Compilation Optimization Pass Selection Using Gate Graph Attention Neural Network for Reliability Improvement
Compilation Optimization Pass Selection Using Gate Graph Attention Neural Network for Reliability Improvement
复制标题
使用门图注意力神经网络进行编译优化通道选择以提高可靠性
DOI：
10.1109/access.2020.3016758
复制
发表时间：
2020-08
期刊：
影响因子：
3.9
通讯作者：
Li Long
中科院分区：
文献类型：
--
作者：
Wu Jiang;Xu Jianjun;Meng Xiankai;Zhang Haoyu;Zhang Zhuo;Li Long
当处理不同的程序或应用程序时，需要为程序选择合适的编译优化阶段或子序列。机器学习作为解决这一问题的有效技术手段被广泛应用。然而，使用机器学习时最重要的问题是程序特征的提取。在这种情况下，从源代码中获取更多的语义和语法信息以及代码段之间的复杂转换显然是必要的，也是以前的工作可能忽略的领域。确保计划信息的完整性和有效性是解决这个问题的关键。而且，在执行和改进选型时，衡量指标往往是程序性能、代码大小等；这方面对程序可靠性的研究有限，需要最长的测量时间和最复杂的测量方法。因此，本文建立了组合程序特征提取模型，并提出了一种基于图的编译优化通道选择模型，该模型可以学习程序可靠性的启发式方法。本实验使用clang编译框架进行。替代编译优化通道采用C语言标准编译优化通道。与传统的机器学习方法相比，我们的模型在程序可靠性优化通道选择方面的平均准确率提高了 5% 至 11%。我们的实验还证明了我们提出的模型的强大可扩展性。
When dealing with different programs or applications, it is necessary to select the appropriate compilation optimization pass or subsequence for the program. Machine learning is widely used as an efficient technological means of solving this problem. However, the most important problem when using machine learning is the extraction of program features. Obtaining more semantic and syntax information and complex transitions among code segments from the source code are obviously necessary in this context, and is also an area that may have been neglected by previous work. Ensuring the integrity and effectiveness of program information is key to this problem. Moreover, when performing and improving the selection, the measurement indicators are often program performance, code size, etc.; there is limited research on program reliability in this context, which requires both the longest measurement time and the most complicated measurement methods. Accordingly, this paper establishes a combined program feature extraction model and proposes a graph-based compilation optimization pass selection model that learns heuristics for program reliability. This experiment was performed using the clang compilation framework. The alternative compilation optimization pass adopts the C language standard compilation optimization passes. Compared with traditional machine learning methods, our model improves the average accuracy by between 5% and 11% in the optimization pass selection for program reliability. Our experiments also demonstrate the strong scalability of our proposed model.
登录
查看更多内容
DOI：
10.13140/rg.2.2.20274.09921
发表时间：
2016
期刊：
--
影响因子：
--
作者：
Amir H. Ashouri
通讯作者：
Amir H. Ashouri
DOI：
10.1109/dsn.2018.00016
发表时间：
2018-06
期刊：
2018 48th Annual IEEE/IFIP International Conference on Dependable Systems and Networks (DSN)
影响因子：
--
作者：
Guanpeng Li;K. Pattabiraman;S. Hari;Michael B. Sullivan;Timothy Tsai
通讯作者：
Guanpeng Li;K. Pattabiraman;S. Hari;Michael B. Sullivan;Timothy Tsai
DOI：
10.1109/sccc.2011.25
发表时间：
2011-11
期刊：
2011 30th International Conference of the Chilean Computer Science Society
影响因子：
--
作者：
J. Foleiss;A. F. Silva;L. B. Ruiz
通讯作者：
J. Foleiss;A. F. Silva;L. B. Ruiz
影响因子：
2.2
作者：
Hui Liu;Rongcai Zhao;Qi Wang;Yingying Li
通讯作者：
Hui Liu;Rongcai Zhao;Qi Wang;Yingying Li
影响因子：
2.2
作者：
Bruce W. Leverett;R. Cattell;Steven O. Hobbs;J. Newcomer;A. Reiner;Bruce R. Schatz;W. Wulf
通讯作者：
Bruce W. Leverett;R. Cattell;Steven O. Hobbs;J. Newcomer;A. Reiner;Bruce R. Schatz;W. Wulf