作者: François Masdupuy
关键词: Abstract interpretation 、 Data structure 、 Tuple 、 Semantic analysis (machine learning) 、 Integer (computer science) 、 Algorithm 、 Vectorization (mathematics) 、 Dependence analysis 、 Interval (mathematics) 、 Theoretical computer science 、 Computer science
摘要: With the growing use of vector supercomputers, efficient and accurate data structure analyses are needed. What we propose in this paper is to quite general framework Cousot's abstract interpretation for particular analysis multi-dimensional array indexes. While such indexes integer tuples, a relational first required. This results combination existing ones that interval congruence based. Two orthogonal problems directly concerned with an analysis, parallelization/vectorization dependence locality problem used storage management. After introducing algorithm, describes on complete example how it order optimize storage.