Kazuhiko Komatsu

Assistant Professor
Cyberscience Center, Tohoku University, Japan
Research Interest: high-performance computing, computational fluid dynamics, physically-based rendering

History

Apr. 1998 - Mar. 2002     Bachelor Degree (Mechanical Engineering), Tohoku University, Japan.
Apr. 2002 - Mar. 2004     Master Degree (Information Sciences), Tohoku University, Japan.
Apr. 2004 - Mar. 2008     Doctor Degree/Ph.D. (Information Sciences), Tohoku University, Japan.
Aug. 2004 - Aug. 2004     Overseas Student, Chonbuk National University, Korea.
Apr. 2005 - Mar. 2006     Overseas Education, Central Michigan University, MI, USA.
Apr. 2008 - Mar. 2012     Post-doctoral fellow at Cyberscience Center, Tohoku University, Japan.
Oct. 2010 - Dec. 2010     Visiting Researcher, HLRS, Stuttgart University, Germany.
Apr. 2012 - Present         Assistant Professor, Cyberscience Center, Tohoku University, Japan.

Publications

Invited Talks

[5]	Kazuhiko Komatsu, Ryusuke Egawa, Hiroyuki Takizawa, and Hiroaki Kobayashi. Toward High Performance-Portabilities on Modern HPC Systems. To appear in 16th Workshop on Sustained Simulation Performance, Dec. 2012. [ bib ]
[4]	Kazuhiko Komatsu, Ryusuke Egawa, Hiroyuki Takizawa, and Hiroaki Kobayashi. Performance of Practical Applications on Modern Supercomputing Systems. In SC12 NEC booth presentation, Nov 2012. [ bib ]
[3]	Kazuhiko Komatsu, Takashi Soga, Ryusuke Egawa, Hiroyuki Takizawa, and Hiroaki Kobayashi. Performance Evaluation of a CFD using Cartesian Meshes on Various Supercomputing Systems. In NUG XXIV, June 2012. [ bib ]
[2]	Kazuhiko Komatsu, Takashi Soga, Ryusuke Egawa, Hiroyuki Takizawa, and Hiroaki Kobayashi. Performance evaluation of a next-generation CFD on various supercomputing systems. In Teraflop workshop 14, December 2011. [ bib ]
[1]	Kazuhiko Komatsu, Kentaro Koyama, Katsuto Sato, Hiroyuki Takizawa, and Hiroaki Kobayashi. A High-level Programming Framework for Efficient Hybrid-architecture Computing. In 14th SIAM Conference on Parallel Processing for Scientific Computing Minisymposium, February 2010. [ bib ]

