Semantic Reification for Robust Random Program Generation

This research presentation focuses on compiler reliability, leveraging semantic reification as a novel approach to improve robustness. Existing randomised testing tools like Csmith are reevaluated, highlighting their limitations. Methods involve generating random control flow graphs (CFGs), sampling execution paths, and implementing symbolic program functions to detect compiler bugs in GCC, LLVM, and JVM. Key contributions include introducing semantic reification, handling arbitrary control...