Phaedrus: Predicting Dynamic Application Behavior with Lightweight Generative Models and LLMs
Proceedings of the ACM on Programming Languages · 10.04.2026 · 10 (OOPSLA1) · 707-736
DOI: 10.1145/3798225 ↗Kurz gefasst
Fachartikel in der Zeitschrift Proceedings of the ACM on Programming Languages, online erschienen am 10.04.2026. Verfasst von Bodhisatwa Chatterjee und 2 weiteren Autor:innen. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 83 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Application profiling is an indispensable technique for many software development tasks, such as code and memory layout optimizations, where optimization decisions are tailored to specific program profiles. Unfortunately, modern application codebases exhibit highly variant behavior across different inputs, creating challenges for conventional profiling approaches that rely on a single representative execution instance. In this paper, we propose Phaedrus , a new compiler-assisted deep learning framework designed to predict dynamic program behavior across varied execution instances, specifically focusing on dynamic function call prediction. These predicted call sequences are subsequently used to guide input-specific compiler optimizations, producing code specialized for each execution instance. Traditional profile-guided optimization methods struggle with the input-dependent variability of modern applications, where profiling on different inputs yields divergent application behaviors. To address this, Phaedrus proposes two new approaches: Application Profile Synthesis (Dynamis) , a profile-less approach where Large Language Models (LLMs) directly infer dynamic functions based on source code & static compiler analysis, bypassing the need for traditional profiling, and Application Profile Generalization (Morpheus) , which uses generative models trained on compressed and augmented Whole Program Path (WPP) based function profiles to predict application behavior under unseen inputs. Our experiments show that Phaedrus accurately identifies the most frequently executed and runtime-dominated hotspot functions, accounting for up to 85–99% of total execution time. Leveraging these predictions, Phaedrus enables superior profile-guided optimizations, delivering an average speedup of 6% (up to 25%) and a binary size reduction of 5.19% (up to 19%), without any program execution. In addition, Phaedrus reduces WPP function profile sizes by up to 10 7 ×.
Zitieren
APA
Chatterjee, B., Jadhav, N., & Pande, S. (2026). Phaedrus: Predicting Dynamic Application Behavior with Lightweight Generative Models and LLMs. Proceedings of the ACM on Programming Languages, 10(OOPSLA1), 707-736. https://doi.org/10.1145/3798225
Vancouver
Chatterjee B, Jadhav N, Pande S. Phaedrus: Predicting Dynamic Application Behavior with Lightweight Generative Models and LLMs. Proceedings of the ACM on Programming Languages. 2026;10(OOPSLA1):707-736. doi:10.1145/3798225
BibTeX
@article{Chatterjee2026,
author = {Chatterjee, Bodhisatwa and Jadhav, Neeraj and Pande, Santosh},
title = {Phaedrus: Predicting Dynamic Application Behavior with Lightweight Generative Models and LLMs},
journal = {Proceedings of the ACM on Programming Languages},
year = {2026},
volume = {10},
number = {OOPSLA1},
pages = {707--736},
doi = {10.1145/3798225}
}
EUVIMED