Performance Evaluations of Parallelized DFT Calculations with SMASH on Intel Xeon Phi Processor
-
- SAITOU Sona
- Department of Chemistry and Research Center for Smart Molecules, College of Science, Rikkyo University, 3-34-1 Nishi-ikebukuro, Toshima-ku, Tokyo 171-8501, Japan
-
- MOCHIZUKI Yuji
- Department of Chemistry and Research Center for Smart Molecules, College of Science, Rikkyo University, 3-34-1 Nishi-ikebukuro, Toshima-ku, Tokyo 171-8501, Japan Institute of Industrial Science, The University of Tokyo, 4-6-1 Komaba, Meguro-ku, Tokyo 153-8503, Japan
-
- YAMAZAKI Yutaka
- Department of Chemistry and Research Center for Smart Molecules, College of Science, Rikkyo University, 3-34-1 Nishi-ikebukuro, Toshima-ku, Tokyo 171-8501, Japan
-
- ISHIMURA Kazuya
- Institute of Molecular Science, 38 Saigonaka, Myodaiji-cho, Okazaki, Aichi 444-8585, Japan
Bibliographic Information
- Other Title
-
- Intel Xeon Phi上でのSMASHによる並列化DFT計算の性能評価
Abstract
<p>Intel's Xeon Phi processors have attracted considerable attention as promising many-core processors for scientific computing and simulation. We report the performance results of parallelized DFT calculations with the SMASH program on a first-generation chip (Knights Corner) in a native mode of executions. As a result, an acceleration of 1.4 was observed for a multi-threading increment from 60 cores to 120 cores. This fact implies a preferable performance on the second-generation Phi (Knights Landing) processors.</p>
Journal
-
- Journal of Computer Chemistry, Japan
-
Journal of Computer Chemistry, Japan 15 (4), 92-96, 2016
Society of Computer Chemistry, Japan
- Tweet
Details 詳細情報について
-
- CRID
- 1390282680157064576
-
- NII Article ID
- 130005434728
-
- ISSN
- 13473824
- 13471767
-
- Text Lang
- ja
-
- Data Source
-
- JaLC
- Crossref
- CiNii Articles
- KAKEN
-
- Abstract License Flag
- Disallowed