Reviewed Papers

[28]	Kazuhiko Komatsu, Takashi Soga, Ryusuke Egawa, Hiroyuki Takizawa, and Hiroaki Kobayashi. Performance evaluation of a next-generation CFD on various supercomputing systems. In Sustained Simulation Performance 2012, Springer Berlin Heidelberg, pages 123-132, 2012. [ bib ]
[27]	Alfian Amrizal, Shoichi Hirasawa, Kazuhiko Komatsu, Hiroyuki Takizawa, Hiroaki Kobayashi. Improving the Scalability of Transparent Checkpointing for GPU Computing Systems. In TENCON2012, Nov. 2012. [ bib ]
[26]	Kazuhiko Komatsu, Takashi Soga, Ryusuke Egawa, Hiroyuki Takizawa, Hiroaki Kobayashi, Shun Takahashi, Daisuke Sasaki, and Kazuhiro Nakahashi. Performance Evaluation of BCM on Various Supercomputing Systems. In 24th International Conference on Parallel Computational Fluid Dynamics, 2012. [ bib ]
[25]	佐藤功人, 小松一彦, 滝沢寛之, 小林広明. OpenCLにおけるタスク並列化支援のための実行時依存関係解析手法. In IPSJ Transactions on Advanced Computing Systems Vol.5 No.1 (ACS36), pages 53-67, 2011. [ bib ]
[24]	Jiali Yao, Kazuhiko Komatsu, Ryusuke Egawa, Hiroyuki Takizawa, Hiroaki Kobayashi. A Patch-Based Bit Mask Filtering Method for Micropolygon Rasterization. In High-Performance Graphics(HPG) Poster, Aug. 2011. [ bib ]
[23]	Katsuto Sato, Kazuhiko Komatsu, Hiroyuki Takizawa, and Hiroaki Kobayashi. A History-based Performance Prediction Model with Profile Data Classification for Automatic Task Allocation in Heterogeneous Computing Systems. In the 9th IEEE International Symposium on Parallel and Distributed Processing with Applications, May 2011. [ bib ]
[22]	小山賢太郎, 佐藤功人, 小松一彦, 村田善智, 滝沢寛之, 小林広明. マイグレーションによる複合型計算システム向けジョブスケジューリング. In IPSJ Transactions on Advanced Computing Systems Vol.4 No.4 (ACS35), pages 203-213, Oct. 2011.
[21]	小山賢太郎, 佐藤功人, 小松一彦, 村田善智, 滝沢寛之, 小林広明. マイグレーションによる複合型計算システム向けジョブスケジューリング. In SACSIS2011 先進的計算基盤システムシンポジウム, May 2011. [ bib ]
[20]	Kazuhiko Komatsu, Hiroyuki Takizawa, and Hiroaki Kobayashi. Hierarchical Parallel Processing of Ray-Frustum Intersection Tests on a Cell Cluster. To be published in Scientific Programming Special issue on IWSV.
[19]	Hiroyuki Takizawa, Kentaro Koyama, Katsuto Sato, Kazuhiko Komatsu, and Hiroaki Kobayashi. CheCL:Transparent Checkpointing and Process Migration of OpenCL Applications Platform. In the 25th IEEE International Parallel & Distributed Processing Symposium, May 2011. [ bib ]
[18]	Kazuhiko Komatsu, Takashi Soga, Ryusuke Egawa, Hiroyuki Takizawa, Hiroaki Kobayashi, Shun Takahashi, Daisuke Sasaki, and Kazuhiro Nakahashi. Parallel Processing of the Building-Cube Method on the GPU Platform. In Computers & Fluids Special Issue "22nd International Conference on Parallel Computational Fluid Dynamics," Vol. 45, No. 1, pp. 122-128, June 2011. [ bib ]
[17]	Takashi Soga, Akihiro Musa, Koki Okabe, Kazuhiko Komatsu, Ryusuke Egawa, Hiroyuki Takizawa, Hiroaki Kobayashi, Shun Takahashi, Daisuke Sasaki, and Kazuhiro Nakahashi. Performance of SOR Methods on Modern Vector and Scalar Processors. In Computers & Fluids Special Issue "22nd International Conference on Parallel Computational Fluid Dynamics", Vol. 45, No. 1, pp. 215-221, June 2011. [ bib ]
[16]	Kazuhiko Komatsu, Katsuto Sato, Yusuke Arai, Kentaro Koyama, Hiroyuki Takizawa, Hiroaki Kobayashi. Evaluating Performance and Portability of OpenCL Programs. In the Fifth International Workshop on Automatic Performance Tuning, June 2010. [ bib ]
[15]	Kazuhiko Komatsu, Takashi Soga, Ryusuke Egawa, Hiroyuki Takizawa, Hiroaki Kobayashi, Shun Takahashi, Daisuke Sasaki, and Kazuhiro Nakahashi. Efficient Data Management for the Building Cube Method using Cartesian Meshes on the GPU Platform. In ISC'10 Poster (International Supercomputing Conference 2010), June 2010. [ bib ]
