2015/01/15 by A. V. Klimov, Klimov, Arkady
Computer Science · #Embedded Systems Design Techniques #FOS: Computer and information sciences #Formal Methods in Verification #Parallel Computing and Optimization Techniques #Programming Languages (cs.PL)
paper · pdf · doi:10.48550/arxiv.1501.03839
openalex publication_date 2015/01/15 · openalex created_date 2022/10/03 · openalex updated_date 2026/07/28
Exact polyhedral model (PM) can be built in the general case if the only\ncontrol structures are tt do-loops and structured tt ifs, and if loop\ncounter bounds, array subscripts and tt if-conditions are affine expressions\nof enclosing loop counters and possibly some integer constants. In more general\ndynamic control programs, where arbitrary tt ifs and tt whiles are\nallowed, in the general case the usual dataflow analysis can be only fuzzy.\nThis is not a problem when PM is used just for guiding the parallelizing\ntransformations, but is insufficient for transforming source programs to other\ncomputation models (CM) relying on the PM, such as our version of dataflow CM\nor the well-known KPN.\n The paper presents a novel way of building the exact polyhedral model and an\nextension of the concept of the exact PM, which allowed us to add in a natural\nway all the processing related to the data dependent conditions. Currently, in\nour system, only arbirary tt ifs (not tt whiles) are allowed in input\nprograms. The resulting polyhedral model can be easily put out as an equivalent\nprogram with the dataflow computation semantics.\n