請用此 Handle URI 來引用此文件：
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/61677完整後設資料紀錄
| DC 欄位 | 值 | 語言 |
|---|---|---|
| dc.contributor.advisor | 廖世偉(Shih-wei Liao) | |
| dc.contributor.author | Jr-Yuan Chiou | en |
| dc.contributor.author | 邱致遠 | zh_TW |
| dc.date.accessioned | 2021-06-16T13:09:19Z | - |
| dc.date.available | 2013-08-14 | |
| dc.date.copyright | 2013-08-14 | |
| dc.date.issued | 2013 | |
| dc.date.submitted | 2013-07-31 | |
| dc.identifier.citation | [1] Clang: A C language family frontend for LLVM.
[2] G. Diamos. The Design and Implementation Ocelot ’ s Dynamic Binary Translator from PTX to Multi-Core x86. Computer Engineering, 9:22–39, 2009. [3] Google. Renderscript. [4] Google. Renderscript Compute. [5] K. Group. OpenCL. [6] M. J. Harvey and G. De Fabritiis. Swan: A tool for porting CUDA programs to OpenCL. Computer Physics Communications, 182(4):1093–1099, 2011. [7] C.-W. Huang. Android-x86 project, 2010. [8] C. Lattner and V. Adve. LLVM: A compilation framework for lifelong program analysis & transformation. International Symposium on Code Generation and Optimization 2004 CGO 2004, 53706(c):75–86, 2004. [9] G. Martinez, M. Gardner, and W. chun Feng. CU2CL: A CUDA-to-OpenCL Translator for Multi- and Many-Core Architectures. In Parallel and Distributed Systems (ICPADS), 2011 IEEE 17th International Conference on, pages 300 –307, December 2011. [10] J. Sams. Levels in Renderscript. [11] J. Stratton, S. Stone, and W. Hwu. MCUDA: An efficient implementation of CUDA kernels for multi-core CPUs. Languages and Compilers for Parallel Computing, pages 16–30, 2008. | |
| dc.identifier.uri | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/61677 | - |
| dc.description.abstract | With 500 million units and counting, mobile Android devices contribute
to more than 50% of the smartphone market this year. As Android platform matures, the demand for parallel computing increases. In response, Google released Renderscript to leverage parallel computing while maintaining portability. However, because many more GPU vendors support OpenCL than Renderscript, a Renderscript-to-OpenCL translator will enable GPUs to be utilized for compute. Thus, we develop the RS2CL translator to enable GPGPU for Android systems. We show that RS2CL automatically generates OpenCL kernel programs successfully. In addition, we present comprehensive comparisons between Renderscript and OpenCL models, in order to validate RS2CL. Because RS2CL enables GPU compute on GPUs with OpenCL support, we achieve a speedup of 4.17 on two more time-consuming image processing applications. | en |
| dc.description.provenance | Made available in DSpace on 2021-06-16T13:09:19Z (GMT). No. of bitstreams: 1 ntu-102-R00922120-1.pdf: 497653 bytes, checksum: 2ac0149a5e073d6ae982550962a97a3a (MD5) Previous issue date: 2013 | en |
| dc.description.tableofcontents | 口試委員會審定書i
Acknowledgments iii Abstract v 1 INTRODUCTION 1 2 RELATED WORK 3 3 RENDERSCRIPT AND OPENCL OVERVIEW 5 3.1 Renderscript . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 3.2 OpenCL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 3.3 Differences between Renderscript and OpenCL . . . . . . . . . . . . . . 6 4 ARCHITECTURE OF RENDERSCRIPT SYSTEM 9 5 DESIGN AND IMPLEMENTATION 11 6 ISSUES IN RENDERSCRIPT TO OPENCL TRANSLATION 13 6.1 Support for domain-specific language (DSL) via pragma . . . . . . . . . 13 6.2 Global declarations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 6.3 Renderscript special type and qualifiers . . . . . . . . . . . . . . . . . . 14 6.4 Type-based rewriting of Renderscript functions . . . . . . . . . . . . . . 14 6.4.1 Map Functions directly from Renderscript to OpenCL . . . . . . 14 6.4.2 Non-Directly Map Functions from renderscript to OpenCL . . . . 14 6.5 Different models in Renderscript and OpenCL . . . . . . . . . . . . . . . 16 7 EXPERIMENTAL RESULTS 19 7.1 RS2CL translation time performance . . . . . . . . . . . . . . . . . . . . 20 7.2 Comparison of OpenCL code runtime performance translated by RS2CL and written by manual run on CPU . . . . . . . . . . . . . . . . . . . . . 20 7.3 Comparison of Renderscript runtime performance and OpenCL runtime performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 8 CONCLUSION AND FUTURE WORK 23 Bibliography 25 | |
| dc.language.iso | en | |
| dc.title | RS2CL: Renderscript 轉 OpenCL 轉換器 | zh_TW |
| dc.title | RS2CL: A Renderscript-to-OpenCL translator | en |
| dc.type | Thesis | |
| dc.date.schoolyear | 101-2 | |
| dc.description.degree | 碩士 | |
| dc.contributor.oralexamcommittee | 徐慰中,陳呈瑋(Cheng-Wei Chen),黃維中(Wei-Chung Hwang) | |
| dc.subject.keyword | Clang,Renderscript,OpenCL,轉換器, | zh_TW |
| dc.subject.keyword | Clang,Renderscript,OpenCL,translator, | en |
| dc.relation.page | 25 | |
| dc.rights.note | 有償授權 | |
| dc.date.accepted | 2013-08-01 | |
| dc.contributor.author-college | 電機資訊學院 | zh_TW |
| dc.contributor.author-dept | 資訊工程學研究所 | zh_TW |
| 顯示於系所單位： | 資訊工程學系 | |
文件中的檔案：
| 檔案 | 大小 | 格式 | |
|---|---|---|---|
| ntu-102-1.pdf 目前未授權公開取用 | 485.99 kB | Adobe PDF |
系統中的文件，除了特別指名其著作權條款之外，均受到著作權保護，並且保留所有的權利。