[14]	Kazuhiko Komatsu, Takashi Soga, Ryusuke Egawa, Hiroyuki Takizawa, Hiroaki Kobayashi, Shun Takahashi, Daisuke Sasaki, and Kazuhiro Nakahashi. Parallel Processing of the Building-Cube Method on the GPU Platform. In 22nd International Conference on Parallel Computational Fluid Dynamics, May 2010. [ bib ]
[13]	Takashi Soga, Akihiro Musa, Koki Okabe, Kazuhiko Komatsu, Ryusuke Egawa, Hiroyuki Takizawa, Hiroaki Kobayashi, Shun Takahashi, Daisuke Sasaki, and Kazuhiro Nakahashi. Performance of SOR Methods on Vector Processor SX-9. In 22nd International Conference on Parallel Computational Fluid Dynamics, May 2010. [ bib ]
[12]	Katsuto Sato, Hiroyuki Takizawa, Kazuhiko Komatsu, and Hiroaki Kobayashi. Automatic Tuning of CUDA Execution Parameters for Stencil Processing. In Software Automatic Tuning: From Concepts to State-of-the-Art Results, Springer, pp. 209-228, September 2010. [ bib ]
[11]	Ken-ichi Suzuki, Yoshiyuki Kaeriyama, Kazuhiko Komatsu, Ryusuke Egawa, Nobuaki Ohba, and Hiroaki Kobayashi. A Fast Ray-Tracing using Bounding Spheres and Frustum Rays for Dynamic Scene Rendering. In IEICE Transactions on Information and Systems, Vol. E93-D, No. 4, pp. 891-902, 2010. [ bib ]
[10]	Hiroyuki Takizawa, Katuto Sato, Kazuhiko Komatsu, and Hiroaki Kobayashi. CheCUDA: A Checkpoint/Restart Tool for CUDA Applications. International Conference on Parallel and Distributed Computing, Applications and Technologies, 2009. [ bib ]
[9]	Kazuhiko Komatsu, Hiroyuki Takizawa, and Hiroaki Kobayashi. Hierarchical Parallel Processing of Ray Tracing on a Cell Cluster. In 1st International Workshop on Super Visualization (IWSV08), CD-ROM, 2008. [ bib ]
[8]	Kazuhiko Komatsu, Yoshiyuki Kaeriyama, Kenichi Suzuki, Hiroyuki Takizawa, and Hiroaki Kobayashi. An Efficient Intersection Algorithm Design of Ray Tracing For Many-Core Graphics Processors. In Computer Graphics and Imaging (CGIM2008), pages 315-320, 2008. [ bib ]
[7]	Kazuhiko Komatsu, Yoshiyuki Kaeriyama, Kenichi Suzuki, Hiroyuki Takizawa, and Hiroaki Kobayashi. A Fast Ray Frustum-Triangle Intersection Algorithm with Precomputation and Early Termination. In IPSJ Transactions on Advanced Computing Systems Vol.1 No.1 (ACS22), pages 85-95, June 2008. [ bib ]
[6]	Kazuhiko Komatsu, Yoshiyuki Kaeriyama, Kenichi Suzuki, Hiroyuki Takizawa, and Hiroaki Kobayashi. A Fast Ray Frustum-Triangle Intersection Algorithm with Precomputation and Early Termination. In High Performance Computing Symposium (HPCS2008), pages 81-88, 2008. [ bib ]
[5]	Kazuhiko Komatsu, Yoshiyuki Kaeriyama, Kenichi Suzuki, Hiroaki Kobayashi, and Tadao Nakamura. Intersection Algorithm for Ray Tracing Using Packets. In FIT2007: Forum on Information Technology Letters, pages 265-268, 2007. [ bib ]
[4]	Kazuhiko Komatsu, Yoshiyuki Kaeriyama, Daichi Zaitsu, Kenichi Suzuki, Nobuyuki Ohba, and Tadao Nakamura. Packet-Primitive Intersection Method. In RT '06: Poster Compendium of IEEE Symposium on Interactive Ray Tracing, page 6, 2006. [ bib ]
[3]	Yoshiyuki Kaeriyama, Daichi Zaitsu, Kazuhiko Komatsu, Kenichi Suzuki, Nobuyuki Ohba, and Tadao Nakamura. Ray tracing hardware system using plane-sphere intersections. In FPL '06: Proceedings of International Conference on Field Programmable Logic and Applications, pages 315-320, 2006. [ bib ]
[2]	Yoshiyuki Kaeriyama, Daichi Zaitsu, Kazuhiko Komatsu, Kenichi Suzuki, Nobuyuki Ohba, and Tadao Nakamura. Hardware for a Ray Tracing Technique Using Plane-Sphere Intersections. In EGPGV '06: Symposium Short Papers Proceedings of Eurographics Symposium on Parallel Graphics and Visualization, pages 9-12, 2006. [ bib ]
[1]	Yoshiyuki Kaeriyama, Daichi Zaitsu, Kazuhiko Komatsu, Kenichi Suzuki, Nobuyuki Ohba, and Tadao Nakamura. Programmable Graphics Hardware for Image Synthesis Using the Global Illumination Model. In COOL Chips IX: Proceedings of International Symposium on Low-Power and High-Speed Chips, pages 183-185, 2006. [ bib ]

