Study the performance of the DVMH-programs on clusters with multicore processors and accelerators


V.A. Bakhtin, N.A. Kataev, A.S. Kolganov, V.A. Krukov, N.V. Podderugina, M.N. Pritula, A.A. Smirnov


The report analyzes the performance  of Fortran DVMH (FDVMH) programs on different architecture clusters  (using universal multcore processors, NVIDIA graphics accelerators and Intel Xeon Phi coprocessors in their nodes) . The characteristics of the programs written on FDVMH and their implementations on other languages are compared. The  effect of  various possible optimizations on the performance of DVMH-programs when mapping  on the cluster is studied.