"An approach to the programming of supercomputers on the basis of multicore multithreaded chips"
Korneev V.V.

An approach to the development of parallel programs adapted to the architecture of supercomputers is proposed on the basis of multicore multithreaded chips. This approach is directed to overcome the negative effect of memory latency on the efficiency of parallel programs and can be considered as a method to develop scalable parallel programs.

Keywords: architecture of supercomputers based on multicore multithreaded chips, scalable parallel programs, efficiency of parallel computing

Korneev V.V.     e-mail: korv@rdi-kvant.ru