Development of Parallel Applications Using DVM-system
DOI:
https://doi.org/10.14529/cmse190106Keywords:
automation the development of parallel programs, DVM-system, parallelism specification directives, accelerator, GPU, coprocessor, Fortran, CAbstract
DVM-system was designed to create parallel programs of scientific-technical computations in C-DVMH and Fortran-DVMH languages. These languages use the same model of parallel programming (DVMH-model) and are the extensions of standard C and Fortran languages by parallelism specifications, implemented as compiler directives. DVMH-model allows creating efficient parallel programs for heterogeneous computational clusters, the nodes of which use as computing devices not only universal multi-core processors but also can use attached accelerators (GPUs or Intel Xeon Phi coprocessors). This article describes the experience of parallelizing various application programs using DVM-system. The method of incremental or partial parallelization, the system's capabilities for working with unstructured grids, new tools for mapping MPI-programs to multi-core processors and accelerators are considered. The efficiency of parallel DVMH-programs on heterogeneous computing clusters K-10, K-100, Lomonosov and MVS-10P is investigated. The main advantages of DVM-approach for the development of parallel programs are described. The main features of DVM-system tools for performance analysis and functional debugging of parallel programs are presented. The directions for further development of DVM-system are determined.
References
Bakhtin V.A., Kolganov A.S., Krukov V.A., Podderugina N.V., Pritula M.N. Methods of Dynamic Tuning of DVMH Programs on Clusters with Accelerators. Superkompjuternye dni v Rossii: Trudy mezhdunarodnoj konferencii (Moskva, 28–29 sentjabrja 2015) [Russian Supercomputing Days: Proceedings of the International Scientific Conference (Moscow, Russia, September 28–29, 2015)]. Moscow, Publishing of the Moscow State University, 2015. pp. 257–268. (in Russian)
Aleksahin V.F., Bakhtin V.A., Kolganov A.S., Krukov V.A., Podderugina N.V., Pritula M.N., Savitskaya O.A., Shubert A.V., Zhukova O.F. GPU Parallelization of NAS NPB3.3.1 Benchmarks Using Fortran DVMH Programming Language. Vestnik Ufimskogo gosudarstvennogo aviacionnogo tehnicheskogo universiteta [Bulletin of the Ufa State Aviation Technical University]. Ufa, 2015. vol. 19, no. 67. pp. 240–250. (in Russian)
Aleksahin V.F., Bakhtin V.A., Kolganov A.S., Krukov V.A., Ostrovskaya I.P., Podderugina N.V., Pritula M.N., Savitskaya O.A., Zhukova O.F. Parallelization of NAS NPB3.3.1 Tests on Fortran-DVMH for Intel Xeon Phi Coproccessor. Parallelnye vychislitelnye tehnologii (PaVT'2015): trudy mezhdunarodnoj nauchnoj konferencii (Ekaterinburg, 31 marta–2 aprelja 2015) [Parallel computational technologies 2015: Proceedings of the International Scientific Conference (Ekaterinburg, Russia, March, 31–April, 2, 2015)]. Chelyabinsk, Publishing of the South Ural State University, 2015. pp. 19–30. (in Russian)
Bakhtin V.A., Klinov M.S., Krukov V.A., Podderugina N.V., Pritula M.N., Smirnov A.A. Usage of Fortran DVMH Language for Solving Hydrodynamics Problems on Hybrid Computing Systems. Vestnik Yuzho-Uralskogo gosudarstvennogo universiteta. Seriya "Vychislitelnaja matematika i informatika" [Bulletin of the South Ural State University. Computation Mathematics and Software Engineering]. Chelyabinsk, Publishing of the South Ural State University, 2013. vol. 2, no. 3. pp. 106–120. (in Russian). DOI: 10.14529/cmse130308.
Bakhtin V.A., Korolev A.V., Podderugina N.V. Parallel Computations for Compositional Flow Simulation During Oil and Gas Fields Development. Mezhdunarodnaja konferencija "Matematika i informacionnye tehnologii v neftegazovom komplekse": tezisy dokladov (Surgut, 16–20 maja 2016) [Mathematics and Informational Technologies for Oil and Gas Industry: Abstracts of the International Scientific Conference (Surgut, Russia, May, 16–20, 2016)]. Surgut: Publishing center SurSU, 2016. pp. 164–166. (in Russian)
Keldysh Institute of Applied Mathematics. Computing Resources. Available at: http://www.kiam.ru/MVS/resourses/ (accessed: 18.05.2018).
Joint Supercomputer Center of the Russian Academy of Sciences. Available at: http://www.jscc.ru/scomputers.html (accessed: 18.05.2018).
Supercomputing Center of Lomonosov Moscow State University. Available at: http://parallel.ru/cluster/ (accessed: 18.05.2018).
Bakhtin V.A., Kolganov A.S., Krukov V.A., Podderugina N.V., Polyakov S.V., Pritula M.N. DVMH Model Extension for Operation with Irregular Grids. Parallelnye vychislitelnye tehnologii (PaVT'2016): trudy mezhdunarodnoj nauchnoj konferencii (Arhangelsk, 28 marta–1 aprelja 2016) [Parallel computational technologies 2015: Proceedings of the International Scientific Conference (Archangelsk, Russia, March, 28–April, 1, 2016)]. Chelyabinsk, Publishing of the South Ural State University, 2016. pp. 757. (in Russian)
C-DVMH Language, C-DVMH Compiler, Compilation, Execution and Debugging of DVMH Programs. Available at: http://dvm-system.org/static_data/docs/CDVMHreference-en.pdf (accessed: 18.05.2018).
Fortran-DVMH Language. Fortran-DVMH Compiler. The Compilation, Execution and Debugging of DVMH-programs. Available at: http://dvm-system.org/static_data/docs/FDVMH-user-guide-en.pdf (accessed: 18.05.2018)
