LIS25-137 Using LLMs to optimize CI performance and analysis

• Wednesday, 14 May 15:30 - 15:55
• Room: Session room 2 | Opala I & II

In this talk, we will go through the use of readily available Large Language Models (LLMs) to facilitate engineering workflow when running Continuous Integration (CI) loops and analyzing regressions. For this presentation, Anthropic’s Claude Sonnet 3.5 and the LLVM buildbots will be used as an example. We will cover: - How to quickly set up an off-the-shelf LLM to use in software development. - Using LLMs to pinpoint problematic commits and avoid time-consuming bisection process. - Using LLMs to analyze commits when looking for the cause of odd regressions. - Future improvements/possibilities of use in performance analysis and CI automation in open source projects. - Current limitations.

Add to My Agenda