WEKO3
アイテム
Design and Preliminary Implementation of a Particle Simulation Machine for Efficient Short - range Interaction Computations
https://ipsj.ixsq.nii.ac.jp/records/18582
https://ipsj.ixsq.nii.ac.jp/records/18582126f229a-6ed0-489c-bc9b-33d3e15290ec
| 名前 / ファイル | ライセンス | アクション |
|---|---|---|
|
|
Copyright (c) 2003 by the Information Processing Society of Japan
|
|
| オープンアクセス | ||
| Item type | Trans(1) | |||||||
|---|---|---|---|---|---|---|---|---|
| 公開日 | 2003-05-15 | |||||||
| タイトル | ||||||||
| タイトル | Design and Preliminary Implementation of a Particle Simulation Machine for Efficient Short - range Interaction Computations | |||||||
| タイトル | ||||||||
| 言語 | en | |||||||
| タイトル | Design and Preliminary Implementation of a Particle Simulation Machine for Efficient Short - range Interaction Computations | |||||||
| 言語 | ||||||||
| 言語 | eng | |||||||
| キーワード | ||||||||
| 主題Scheme | Other | |||||||
| 主題 | 高性能アーキテクチャ | |||||||
| 資源タイプ | ||||||||
| 資源タイプ識別子 | http://purl.org/coar/resource_type/c_6501 | |||||||
| 資源タイプ | journal article | |||||||
| 著者所属 | ||||||||
| Graduate School of Information Systems University of Electro - Communications/Image Technology Laboratory Corporation | ||||||||
| 著者所属 | ||||||||
| Graduate School of Information Systems University of Electro - Communications | ||||||||
| 著者所属 | ||||||||
| Graduate School of Information Systems University of Electro - Communications | ||||||||
| 著者所属 | ||||||||
| Graduate School of Information Systems University of Electro - Communications | ||||||||
| 著者所属(英) | ||||||||
| en | ||||||||
| Graduate School of Information Systems, University of Electro - Communications/Image Technology Laboratory Corporation | ||||||||
| 著者所属(英) | ||||||||
| en | ||||||||
| Graduate School of Information Systems, University of Electro - Communications | ||||||||
| 著者所属(英) | ||||||||
| en | ||||||||
| Graduate School of Information Systems, University of Electro - Communications | ||||||||
| 著者所属(英) | ||||||||
| en | ||||||||
| Graduate School of Information Systems, University of Electro - Communications | ||||||||
| 著者名 |
Ryo, Takata
Kenji, Kise
Hiroki, Honda
Toshitsugu, Yuba
× Ryo, Takata Kenji, Kise Hiroki, Honda Toshitsugu, Yuba
|
|||||||
| 著者名(英) |
Ryo, Takata
Kenji, Kise
Hiroki, Honda
Toshitsugu, Yuba
× Ryo, Takata Kenji, Kise Hiroki, Honda Toshitsugu, Yuba
|
|||||||
| 論文抄録 | ||||||||
| 内容記述タイプ | Other | |||||||
| 内容記述 | This paper describes the architecture of a high performance particle simulation machine DEM-1 (Dentsudai EAM Machine-1) for short-range particle interaction computations. All existing particle simulation machines have specialized pipelines to calculate long-range particle interactions effectively. However their ability to perform particle simulations efficiently diminishes with short-range interactions. Communication cost in particle simulations will play a significant role in performance when the computation cost becomes O(N) where Nis the number of particles. In DEM-1 a three-dimensional torus high-speed network reduces this cost while 2048 local processors perform time integration. Specially designed cutoff units reduce the number of calculations in DEM-1. Each specialized pipeline consists of a dedicated data path supported by position vector pre-fetch dual ported memory (pre-fetch DPM). A prototype pipeline has been implemented on an FPGA test-bed. Performance of the DEM-1 pipeline has been measured and total system performance has been estimated with very large-scale Embedded Atom Method (EAM) molecular dynamics simulations. | |||||||
| 論文抄録(英) | ||||||||
| 内容記述タイプ | Other | |||||||
| 内容記述 | This paper describes the architecture of a high performance particle simulation machine,DEM-1 (Dentsudai EAM Machine-1), for short-range particle interaction computations. All existing particle simulation machines have specialized pipelines to calculate long-range particle interactions effectively. However, their ability to perform particle simulations efficiently diminishes with short-range interactions. Communication cost in particle simulations will play a significant role in performance when the computation cost becomes O(N), where Nis the number of particles. In DEM-1 a three-dimensional torus high-speed network reduces this cost while 2048 local processors perform time integration. Specially designed cutoff units reduce the number of calculations in DEM-1. Each specialized pipeline consists of a dedicated data path supported by position vector pre-fetch dual ported memory (pre-fetch DPM). A prototype pipeline has been implemented on an FPGA test-bed. Performance of the DEM-1 pipeline has been measured and total system performance has been estimated with very large-scale Embedded Atom Method (EAM) molecular dynamics simulations. | |||||||
| 書誌レコードID | ||||||||
| 収録物識別子タイプ | NCID | |||||||
| 収録物識別子 | AA11833852 | |||||||
| 書誌情報 |
情報処理学会論文誌コンピューティングシステム(ACS) 巻 44, 号 SIG06(ACS1), p. 96-112, 発行日 2003-05-15 |
|||||||
| ISSN | ||||||||
| 収録物識別子タイプ | ISSN | |||||||
| 収録物識別子 | 1882-7829 | |||||||
| 出版者 | ||||||||
| 言語 | ja | |||||||
| 出版者 | 情報処理学会 | |||||||