CSR: Medium: Understanding and Automatically Adjusting Performance Sensitive Software Configurations

CSR：中：了解并自动调整性能敏感的软件配置

• 批准号: 1764039
• 负责人: Henry Hoffmann
• 金额: $ 114.91万
• 依托单位: University of Chicago
• 依托单位国家: 美国
• 项目类别: Continuing Grant
• 财政年份: 2018
• 资助国家: 美国
• 起止时间: 2018-09-01 至 2023-08-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1764039&HistoricalAwards=false
• 关键词: CSR; Medium; Understanding; Automatically; Adjusting

Every day, people interact with complex software systems on their mobile devices and in the cloud. Experts develop this software, but they cannot account for all possible uses, so they make the software configurable using redefinable parameters. Most of these parameters affect software's performance. These performance-related configuration parameters are growing exponentially. Unfortunately, so is the difficulty of setting them. Performance-related misconfigurations are a leading cause of system failures (e.g., web services crashing) and lag (e.g., delayed response to user requests). This project will develop AutoConf to automatically set performance-related configuration parameters and eliminate these sources of failure in critical software infrastructure.To better understand and manage performance-related configurations, AutoConf has four major research objectives. The first is studying existing configuration parameters to understand why and when they lead to failures. The second is designing a new interface where users set high-level goals in terms of desired performance rather than setting specific parameter values. The third is developing techniques that automatically and dynamically configure software to meet these goals. The final objective is constructing tools that ensure the interface's proper usage to produce the desired behavior.If successful, AutoConf will change the way users deploy software. Instead of setting 100s of configuration parameters, they will state their goals. AutoConf will automatically adjust performance-related configurations to ensure the goals are met, resulting in faster and more robust software deployment. Additionally, the AutoConf project will create many research opportunities that will be used to broaden participation in computer science.Documentation and software releases will be posted at http://autoconf.cs.uchicago.edu/.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

每天，人们都会在其移动设备和云中与复杂的软件系统进行互动。 专家开发此软件，但无法解释所有可能的用途，因此它们可以使用可重新定义的参数使软件可配置。这些参数大多数都会影响软件的性能。这些与性能相关的配置参数正在成倍增长。 不幸的是，设置它们的困难也是如此。 与性能相关的错误配置是系统故障（例如Web服务崩溃）和滞后（例如，延迟对用户请求的响应）的主要原因。 该项目将开发自动设置与性能相关的配置参数并消除关键软件基础架构中的这些故障来源。为了更好地理解和管理与性能相关的配置，AutoConf具有四个主要的研究目标。 首先是研究现有的配置参数，以了解原因以及何时导致故障。 第二个是设计一个新界面，用户在其中设定了所需性能而不是设置特定参数值的高级目标。 第三个是开发自动和动态配置软件以满足这些目标的技术。 最终的目标是构建工具，以确保接口的适当用法产生所需的行为。如果成功，AutoConf将改变用户部署软件的方式。 他们将陈述目标，而不是设置100张配置参数。 AutoConf将自动调整与性能相关的配置，以确保实现目标，从而更快，更强大的软件部署。 Additionally, the AutoConf project will create many research opportunities that will be used to broaden participation in computer science.Documentation and software releases will be posted at http://autoconf.cs.uchicago.edu/.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

项目成果

Orthogonalized SGD and Nested Architectures for Anytime Neural Networks

适用于随时神经网络的正交化 SGD 和嵌套架构

• 发表时间: 2020
• 期刊: 37th International Conference on Machine Learning
• 作者: Wan, Chengcheng;Hoffmann, Henry;Lu, Shan;Maire, Michael
• 通讯作者: Maire, Michael

Proxima: accelerating the integration of machine learning in atomistic simulations

• DOI: 10.1145/3447818.3460370
• 发表时间: 2021-06
• 期刊: Proceedings of the 35th ACM International Conference on Supercomputing
• 作者: Yuliana Zamora;Logan T. Ward;G. Sivaraman;Ian T. Foster;H. Hoffmann

Dynamical Systems Theory for Causal Inference with Application to Synthetic Control Methods

因果推理的动力系统理论及其在综合控制方法中的应用

• 发表时间: 2020
• 期刊: Proceedings of the Twenty Third International Conference on Artificial Intelligence and Statistics
• 作者: Ding, Yi;Toulis, Panos
• 通讯作者: Toulis, Panos

Generalizable and interpretable learning for configuration extrapolation

• DOI: 10.1145/3468264.3468603
• 发表时间: 2021-08
• 期刊: Proceedings of the 29th ACM Joint Meeting on European Software Engineering Conference and Symposium on the Foundations of Software Engineering
• 作者: Yi Ding;Ahsan Pervaiz;Michael Carbin;H. Hoffmann

GOAL: Supporting General and Dynamic Adaptation in Computing Systems

• DOI: 10.1145/3563835.3567655
• 发表时间: 2022-11
• 期刊: Proceedings of the 2022 ACM SIGPLAN International Symposium on New Ideas, New Paradigms, and Reflections on Programming and Software
• 作者: Ahsan Pervaiz;Yao-Hsiang Yang;Adam Duracz;F. Bartha;R. Sai;Connor Imes;Robert Cartwright;K. Palem;Shan Lu;Henry Hoffmann