WEKO3
アイテム
スーパーコンピュータの高速化技術
https://jaxa.repo.nii.ac.jp/records/43184
https://jaxa.repo.nii.ac.jp/records/4318432a782b6-7f6e-48f2-8363-b4917f1ef0af
| 名前 / ファイル | ライセンス | アクション |
|---|---|---|
|
|
|
| Item type | テクニカルレポート / Technical Report(1) | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| 公開日 | 2015-03-26 | |||||||||
| タイトル | ||||||||||
| タイトル | スーパーコンピュータの高速化技術 | |||||||||
| 言語 | ||||||||||
| 言語 | jpn | |||||||||
| 資源タイプ | ||||||||||
| 資源タイプ識別子 | http://purl.org/coar/resource_type/c_18gh | |||||||||
| 資源タイプ | technical report | |||||||||
| その他のタイトル(英) | ||||||||||
| その他のタイトル | Techniques to Achieve High-speed Computations in Supercomputers | |||||||||
| 著者 |
渡辺, 貞
× 渡辺, 貞
× Watanabe, Tadashi
|
|||||||||
| 著者所属 | ||||||||||
| 日本電気株式会社 | ||||||||||
| 出版者 | ||||||||||
| 出版者 | 航空宇宙技術研究所 | |||||||||
| 出版者(英) | ||||||||||
| 出版者 | National Aerospace Laboratory(NAL) | |||||||||
| 書誌情報 |
航空宇宙技術研究所特別資料 en : Special Publication of National Aerospace Laboratory SP-5 巻 5, p. 19-26, 発行日 1985-11 |
|||||||||
| 抄録(英) | ||||||||||
| 内容記述タイプ | Other | |||||||||
| 内容記述 | Various techniques to achieve high-speed computations in supercomputers are described in this paper. In particular, architectural approaches to augment the concurrency are described. Those approaches include multi-processors, parallel processors and pipelined processors. A multi-processor system is one in which plural processors share the main memory and different programs are executed on each processor. The multi-processor enhances the system throughput by the concurrent execution of programs. In a parallel processor system, plural processors are connected to each other directly or through a memory unit, and all the processors execute the identical instruction simultaneously under the control of the control unit. In a pipelined processor, instruction execution is streamlined like an assembly line of an automobile factory or the processing procedure in a chemical plant. Next, vector processing for achieving high-speed computations is outlined. Finally, as an example of supercomputers, the vector processing feature of the NEC supercomputer SX system is introduced. The SX system employs multiple-parallel pipelines and 512 way-interleaved memory system to achieve a vector processing speed of 1.3 G flops. | |||||||||
| ISSN | ||||||||||
| 収録物識別子タイプ | ISSN | |||||||||
| 収録物識別子 | 0289-260X | |||||||||
| 書誌レコードID | ||||||||||
| 収録物識別子タイプ | NCID | |||||||||
| 収録物識別子 | AN10097345 | |||||||||
| 資料番号 | ||||||||||
| 内容記述タイプ | Other | |||||||||
| 内容記述 | 資料番号: NALSP0005004 | |||||||||
| レポート番号 | ||||||||||
| 内容記述タイプ | Other | |||||||||
| 内容記述 | レポート番号: NAL SP-5 | |||||||||