が得られている一方、A64FXではキャッシュ性能以外にも考慮すべき点がいくつかある(演算レイテンシと、それを隠すために必要なパイプライン実行の段数、パイプライン実行に必要なレジスタ数など)ことが分析によりわかってきた。さらに分析を進め、アプリケーション側のパラメータを含め、性能モデルの構築を目指している。
    また実アプリケーションとして、気象庁の天気予報シミュレーションプログラムであるasucaへの時空間ブロッキングの適用を検討している。時空間ブロッキングは、キャッシュを有効利用することにより、冗長な計算を許容しつつもメモリへの負荷を下げる手法であるため、メモリ性能を十分に引き出しいてるプログラムでなければ適用効果が見込めないため、時空間ブロッキング適用に向けたasucaの最適化を行なった。最適化の効果により1.5倍程度の高速化が得られたが、時空間ブロッキングの適用には更なる最適化が必要であることが判明した。
    2023年1月に、課題代表者が東京大学情報基盤センター助教から名古屋大学情報基盤センター准教授へと転職したことによる予算執行の遅れや、新たに受け持つことになった授業の準備などの影響で、当初計画より実装や実験の準備が遅れている。当初計画では、2022年度中に国内研究会で成果発表を行う予定であったが、成果発表が2023年度にずれ込むこととなったため、やや遅れていると評価した。
    今後の方針として、より新しいCPUを用いた性能評価が必要となる。本研究と関連して、学際大規模情報基盤共同利用・共同研究拠点の2023年度課題として「近代的メニーコアシステムにおける性能モデリング手法」に採択されたため、2023年度は京都大学に新たに設置されたHBMメモリを持つXeon CPU等での評価・分析を行い、より多様なプロセッサを用いての性能モデルの開発に取り組む。
    また実アプリケーションでの効果の検証を行うため、気象庁プログラムであるasucaの他、地震シミュレーションコード、大気海洋シミュレーションコードなどへも、それぞれの専門家との共同研究により適用を検討したい。

  2. 格子H行列に基づく数値線形代数の構築と最新アーキテクチャへの高性能実装法

    研究課題/研究課題番号:21H03447  2021年4月 - 2024年3月

    日本学術振興会  科学研究費助成事業  基盤研究(B)

    伊田 明弘, 横田 理央, 塙 敏博, 岩下 武史, 大島 聡史, 星野 哲也, 平石 拓, 横田 理央, 塙 敏博, 岩下 武史, 大島 聡史, 星野 哲也, 平石 拓

      詳細を見る

    担当区分:研究分担者 

    格子H行列を用いた数値線形代数系の構築を目的とし、固有値計算、LU分解、QR分解などを実行する新しいアルゴリズムの研究・開発を行った。多くの成果が得られた中でも、主要なものとして、BLR(Block Low-Rank)行列の固有値計算法が挙げられる。格子H行列の特殊な場合であるBLR行列について、全固有値を計算するアルゴリズムを開発した。開発したアルゴリズムの計算複雑度を、BLR行列を特徴付ける行列サイズ・ブロックサイズ・各ブロックの階数(ランク)を変数として見積り、最適な条件を検討した。最適条件下で提案アルゴリズムの計算量が従来の密行列の場合に比べて大幅に低減させられることを理論的に示した。また、数値実験を用いて、計算時間が理論通り計算量に比例すること、および、固有値と固有ベクトルの誤差が階数を増やすにつれて密行列に近づいていくことを確かめた。
    格子H行列法の適用範囲の拡大に取り組んだ。従来、格子H行列を含む低ランク構造行列法は境界要素解析(時間項を含まない空間領域の積分方程式法)への適用が想定されていた。時空間領域積分方程式法に適用範囲を拡張すべく、FDP(Fast Domain Partitioning)法と格子H行列法を組み合わせた手法を開発し、従来手法の計算量を開発手法では大幅に低減させられることを理論的に示した。さらに、3次元弾性波動伝播解析を提案手法で行う計算コードを開発し、理論に近い計算時間で解析が行えることを確かめた。
    格子H行列法の高性能実装に関する研究においても多くの成果が得られた。主要な成果としては、タスク並列言語Tascellを拡張し、分散メモリ環境における行列分割構造生成の並列実装を提案した。3次元電場解析に対する約1億要素を用いた数値実験において、最大8ノード×36ワーカーで良好な高速化を達成した。

