Incremental Banerjee test conditions committing for robust parallelization framework

Abstract

This paper describes the design of an automatic parallelization framework. The kernel supplied at itsfront end was suggested as an instrument for parallel potential assessment. It was used to measure the maximumachievable speedups in the major set of the CHStone benchmark suite programs. In such framework, we suggestedthe liberation of parallelism incrementally. We proposed a data dependency heuristic-based transformation method tomake true dependences dissociation. We generated an internal representation ( IR2), where the Banerjee test conditionsare met. Two among three of Banerjee test conditions came to be committed. In shared memory many/multicoreplatforms, the third condition could be satisfied by privatization. We would be able to choose the safe and the opportunepairwise (mapping-privatization) scheme among a number of threads mapping scenarios that become available in theIR2structure. Instrumentation on a subset of CHStone benchmark was carried out as a validity proof of our proposal,and the results confirmed that our framework kernel is robust.