Методы распределения нагрузки на многопроцессорную систему

Approaches to distribute load among different processors of a homogeneous computer system are widespread and are efficiently implemented via parallel programming techniques such as OpenMP, however, analogous approaches, which take into account not only performance of a processor but also performance of other devices of a computer system involved in a computation (such as GPUs and storage devices), are used in isolation and are not general. General approach was implemented and through a series of experiments it was shown that it exhibits high performance in a problem of wavy ocean surface generation. This approach led to program structure being simplified as parallel tasks and pipeline links were factorised into abstract unified components.

• Git
• Full text

@article{gankevich2014meto,
  title={Методы распределения нагрузки на многопроцессорную систему},
  author={И. Г. Ганкевич and А. Б. Дегтярев},
  journal={Процессы управления и устойчивость},
  year={2014},
  month={01},
  volume={1},
  pages={295--300},
  number={17},
  language={russian},
  type={article}
}