Informatics and Applications
2015, Volume 9, Issue 4, pp 29-36
DATAFLOW COMPUTING MODEL AS A PARADIGM OF FUTURE MAINSTREAM OF SOFTWARE DEVELOPMENT
- D. N. Zmejev
- A. V. Klimov
- N. N. Levchenko
- A. S. Okunev
- A. L. Stempkovsky
Abstract
The article describes the parallel dataflow computing system with dynamic forming of context and the architectural features of its implementation. The dataflow computing model allows to solve the problems that arise when creating and using supercomputers. One of these problems is the difficulty of loading the rising number of functional units of cores, while remaining in the borders of traditional programming. The article describes
the advantages of the proposed computing model. The article compares the paradigm of "distribution" with the traditional paradigm of "collection." The dataflow computing model works in the paradigm of "distribution." The description of basic features of the architecture of the "Buran" parallel dataflow computing system and its differences from the classical flow computing systems are given. The studies give hope that the proposed computing model will become in the future the main programming paradigm for large-scale parallel computing.
[+] References (18)
- Klimov, A.V., N. N. Levchenko, and A. S. Okunev. 2012. Preimushchestva potokovoy modeli vychisleniy v usloviyakh neodnorodnykh setey [The advantages of dataflow calculation model in heterogeneous networks conditions]. Zh. Informatsionnye Tekhnologii i Vychisli- tel'nye Sistemy [J. Information Technologies Computing Systems] 2:36-45.
- Levchenko, N. N., A. S. Okunev, and A. L. Stempkovskiy. 2012. Ispol'zovanie modeli vychisleniy s upravleniem po- tokom dannykh i realizuyushchey ee arkhitektury dlya sis- tem eksaflopsnogo urovnya proizvoditel'nosti [Dataflow calculation model and its architecture usage for exa- flop's computing systems]. Vseross. nauch.-tekhnich. konf. "Problemy Razrabotki Perspektivnykh Mikro- i Nanoelek- tronnykh Sistem" (MES-2012): Sb. tr. [5th All-Russia Scientific and Technical Conference "Problems of Advanced Micro- and Nanoelectronic Systems Development" (MES-2012) Proceedings]. Moscow. 459-462.
- Klimov, A. V., N. N. Levchenko, A. S. Okunev, and A. L. Stempkovskiy. 2014. Superkomp'yutery, ierarkhiya pamyati i potokovaya model' vychisleniy [Supercomputers, memory hierarchy, and the dataflow computing system]. Programmnye Sistemy: Teoriya i Prilozheniya. Elektron. Nauchn. Zh. [Program Systems: Theory and Applications. Electron. Sci. J.] 5:1(19):15-36.
- Stempkovskiy, A.L., N. N. Levchenko, A. S. Okunev, and V. V. Tsvetkov. 2008. Parallel'naya potokovaya vy- chislitel'naya sistema - dal'neyshee razvitie arkhitektury
i strukturnoy organizatsii vychislitel'noy sistemy s av- tomaticheskim raspredeleniem resursov [Parallel dataflow computing system: Further development of architecture and structural organization ofthe computing system with automatic distribution of resources]. Zh. Informatsionnye Tekhnologii [J. Inform. Technol.] 10:2-7.
- Silc, J., B. Robic, and T. Ungerer. 1998. Asynchrony in parallel computing: From dataflow to multithreading. Parallel Distributed Computing Practices 1(1):3-30.
- Stempkovskiy, A. L., A. V. Klimov, N. N. Levchenko, and A. S. Okunev. 2009. Metody adaptatsii parallel'noy po- tokovoy vychislitel'noy sistemy pod zadachi otdel'nykh klassov [Methods of parallel dataflow computing system adaptation for problems of individual classes]. Zh. Infor-matsionnye Tekhnologii i Vychislitel'nye Sistemy [J. Inform. Technol. Comput. Syst.] 3:12-21.
- Stempkovskiy, A. L., N. N. Levchenko, and A. S. Okunev.
2010. Arkhitektura sverkhpetaflopnoy vychislitel'noy sis-temy s vysokoy real'noy proizvoditel'nost'yu, bazi- ruyushcheysya na netraditsionnoy modeli vychisleniy
[Abovepetaflop's computer systems architecture with high real performance, based on nontraditional computing model]. Mat-ly nauchn. konf. "Rezul'taty Tselevykh Ori- entirovannykh Fundamental'nykh Issledovaniy i Ikh Is-pol'zovanie v Rossiyskoy Promyshlennosti" [Annual Scientific Conference "Results of the Target-Oriented Fundamental Researches and Their Usage in Russian Industry" Proceedings]. Taganrog, Russia. 68-72.
- Klimov, A. V., N. N. Levchenko, A. S. Okunev, and A. L. Stempkovskiy. 2011. Avtomaticheskoe rasparalle- livanie dlya gibridnoy sistemy s potokovym uskoritelem [Automatic parallelization of sequential programs for hybrid systems with accelerator-based dataflow]. Zh. Informatsionnye Tekhnologii i Vychislitel'nye Sistemy [J. Inform. Technol. Comput. Syst.] 2:3-11.
- Zmeev, D.N., N. N. Levchenko, A. S. Okunev, and A. V. Klimov. 2014. Sposoby regulirovaniya vychisleniy v parallel'noy potokovoy vychislitel'noy sisteme [Methods of control of computations in the parallel dataflow computing system]. Problemy Razrabotki Perspektivnykh Mikro- i Nanoelektronnykh Sistem-2014: Sb. tr. [All- Russia Scientific and Technical Conference "Problems of Advanced Micro- and Nanoelectronic Sys- temsDevelopment-2014" Proceedings]. Ed. A. L. Stempkovskiy. Moscow. IV:79-82.