  3. 余剰コアを活用する高性能計算・データ解析支援

    研究課題/研究課題番号:20H00580  2020年4月 - 2023年3月

    日本学術振興会  科学研究費助成事業  基盤研究(A)

    塙 敏博, 下川辺 隆史, 星野 哲也, 三木 洋平, 伊田 明弘, 下川辺 隆史, 星野 哲也, 三木 洋平, 伊田 明弘

      詳細を見る

    担当区分:研究分担者 

    ①ユーザプログラムを再コンパイルせずに、動的プロファイリングを実現する手法を検討した。SystemTapを用いて,測定したい関数部分に着目したプロファイル、OpenMP並列数の動的な変更をほぼオーバヘッドなく実現した。主計算と同時に他の処理を実行した際の影響を測定した。
    ②MPI+OpenMP並列化されたアプリケーションを効率的に動作させることを目指して、MPIプロセス毎に割り付けられるコア数(OpenMPスレッド数)をコントロールし、各コアの負荷を均一化する手法について検討した。検討手法をDCBライブラリとして実装し、アプリケーションから簡単なAPIコールのみで利用できるようにした。DCBライブラリを格子H行列法コードに適用し、15.5%の計算速度向上と8.0%の消費電力削減を実現した。
    ③In Situデータ解析の実現に向けた準備として,宇宙物理分野のアプリケーションにおいて時間進化の計算中に解析処理を実行する機能を追加した.特定条件を満たしたデータだけを高頻度に単一ファイルへと追加出力する機能も追加し,非同期ファイルIO機能の予備評価に利用可能にした。
    ④流体計算において頻出するステンシル計算は一般的にメモリ律速となることが知られており、余剰な演算器が発生しやすい演算パターンである。キャッシュメモリを利用し数ステップ分の演算をメモリに書き戻すことなく実行することで、余剰な演算器を活用する手法として知られるテンポラルブロッキングを3次元の拡散方程式カーネルに適用し、最新のプロセッサにおいて評価した結果,最大4.99倍の性能向上を実現した。
    ⑤ステンシル計算のうち局所的に高精細にできる適合格子細分化法(AMR法)では、データ構造の複雑さから通信が性能低下につながる。余剰コアを活用するという観点からテンポラルブロッキングを最新のプロセッサで効率的に利用する方法について検討を進めた。

  4. アプリケーションのデータ構造に着目したメニーコア向け自動最適化フレームワーク

    研究課題/研究課題番号:16H06679  2016年8月 - 2018年3月

    日本学術振興会  科学研究費助成事業  研究活動スタート支援

    星野 哲也

      詳細を見る

    近年増加傾向にあるメニーコアプロセッサを用いた計算環境において、その性能を引き出すためにはVector Processing Unit (VPU)を効率良く利用することが重要である。しかし、VPUの効率的な利用にはハードウェアやコンパイラに関する知識が必要であり、またプログラムのデータ構造の変更などが往々にして必要となる。
    本研究では、データ構造を抽象化するためのコンパイラ指示文の提案と、その指示文を解釈するトランスレータの開発、自動ベクトル化を促進するフレームワークデザインの提案と、そのデザインに則った境界要素法向けのフレームワークの開発を行った。

 

担当経験のある科目 (本学) 3

  1. 大規模並列数値計算特論

    2023

  2. 大規模計算特論B

    2023

  3. プログラミング2

    2023

 

社会貢献活動 1

  1. 最近のFortran向けGPUプログラミング事情(JAXA内部講習会)

    役割:講師

    2023年12月

学術貢献活動 2

  1. HPC Asia 2024 Local Arrangement Chair

    役割:企画立案・運営等

    2024年1月

     詳細を見る

    種別:学会・研究会等 

  2. xSIG 2023 プログラム委員

    役割:査読

    2023年8月

     詳細を見る

    種別:査読等