350 rub
Journal Radioengineering №11 for 2014 г.
Article in number:
Implementation of Viterbi decoding on massively parallel computing devices
Authors:
A.V. Bashkirov - Ph.D. (Eng.), design and manufacture of radio, Voronezh State Technical University. E-mail: kipr@vorstu.ru
Y.S. Naumenko - post-graduate student, department of design and manufacture of radio, Voronezh State Technical University. E-mail: kipr@vorstu.ru
O.E. Sobolev - designer, JSC "Concern "Sozvezdie". E-mail: kipr@vorstu.ru
Yu.S. Sukhorukov - Dr.Sc. (Eng.), professor
Y.S. Naumenko - post-graduate student, department of design and manufacture of radio, Voronezh State Technical University. E-mail: kipr@vorstu.ru
O.E. Sobolev - designer, JSC "Concern "Sozvezdie". E-mail: kipr@vorstu.ru
Yu.S. Sukhorukov - Dr.Sc. (Eng.), professor
Abstract:
This paper deals with the problem of resource-modeling error correction coding / decoding, produced in the preliminary stages of the development of modern telecommunication systems.
Considerable time at this stage associated with a large number of simulations produced, due to the variation of parameters necessary codecs and statistical verification. As an approach to the problem of accelerating this process provides the use of parallel computing techniques on GPUs (GPGPU). Proposes the use of untapped resources of heterogeneous computer systems: in addition to CPU resource (central processing unit, CPU), GPU computing resources ( graphics processing unit, GPU).
A possible parallel implementation of Viterbi decoding, adapted for execution on massively parallel computing devices in accordance with the open standard OpenCL.
The results of application of this technique to demonstrate increase in computing speed using GPU resources with respect to classical simulation on the CPU.
Pages: 20-24
References
- Podval'ny'j S.L., Ledeneva T.M. Intellektual'ny'e sistemy' modelirovaniya: princzipy' razrabotki // Sistemy' upravleniya i informaczionny'e texnologii. 2013. T. 51. № 1. S. 4-10.
- Podval'ny'j S.L., Xolopkina L.V., Popov D.V. Chislenny'e metody' i vy'chislitel'ny'j e'ksperiment // Ufim. gos. aviacz. texn. un-t. 2005. 224 s.
- Bashkirov A.V., Klimov A.I., Muratov A.V., Naumenko Ju.S., Cy'mbalyuk V.S. Perspektivy' modelirovaniya parametrov algoritmov pomexoustojchivogo kodirovaniya s vy'sokoj stepen'yu parallelizma pri pomoshhi apparatnoj platformy' na baze GPU // Radiotexnika. 2013. № 12. S. 26-29.
- Naumenko Ju.S. Problemy' modelirovaniya pomexoustojchivy'x kodekov v geterogenny'x sistemax // Radiotexnika. 2014. № 3. S. 80-82.
- Naumenko Ju.S. Massivny'e parallel'ny'e vy'chisleniya v geterogenny'x sistemax pri modelirovanii nizkoplotnostny'x kodekov // Radiotexnika. 2014. № 6. S. 43-46.
- Boreskov A.V., Xarlamov A.A. Osnovy' raboty' s texnologiej CUDA. M.: DMK Press. 2010. 232 s. : il.
- Khronos OpenCL Working Group. "The OpenCL Specification". Version: 2.0. Document Revision : 19. 11/14/2013. 283 p.