- Zmeev, D. N., N. N. Levchenko, A. S. Okunev, and A. V. Klimov. 2014. Arkhitektura planirovshchika pro- tsessora sopostavleniya PPVS "Buran" [Architecture of the scheduler of mapping processor of PDCS "Buran"]. Problemy Razrabotki Perspektivnykh Mikro- i Nanoelektronnykh Sistem-2014. Sb. tr. [All-Russia Scientific and Technical Conference "Problems of Advanced Micro- and Nanoelectronic Systems Development-2014" Proceedings]. Ed. A. L. Stempkovskiy. Moscow. IV:75-78.
- Levchenko, N. N., A. S. Okunev, and A. V. Klimov. 2009. Realizatsiya zadachi "Povyshenie kontrastnosti izobra- zheniya" na parallel'noy potokovoy vychislitel'noy sis- teme [Realization of the task of "Improving the contrast of the image" on the parallel dataflow computing system]. Scientific and Technical Conference (International) "Mul-tiprocessor Computing and Control Systems (MCCS-2009) " Proceedings. Divnomorskoe, Russia. 1:197-200.
- Levchenko, N. N., and A. S. Okunev. 2009. Analiz prokhozhdeniya zadachi "obnaruzhenie defektov" na pa-rallel'noy potokovoy vychislitel'noy sisteme [Analysis of the running of defects detection tasks on the parallel dataflow computing system]. Tr. III Vseross. nauchn. konf. Metody i Sredstva Obrabotki Informatsii [3rd All-Russia Scientific Conference "Methods and Tools of Information Processing" Proceedings]. Moscow. 346-350.
- Levchenko, N. N., and A. S. Okunev. 2010. Spetsializa- tsiya arkhitektury mnogoyadernoy parallel'noy potokovoy vychislitel'noy sistemy dlya resheniya zadachi bystrogo preobrazovaniya Fur'e [Specialization of the architecture of multicore parallel dataflow computing system to compute the task of Fast Fourier Transform]. Sb. nauch. tr. IV Vseross. nauchn.-tekhnich. konf. "Proble- my Razrabotki Perspektivnykh Mikro- i Nanoelektronnykh Sistem-2010 [4th All-Russia Scientific and Technical Conference "Problems of Advanced Micro- and Na- noelectronic Systems Development-2010" Proceedings]. Moscow. 458-461.
- Klimov, A. V., N. N. Levchenko, A. S. Okunev,
D. N. Zmeev, and A. L. Stempkovskiy. 2014. Issledovanie vozmozhnosti asinkhronnoy realizatsii zadachi moleku- lyarnoy dinamiki na PPVS "Buran" [The research of the asynchronous implementation ofmolecular dynamics on PDCS "Buran"]. Zh. Kachestvo. Innovatsii. Obrazovanie [J. Quality. Innovation. Education] 10:46-51.
- Zmeev, D. N., A. V. Klimov, N. N. Levchenko, and A. S. Okunev. 2014. Primenenie parallel'noy potokovoy vychislitel'noy sistemy "Buran" dlya zadach molekulyar- noy dinamiki [Application of the parallel dataflow computing system "Buran" for molecular dynamics]. Mat-ly 3-y Vseross. nauchn.-tekhnich. konf. "Superkomp'yuternye Tekhnologii" (SKT-2014) [3rd All-Russia Scientific and
Technical Conference "Supercomputer Technologies" (SCT-2014) Proceedings] 2:222-224.
- Levchenko, N. N., A. S. Okunev, and D. N. Zmeev. 2008. Rasshirenie vozmozhnostey povedencheskoy blochno- registrovoy modeli parallel'noy potokovoy vychislitel'noy sistemy [Enhancement of capability of the behavioral block-rated model of the parallel dataflow computing system]. 5th Youth Research and Technical Conference (International) "High-Performance Computing Systems (HPCS 2008)" Proceedings. Taganrog, Russia. 371-374.
- Zmeev, D. N., A. S. Okunev, N. N. Levchenko, and A. V. Klimov. 2013. Realizatsiya parallel'noy modeli vy- chisleniy s upravleniem potokom dannykh na klasternykh superkomp'yuterakh [The implementation of parallel computing model with dataflow control management on cluster supercomputers]. Supercomuter Conference (Inter-national) "Scientific Service in the Internet: All Sides of Parallelism"Proceedings. Moscow: MSU. 375-377.
- Zmeev, D. N., N. N. Levchenko, and A. S. Okunev. 2014. Issledovanie printsipov raboty parallel'noy potokovoy vy-chislitel'noy sistemy na klasternykh vychislitel'nykh siste- makh [Study of the principles of parallel dataflow computing system on the cluster computing systems]. Mat-ly
3- y Vseross. nauchn.-tekhnich. konf. "Superkomp'yuternye Tekhnologii" (SKT-2014) [3rd All-Russia Scientific and Technical Conference "Supercomputer Technologies" (SCT-2014) Proceedings]. 1:45-48.
[+] About this article
Title
DATAFLOW COMPUTING MODEL AS A PARADIGM OF FUTURE MAINSTREAM OF SOFTWARE DEVELOPMENT
Journal
Informatics and Applications
2015, Volume 9, Issue 4, pp 29-36
Cover Date
2015-11-30
DOI
10.14357/19922264150403
Print ISSN
1992-2264
Publisher
Institute of Informatics Problems, Russian Academy of Sciences
Additional Links
Key words
dataflow computing system; paradigm of distribution; content-addressable memory; localization of computation
Authors
D. N. Zmejev , A. V. Klimov ,
N. N. Levchenko , A. S. Okunev , and A. L. Stempkovsky
Author Affiliations
Institute for Design Problems in Microelectronics, Russian Academy of Sciences, 3 Sovetskaya Str., Moscow 124365, Russian Federation
|