Title:LLM-Based Program Generation for Triggering Numerical Inconsistencies Across Compilers

Abstract:Floating-point inconsistencies across compilers can undermine the reliability of numerical software. We present LLM4FP, the first framework that uses Large Language Models (LLMs) to generate floating-point programs specifically designed to trigger such inconsistencies. LLM4FP combines Grammar-Based Generation and Feedback-Based Mutation to produce diverse and valid programs. We evaluate LLM4FP across multiple compilers and optimization levels, measuring inconsistency rate, time cost, and program diversity. LLM4FP detects over twice as many inconsistencies compared to the state-of-the-art tool, Varity. Notably, most of the inconsistencies involve real-valued differences, rather than extreme values like NaN or infinities. LLM4FP also uncovers inconsistencies across a wider range of optimization levels, and finds the most mismatches between host and device compilers. These results show that LLM-guided program generation improves the detection of numerical inconsistencies.