Technical Papers

[26]	菅原誠, 平澤将一, 小松一彦, 滝沢寛之, 小林広明.. ナノ粒子群形成アプリケーションのOpenACCによる実装と性能評価． To appear in 第26回数値流体力学シンポジウムCFD2012, Dec. 2012. [ bib ]
[25]	小松一彦, 曽我隆, 江川隆輔, 滝沢寛之, 小林広明. 大規模計算システムにおけるBuilding Cube Methodの性能評価． To appear in 第26回数値流体力学シンポジウムCFD2012, Dec. 2012. [ bib ]
[24]	安田一平, 小松一彦, 江川隆輔, 小林広明. 大規模並列システムのノード間通信を考慮した性能モデルに関する一検討． To appear in 第194回ARC・第137回HPC合同研究発表会(HOKKE-20), Dec. 2012. [ bib ]
[23]	菅原誠, 平澤将一, 小松一彦, 滝沢寛之, 小林広明. ナノ粒子群形成アプリケーションのOpenACCによる実装と性能評価． In 第136回HPC研究会, Oct. 2012. [ bib ]
[22]	小松一彦, 江川隆輔, 安田一平, 撫佐昭裕, 松岡浩司, 小林広明. HPCアプリケーションの性能可搬性に関する一検討． In 第136回HPC研究会, Oct. 2012. [ bib ]
[21]	小松一彦, 江川隆輔, 安田一平, 撫佐昭裕, 松岡浩司, 小林広明. HPCシステムにおける最適化手法の性能可搬性に関する一検討． In 第7回次世代CFD研究会, Sep. 2012. [ bib ]
[20]	小松一彦, 曽我隆, 江川隆輔, 滝沢寛之, 小林広明. 大規模計算システムにおけるBCMの性能評価． In 東北大学サイバーサイエンスセンター大規模科学計算システム広報 SENAC, ISSN 0286-7419, Vol.45, No.3, July 2012. [ bib ]
[19]	小松一彦, 滝沢寛之, 小林広明. Build Cube Methodによる非圧縮性流体解析のマルチプラットフォーム評価. In 第6回次世代CFD研究会, March 2012. [ bib ]
[18]	Muhammad Alfian Amrizal, Katsuto Sato, Kazuhiko Komatsu, Hiroyuki Takizawa, and Hiroaki Kobayashi. Evaluation of a scalable checkpointing mechanism for heterogeneous computing systems. In 第7回情報処理学会東北支部研究会 March 2012. [ bib ]
[17]	明珎陽介, 小松一彦, 江川隆輔, 滝沢寛之, 小林広明. 直交格子を用いた大規模流体シミュレーションコードの通信最適化に関する検討. In 第7回情報処理学会東北支部研究会, March 2012. [ bib ]
[16]	滝沢寛之, 佐藤功人, 小松一彦, 小林広明. OpenCLアプリケーションの実行時自動チューニング(Runtime Automatic Performance Tuning of OpenCL Applications). To be appeared in 第17回計算工学講演会, 2012. [ bib ]
[15]	Kazuhiko Komatsu, Takashi Soga, Ryusuke Egawa, Hiroyuki Takizawa, Hiroaki Kobayashi, Shun Takahashi, Daisuke Sasaki, and Kazuhiro Nakahashi. Performance of Building Cube Method on Various Platforms. In International Conference on Flow Dynamics 2011, November 2011. [ bib ]
[14]	Katsuto Sato, Kazuhiko Komatsu, Hiroyuki Takizawa, Hiroaki Kobayashi. An Automatic Task Assignment Method for Heterogeneous Computing Systems. In International Conference on Flow Dynamics 2011 Poster, November 2011. [ bib ]
[13]	菅原誠, 佐藤功人, 小松一彦, 滝沢寛之, 小林広明. プログラム自動生成技術に基づくGPUコンピューティングの性能評価. In SWoPP2011, July 2011. [ bib ]
[12]	佐藤功人, 小松一彦, 滝沢寛之, 小林広明. 複合型計算機におけるソフトウェア開発の課題と支援手法の検討. In 日本学術振興会インターネット技術第163委員会情報流通基盤分科会(ITRC/INI)「情報流通基盤分科会ワークショップ」・「先端的ネットワーク＆コンピューティングテクノロジワークショップ」合同ワークショップ, March 2011. [ bib ]
