Synchronization Inference

This lecture covers the automatic synthesis of correct and efficient synchronization in programs, focusing on the goal of generating synchronized programs from an intuitive starting point. The process involves removing redundant atomicity, reordering statements, and adding synchronization metadata. The lecture also explores the challenges of atomicity reduction, the application of transformations, and the generation-check process. Additionally, it delves into concurrent sets, the steps involved in generating and checking synchronization, and the concept of abstraction-guided synthesis for automatic verification. The lecture concludes with a discussion on the IBM Project Paraglide and the implications of abstraction-guided synthesis in program verification.