Документ взят из кэша поисковой машины. Адрес оригинального документа : http://www.cosmos.ru/seminar/2014100103/presentation/Barash2.pdf Дата изменения: Thu Oct 16 20:13:29 2014 Дата индексирования: Sun Apr 10 11:58:13 2016 Кодировка: Поисковые слова: m 87 jet

Multi-GPU -.

.. , ..
. ..
, , 2 2014


- GPU.
1. PRAND RNGSSELIB , , (GM19, GM31, GM61, GM29.1, GM55.4, GQ58.1, GQ58.3, GQ58.4), MRG32K3A, LFSR113, MT19937. 2. : · · · · · ; 10^19 CPU SIMD- SSE-; , - , ; . ; .






L.Yu. B., Europhysics Letters (EPL) 95, 10003 (2011). L.Yu. B., L.N. Shchur, Phys.Rev. E 73 , 036701 (2006).


L.Yu. B., Applying dissipative dynamical systems to pseudorandom number generation: Equidistribution property and statistical independence of bits at distances up to logarithm of mesh size, Europhysics Letters (EPL) 95, 10003 (2011). L.Yu. B., Geometric and statistical properties of pseudorandom number generators based on multiple recursive transformations, Springer Proceedings in Mathematics and Statistics, Volume 23, p. 265-280 (2012). [book chapter in "Monte Carlo and Quasi-Monte Carlo Methods 2010", Ed. by Henryk Wozniakowski and Leszek Plaskota, Springer-Verlag, 2012. xii,732 pp, ISBN 978-3-64227439-8].



: : 1. : 2. : .


1. 2. N , k , k. 3. . 4. , . . 5. . 6. 2­5 M


1. k = T(1)/T(N) N, T(N) ­ ( N) N mpi. «» . «K-100» . .. . MPI CUDA.


2. k = T(1)/T(N) N, T(N) ­ ( N) N mpi. «» . MPI CUDA.


!


1. T. Hahn, CUA -- a library for multidimensional numerical integration, Computer Physics Communications 176 (2007) 712­713. 2. J. Kanzaki, Monte Carlo Integration in GPU, Eur. Phys. J. C (2011)71:1559. 3. L.Yu. Barash, L.N. Shchur, RNGSSELIB: Program library for random number eneration. More generators, parallel streams of random numbers and Fortran compatibility, Computer Physics Communications 184 (2013) 2367. 4. L.Yu. Barash, L.N. Shchur, PRAND: GPU accelerated parallel random number generation library, Computer Physics Communications 185 (2014) 1343. 5. .. , .. , -,