基于OpenCL的连续数据无关访存密集型函数并行与优化研究
文献类型：期刊论文
| 作者 | 蒋丽媛 ; 张云泉 ; 龙国平 ; 贾海鹏 |
| 刊名 | 计算机科学
 |
| 出版日期 | 2013 |
| 卷号 | 40期号:3页码:111-115 |
| 关键词 | GPU OpenCL 向量化 ROI |
| ISSN号 | 1002-137X |
| 其他题名 | parallelism and research on functions with continuously independent data and intensive memory access using opencl |
| 中文摘要 | 连续的数据无关是指计算目标矩阵连续的元素时使用的源矩阵元素之间没有关系且也为连续的,访存密集型是指函数的计算量较小,但是有大量的数据传输操作。在OpenCL框架下,以bitwise函数为例,研究和实现了连续数据无关访存密集型函数在GPU平台上的并行与优化。在考察向量化、线程组织方式和指令选择优化等多个优化角度在不同的GPU硬件平台上对性能的影响之后,实现了这个函数的跨平台性能移植。实验结果表明,在不考虑数据传输的前提下,优化后的函数与这个函数在OpenCV库中的CPU版本相比,在AMD HD 5850GPU达到了平均40倍的性能加速比;在AMD HD 7970GPU达到了平均90倍的性能加速比;在NVIDIA Tesla C2050GPU上达到了平均60倍的性能加速比;同时,与这个函数在OpenCV库中的CUDA实现相比,在NVIDIA Tesla C2050平台上也达到了1.5倍的性能加速。 |
| 英文摘要 | Continuously independent data type means when calculating the continuous elements of destination matrix, the used elements of source matrices are also continuous and there are no relationship among them. Intensive memory access function is the function that has less computation but a lot of data transfer operations. This paper took the bitwise function as the example, studied and implemented the parallel and the optimizing methods of the continuously independent data and intensive memory access function on GPU platforms. Based on the OpenCL framework, this paper studied and compared various optimizing methods,such as vectorizing, threads organizing, and instruction selecting, and finally used these methods to implement the cross-platform transfer of the bitwise function among different platforms. The study tested the function's execution time without data transfer both on AMD GPU and NVIDIA GPU platforms. On the AMD Radeon HD 5850 platform, the performance has reached 40 times faster than the CPU version in OpenCV library, 90 times faster on AMD Radeon HD 7970 platform, and 60 times faster on NVIDIA GPU Tesla C2050 platform. On NVIDIA GPU Tesla C2050 platform, the speedup is 1.5 comparing with the CUDA version in OpenCV library. |
| 学科主题 | Computer Science (provided by Thomson Reuters) |
| 收录类别 | CNKI ; WANFANG ; CSCD |
| 资助信息 | 国家自然科学基金资助项目(60303020,60533020),国家自然科学基金资助重点项目(60503020),国家自然科学基金青年基金课题(61100072)|国家“863”计划基金资助项目(2012AA010902)|ISCAS-AMD联合fusion软件中心资助 |
| 语种 | 中文 |
| CSCD记录号 | CSCD:4778295 |
| 公开日期 | 2013-09-17 |
| 源URL | [http://ir.iscas.ac.cn/handle/311060/15559]  |
| 专题 | 软件研究所_软件所图书馆_期刊论文 |
| 推荐引用方式 GB/T 7714 | 蒋丽媛,张云泉,龙国平,等. 基于OpenCL的连续数据无关访存密集型函数并行与优化研究[J]. 计算机科学,2013,40(3):111-115. |
| APA | 蒋丽媛,张云泉,龙国平,&贾海鹏.(2013).基于OpenCL的连续数据无关访存密集型函数并行与优化研究.计算机科学,40(3),111-115. |
| MLA | 蒋丽媛,et al."基于OpenCL的连续数据无关访存密集型函数并行与优化研究".计算机科学 40.3(2013):111-115. |
入库方式： OAI收割
来源：软件研究所
浏览0
下载0
收藏0
其他版本
除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。