ASE 2024
Sun 27 October - Fri 1 November 2024 Sacramento, California, United States
Tue 29 Oct 2024 16:25 - 16:35 at Carr - Performance and load

This paper describes a formal repair framework for performance bugs in loop programs, which are programming errors that slow down program execution. The approach is developed based on the observation that a program with a performance bug is a semantically correct program, but it may perform inefficiently for some inputs. This observation permits the formal treatment of performance bugs using the idea of program invariants, where the original program is augmented with a number of non-functional variables that are used to assess the efficiency of the patched version vs. the original program using the derived invariants. The proposed approach offers two major advantages compared to the conventional test-based patch validation approach. First, it enables the formal validation of patches using program verifiers. Second, it helps to assess the efficiency boost provided by the generated patches. To the best of our knowledge, the formal treatment of performance bugs has not been studied in the prior literature.

Tue 29 Oct

Displayed time zone: Pacific Time (US & Canada) change

15:30 - 17:00
15:30
15m
Talk
AI-driven Java Performance Testing: Balancing Result Quality with Testing Time
Research Papers
Luca Traini University of L'Aquila, Federico Di Menna University of L'Aquila, Vittorio Cortellessa University of L'Aquila
DOI Pre-print
15:45
15m
Talk
MLOLET - Machine Learning Optimized Load and Endurance Testing: An industrial experience report
Industry Showcase
Arthur Vitui Concordia University, Tse-Hsun (Peter) Chen Concordia University
16:00
15m
Talk
Dynamic Scoring Code Token Tree: A Novel Decoding Strategy for Generating High-Performance Code
Research Papers
Muzi Qu University of Chinese Academy of Sciences, Jie Liu Institute of Software, Chinese Academy of Sciences, Liangyi Kang Institute of Software, Chinese Academy of Sciences, Shuai Wang Institute of Software, Chinese Academy of Sciences, Dan Ye Institute of Software, Chinese Academy of Sciences, Tao Huang Institute of Software at Chinese Academy of Sciences
16:15
10m
Talk
BenchCloud: A Platform for Scalable Performance Benchmarking
Tool Demonstrations
Dirk Beyer LMU Munich, Po-Chun Chien LMU Munich, Marek Jankola LMU Munich
DOI Pre-print Media Attached
16:25
10m
Talk
A Formal Treatment of Performance BugsRecorded Talk
NIER Track
Omar I. Al Bataineh Gran Sasso Science Institute (GSSI)