sayzardPractical Formal Verification for MLIR Programs
이 논문은 MLIR 프로그램의 변환 최적화가 올바르게 수행되었는지 형식적으로 검증하는 방법을 제안한다. 변환 전후 프로그램 쌍의 의미적 동등성을 계산하는 하이브리드 구체-상징적 해석 방식을 도입하여, 프로그램 구현 세부사항에 크게 의존하지 않고 선형 시간 내에 동등성을 증명할 수 있다. 이를 위해 MLIR의 의미 있는 부분집합에 대한 검증기를 개발하고, AMD의 MLIR-AIR 및 MLIR-AIE 툴체인과 표준 mlir-opt에 대해 수백 개 벤치마크 변종을 검증한 결과를 보고한다. 이 연구는 MLIR 기반 컴파일러 최적화의 신뢰성 확보에 실질적 기여를 한다.
https://arxiv.org/abs/2605.01124
#mlir #formalverification #compileroptimization #programanalysis #symbolicexecution
Practical Formal Verification for MLIR Programs
Optimizing compilers have become a cornerstone for high-performance program generation in research and industry. Optimizations, including those implemented manually by a user and those target-specific and non-target-specific, are used to transform programs to achieve good performance. Although these optimizations are necessary for performance, assessing their correctness has remained a major challenge; the risk of incorrect code being deployed increases with unproven optimization flows. In this work, we target the formal verification of correctness of a transformed program by computing whether a pair of programs are semantically equivalent, one being a transformed version of the other. We restrict the class of programs supported to enable a hybrid concrete-symbolic interpretation approach to equivalence, which in turn is mostly agnostic to how the programs are implemented (syntax, schedule, storage, etc.). This approach can show equivalence in linear time with respect to the operations executed by the programs. We develop a verifier for a meaningful subset of MLIR, and report on the verification of the AMD MLIR-AIR and MLIR-AIE toolchains, as well as the standard mlir-opt on hundreds of benchmarks variants.
catatonicprime
Achim D. Brucker
hubertf
Andreas Klopsch
Michael
Ringzer0
Joxean Koret (@matalaz)