arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2206.13619cs.SEcs.AIcs.PF
收藏

DeepPERF:一种基于深度学习的软件性能改进方法

DeepPERF: A Deep Learning-Based Approach For Improving Software Performance

Microsoft(微软)

Spandan Garg, Roshanak Zilouchian Moghaddam, Colin B. Clement, Neel Sundaresan, Chen Wu

摘要:

提高软件性能是软件开发周期中一个重要而又具有挑战性的部分。如今,大多数性能低下的问题都是由性能专家识别和修补的。深度学习方法的最新进展和开放源数据的广泛可用性为自动识别和修补性能问题创造了很好的机会。在本文中,我们提出了DeepPERF,这是一种基于transformer的方法,可以为C#应用程序提供性能改进。我们在英语和源代码语料库上预训练DeepPERF,然后进行微调,以完成为C#应用程序生成性能改进补丁的任务。我们的评估表明,我们的模型可以在约53%的情况下生成与开发人员修复相同的性能改进建议,在我们的专家验证的C#开发人员性能更改数据集中,其中约34%是逐字的。此外,我们使用基准测试和单元测试在GitHub上的50个开源C#存储库上评估了DeepPERF,发现我们的模型能够提出有效的性能改进,可以提高CPU使用率和内存分配。到目前为止,我们已经提交了19个pull-request,其中有28个不同的性能优化,其中11个PR已经得到了项目所有者的批准。

英文摘要:

Improving software performance is an important yet challenging part of the software development cycle. Today, the majority of performance inefficiencies are identified and patched by performance experts. Recent advancements in deep learning approaches and the wide-spread availability of open source data creates a great opportunity to automate the identification and patching of performance problems. In this paper, we present DeepPERF, a transformer-based approach to suggest performance improvements for C# applications. We pretrain DeepPERF on English and Source code corpora and followed by finetuning for the task of generating performance improvement patches for C# applications. Our evaluation shows that our model can generate the same performance improvement suggestion as the developer fix in ~53% of the cases, getting ~34% of them verbatim in our expert-verified dataset of performance changes made by C# developers. Additionally, we evaluate DeepPERF on 50 open source C# repositories on GitHub using both benchmark and unit tests and find that our model is able to suggest valid performance improvements that can improve both CPU usage and Memory allocations. So far we've submitted 19 pull-requests with 28 different performance optimizations and 11 of these PRs have been approved by the project owners.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