[11]	小松一彦, 滝沢寛之, 小林広明. GPUクラスタによるBuilding Cube Methodの性能評価. In 第4回次世代CFD研究会, February 2011. [ bib ]
[10]	小山賢太郎, 佐藤功人, 小松一彦, 村田善智, 滝沢寛之, 小林広明. 複合型計算システムのためのジョブスケジューリングの検討. In 第9回情報シナジー研究会, February 2011. [ bib ]
[9]	Katsuto Sato, Kazuhiko Komatsu, Hiroyuki Takizawa, and Hiroaki Kobayashi. A Runtime Task Reallocation Library for Heterogeneous Computational Environments. In International Conference on Flow Dynamics 2010 Poster, November 2010. [ bib ]
[8]	小松一彦, 小山賢太郎, 佐藤功人, 滝沢寛之, 小林広明. ハイブリッド型計算環境のためのプログラミングフレームワークSPRAT(A High-level Programming Framework for Efficient Hybrid-architecture Computing). In 日本学術振興会インターネット技術第163委員会情報流通基盤分科会(ITRC/INI)「情報流通基盤分科会ワークショップ」・「先端的ネットワーク＆コンピューティングテクノロジワークショップ」合同ワークショップ, March 2010. [ bib ]
[7]	滝沢寛之, 佐藤功人, 小松一彦, 小林広明. CUDAアプリケーション向けチェックポイント・リスタート機能の実装と評価(Implementation and Evaluation of a Checkpoint/Restart Tool for CUDA Applications). In IPSJ SIG Technical Report, October 2009. [ bib ]
[6]	小松一彦, 小山賢太郎, 佐藤功人, 滝沢寛之, 小林広明. プロセッサ自動選択機能を有するBLAS の実現に向けた性能評価(Performance Evaluation towards BLAS with Automatic Processor Selection). In FIT2009: 第8回情報科学技術フォーラム, September 2009. [ bib ]
[5]	Chainan Satayapiwat, Kazuhiko Komatsu, Ryusuke Egawa, Hiroyuki Takizawa, Hiroaki Kobayashi. An Auction-based Resource Allocation Considering Multifaceted Utilities in a Peer-to-Peer Environment. In FIT2009: Forum on Information Technology, September 2009. [ bib ]
[4]	小山賢太郎, 佐藤功人, 小松一彦, 滝沢寛之, 小林広明. GPU向け線形代数ライブラリの性能評価(Early evaluation of linear algebra libraries for GPU computing). In 計算工学講演会論文集第14巻第1号(Proceedings of the Conference on Computational Engineering and Science, Volume 14, No.1), pages 289-292, 2009年5月. [ bib ]
[3]	Yoshitomo Murata, Kazuhiko Komatsu, Yuki Ishimori, Hiroyuki Takizawa, Hiroaki Kobayashi. A Large-scale Distributed Data-Mining System using Idle Time of Game Consoles. In 日本学術振興会インターネット技術第163委員会情報流通基盤分科会(ITRC/INI)「情報流通基盤分科会ワークショップ」・「先端的ネットワーク＆コンピューティングテクノロジワークショップ」合同ワークショップ, February 2009. [ bib ]
[2]	小松一彦, 佐野健太郎, 鈴木健一, 中村維男. 汎用グラフィックスアクセラレータ(GPU)を用いたボリュームレンダリングの将来展望. In 脳神経情報処理研究会, September 2004. [ bib ]
[1]	小松一彦, 佐野健太郎, 鈴木健一, 中村維男. ボリュームデータセグメンテーションの高速処理方式. In 情報処理学会東北支部研究会, pages 43-46, November 2003. [ bib ]