Infrastructures for accelerating the synergy effect of software-hardware co-design
Project/Area Number |
25280041
|
Research Category |
Grant-in-Aid for Scientific Research (B)
|
Allocation Type | Partial Multi-year Fund |
Section | 一般 |
Research Field |
High performance computing
|
Research Institution | Tohoku University |
Principal Investigator |
|
Co-Investigator(Kenkyū-buntansha) |
Kobayashi Hiroaki 東北大学, サイバーサイエンスセンター, 教授 (40205480)
Aoki Takafumi 東北大学, 情報科学研究科, 教授 (80241529)
Sano Kentaro 東北大学, 大学院・情報科学研究科, 准教授 (00323048)
|
Co-Investigator(Renkei-kenkyūsha) |
Egawa Ryusuke 東北大学, サイバーサイエンスセンター, 准教授 (80374990)
Tada Jube 山形大学, 理工学研究科, 助教 (30361273)
Ito Koichi 東北大学, 情報科学研究科, 助教 (70400299)
|
Project Period (FY) |
2013-04-01 – 2016-03-31
|
Project Status |
Completed (Fiscal Year 2015)
|
Budget Amount *help |
¥18,070,000 (Direct Cost: ¥13,900,000、Indirect Cost: ¥4,170,000)
Fiscal Year 2015: ¥4,420,000 (Direct Cost: ¥3,400,000、Indirect Cost: ¥1,020,000)
Fiscal Year 2014: ¥6,110,000 (Direct Cost: ¥4,700,000、Indirect Cost: ¥1,410,000)
Fiscal Year 2013: ¥7,540,000 (Direct Cost: ¥5,800,000、Indirect Cost: ¥1,740,000)
|
Keywords | アクセラレータ / GPU / FPGA / 国際情報交換(米国) |
Outline of Final Research Achievements |
Assuming OpenCL as a standard environment for accelerator programming, we have pointed out some missing features for supporting more various accelerator architectures,and proposed OpenCL extensions. Although OpenCL has gradually become to be used for hardware description, OpenCL C is not necessarily appropriate for describing OpenCL kernels. Hence, we have designed and implemented high productivity languages for typical computations in the fields of image processing and high performance computing. In addition, we have proposed an automatic tuning method for performance parameters, which need to be adjusted for individual accelerators. The proposed method has been implemented for evaluating its performance impacts.
|
Report
(4 results)
Research Products
(47 results)