| Item type |
SIG Technical Reports(1) |
| 公開日 |
2019-07-17 |
| タイトル |
|
|
タイトル |
Towards Performance Portability and Modernization of FLASH via Transpilation with High-Level Intermediate Representation |
| タイトル |
|
|
言語 |
en |
|
タイトル |
Towards Performance Portability and Modernization of FLASH via Transpilation with High-Level Intermediate Representation |
| 言語 |
|
|
言語 |
eng |
| キーワード |
|
|
主題Scheme |
Other |
|
主題 |
大規模システム |
| 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 |
| 著者所属 |
|
|
|
Tokyo Institute of Technology／National Institute of Advanced Industrial Science and Technology |
| 著者所属 |
|
|
|
Argonne National Laboratory／University of Chicago |
| 著者所属 |
|
|
|
National Institute of Advanced Industrial Science and Technology |
| 著者所属 |
|
|
|
Argonne National Laboratory／University of Chicago |
| 著者所属 |
|
|
|
RIKEN Center for Computational Science／Tokyo Institute of Technology／National Institute of Advanced Industrial Science and Technology |
| 著者所属(英) |
|
|
|
en |
|
|
Tokyo Institute of Technology / National Institute of Advanced Industrial Science and Technology |
| 著者所属(英) |
|
|
|
en |
|
|
Argonne National Laboratory / University of Chicago |
| 著者所属(英) |
|
|
|
en |
|
|
National Institute of Advanced Industrial Science and Technology |
| 著者所属(英) |
|
|
|
en |
|
|
Argonne National Laboratory / University of Chicago |
| 著者所属(英) |
|
|
|
en |
|
|
RIKEN Center for Computational Science / Tokyo Institute of Technology / National Institute of Advanced Industrial Science and Technology |
| 著者名 |
Mateusz, Bysiek
Saurabh, Chawdhary
Mohamed, Wahib
Anshu, Dubey
Satoshi, Matsuoka
|
| 著者名(英) |
Mateusz, Bysiek
Saurabh, Chawdhary
Mohamed, Wahib
Anshu, Dubey
Satoshi, Matsuoka
|
| 論文抄録 |
|
|
内容記述タイプ |
Other |
|
内容記述 |
With concurrent increase in application complexity and hardware heterogeneity, large multiphysics code FLASH faces huge challenges to its continued usability on high performance computing platforms. We are building a novel transpilation framework that relies on high-level intermediate representation to confront this challenge and enable FLASH to adapt to accelerator-based architecture via performance-oriented refactoring. Additionally, we use the framework to modernize code by enabling higher level of abstraction in expressing computations. We evaluate the effectiveness of the tool with respect to speedup obtained relative to original code performance, and also quantify productivity gains. |
| 論文抄録(英) |
|
|
内容記述タイプ |
Other |
|
内容記述 |
With concurrent increase in application complexity and hardware heterogeneity, large multiphysics code FLASH faces huge challenges to its continued usability on high performance computing platforms. We are building a novel transpilation framework that relies on high-level intermediate representation to confront this challenge and enable FLASH to adapt to accelerator-based architecture via performance-oriented refactoring. Additionally, we use the framework to modernize code by enabling higher level of abstraction in expressing computations. We evaluate the effectiveness of the tool with respect to speedup obtained relative to original code performance, and also quantify productivity gains. |
| 書誌レコードID |
|
|
収録物識別子タイプ |
NCID |
|
収録物識別子 |
AN10463942 |
| 書誌情報 |
研究報告ハイパフォーマンスコンピューティング（HPC）
巻 2019-HPC-170,
号 30,
p. 1-9,
発行日 2019-07-17
|
| ISSN |
|
|
収録物識別子タイプ |
ISSN |
|
収録物識別子 |
2188-8841 |
| Notice |
|
|
|
SIG Technical Reports are nonrefereed and hence may later appear in any journals, conferences, symposia, etc. |
| 出版者 |
|
|
言語 |
ja |
|
出版者 |
情報処理学会 |
IPSJ-HPC19170030.pdf (1.0 MB)
