| SIMD自动向量化编译优化概述 |
| |
| 引用本文： | 高伟,赵荣彩,韩林,庞建民,丁锐.SIMD自动向量化编译优化概述[J].软件学报,2015,26(6):1265-1284. |
| |
| 作者姓名： | 高伟 赵荣彩 韩林 庞建民 丁锐 |
| |
| 作者单位： | 数学工程与先进计算国家重点实验室(解放军信息工程大学), 河南 郑州 450001,数学工程与先进计算国家重点实验室(解放军信息工程大学), 河南 郑州 450001,数学工程与先进计算国家重点实验室(解放军信息工程大学), 河南 郑州 450001,数学工程与先进计算国家重点实验室(解放军信息工程大学), 河南 郑州 450001,数学工程与先进计算国家重点实验室(解放军信息工程大学), 河南 郑州 450001 |
| |
| 基金项目： | “核高基”国家科技重大专项(2009ZX01036-001-001-2) |
| |
| 摘 要： | SIMD扩展部件是集成到通用处理器中的加速部件,旨在发掘多媒体程序和科学计算程序的数据级并行.首先介绍SIMD扩展部件的背景和研究现状,然后从发掘方法、数据布局、多平台向量化这3个角度介绍了SIMD自动向量化的研究问题、困难和最新研究成果,最后展望了SIMD编译优化未来的研究方向.
|
| 关 键 词： | SIMD扩展部件 自动向量化 数据级并行 编译优化 |
| 收稿时间： | 4/8/2014 12:00:00 AM |
| 修稿时间： | 2014/12/22 0:00:00 |
Research on SIMD Auto-Vectorization Compiling Optimization |
| |
| GAO Wei,ZHAO Rong-Cai,HAN Lin,PANG Jian-Min and DING Rui.Research on SIMD Auto-Vectorization Compiling Optimization[J].Journal of Software,2015,26(6):1265-1284. |
| |
| Authors: | GAO Wei ZHAO Rong-Cai HAN Lin PANG Jian-Min and DING Rui |
| |
| Affiliation: | State Key Laboratory of Mathematical Engineering and Advanced Computing (PLA Information Engineering University), Zhengzhou 450001, China,State Key Laboratory of Mathematical Engineering and Advanced Computing (PLA Information Engineering University), Zhengzhou 450001, China,State Key Laboratory of Mathematical Engineering and Advanced Computing (PLA Information Engineering University), Zhengzhou 450001, China,State Key Laboratory of Mathematical Engineering and Advanced Computing (PLA Information Engineering University), Zhengzhou 450001, China and State Key Laboratory of Mathematical Engineering and Advanced Computing (PLA Information Engineering University), Zhengzhou 450001, China |
| |
| Abstract: | SIMD extension is an acceleration component integrated into the general processor for developing data level parallelism in multimedia and scientific computing applications. Firstly, in this study the background and research status of SIMD extension are introduced. Next, challenges and latest research achievements in SIMD auto-vectorization are discussed from three perspectives: development method, data layout and vectorization in multi-platform. Finally, some future trends in the SIMD compiling optimization are addressed. |
| |
| Keywords: | SIMD extension auto-vectorization data level parallelism compiling optimization |
| 本文献已被 万方数据 等数据库收录！ |
| 点击此处可从《软件学报》浏览原始摘要信息 |
|
点击此处可从《软件学报》下载全文 |
|
