Optimization Problem for Heterogeneous Computing Systems

The article is devoted to the problem of parallel code optimization for execution in heterogeneous computing systems. The difficulty of optimizing is associated with the difference in parallelization technologies for different types of heterogeneous systems, as well as with the difference in the command system for nodes within the system. The article proposes an approach to searching the optimal variant of parallelizing program code in heterogeneous systems taking into account the specifics of the architecture of the system and the technology used.