Murali Srirangam Ramanujam will present a General Exam entitled "App-driven optimizations to boost smartphone resource efficiency" on Monday, October 3, 2022 at 3pm in CS 302. Zoom link: [ https://princeton.zoom.us/my/muralisr92 | https://princeton.zoom.us/my/muralisr92 ] The members of the committee are as follows: Ravi Netravali (adviser), Amit Levy, and Wyatt Lloyd Abstract: [ https://docs.google.com/document/d/1trZhu-zt9wRft6nJO5_BnOCvgQMjd2n4D7s5pbPS... | https://docs.google.com/document/d/1trZhu-zt9wRft6nJO5_BnOCvgQMjd2n4D7s5pbPS... ] Low interaction response times are crucial to the experience that mobile apps provide for their users. These response times are determined by both the network delays incurred during content fetches, as well as the client-side computation delays to parse/render that content and other code in the app. Unfortunately, existing strategies to alleviate the network and compute latencies that hinder app responsiveness fall short in practice. We first present Marauder, a system to address network latencies by synergizing caching and prefetching to improve speedups achieved by each technique while avoiding their inherent limitations. Key to Marauder is our observation that, like web pages, apps handle interactions by downloading and parsing structured text resources that entirely list (i.e., without needing to consult app binaries) the set of other resources to load. Next, we present Floo, a system that aims to automatically reuse (or memoize) computation results during app operation in an effort to reduce the amount of computation needed to handle user interactions. In designing Floo, our primary insight is that app computations exhibit substantial stability over time in that they are entirely performed in rarely-updated codebases within app binaries and the OS. Across a wide range of apps, live networks, phones, and interaction traces, Marauder and Floo reduce median interaction response times by 27.4% and 32.7% respectively. Reading List: https://docs.google.com/document/d/1N_2e1CevxdQVyqnMYqi79TzcnpS5p3R64ybmMdkc...