Scientific journal
Scientific Review. Technical science
ISSN 2500-0799
ПИ №ФС77-57440

Personal portfolio

PIPELINE-PARALLEL PROCESSING OF INTEGRATED DATA STREAMS

Tishchenko I.P. 1 Khachumov V.M. 1
1 Ailamazyan Program Systems Institute of the Russian Academy of Sciences
The objective of this research is an analysis of pipeline-parallel data streams processing methods and software tools for high-performance computing systems that equipped with general-purpose multi-core and graphics processors. The information about solutions and their capabilities is presented. Specifically, the perspective complex of intelligent integrated data processing (telemetry, target and command information) to space purposes earth station, that linking mission control center and spacecraft. GPU-algorithms library for a wide range of applications quickly solving are considered. Highlighted the advantages of architecture CUDA. The proposed cross-platform libraries greatly simplify the programming of modern computing systems, they provide scalability by increasing the number of computing cores and nodes, possibility to data streams pipeline-parallel processing, control and diagnostics problem solving.