Global scheduling independent of control dependencies based on condition vectors

Abstract
The authors present a global scheduling method based on condition vectors. The proposed method exploits global parallelism. The technique can schedule operations independent of control dependencies. It transforms the control structure of the given behavior drastically, while preserving semantics to minimize the number of states in final schedule. The method can parallelize multiple nests of conditional branches and optimize across the boundaries of basic blocks. It can also optimize all possible execution paths. An algorithm is proposed which generates a single finite state machine controller from parallel individual control sequences derived in the global parallelization process. Experimental results prove that the global parallelization is very effective.

This publication has 7 references indexed in Scilit: