Simulation of Multimedia Data Parallelization Process when Processing on CPU and GPU in Online Learning Cloud Platforms

Autor: Denis Parfenov, Vladimir Shardakov, Veronika Zaporozhko
Rok vydání: 2019
Předmět:
Zdroj: 2019 International Russian Automation Conference (RusAutoCon).
DOI: 10.1109/rusautocon.2019.8867796
Popis: The article describes the principle of operation of a parallel algorithm for generating terrain (landscape) and the superposition method based on the architecture of computing unified devices. The presented solutions were implemented in the design and creation of a software module for the implementation and research of virtual models of environmental objects in online courses. An innovative approach is proposed, which allows solving the problem of efficient distribution and use of computing resources in processing large streams of multimedia data by reducing the load on the central processor, video and RAM. A model of the process of multimedia data parallelization is presented when processing on a CPU and GPU in cloud-based online learning platforms. The result of visualization of a three-dimensional landscape obtained as a result of the multimedia data parallelization process during its generation is presented. It was concluded that the developed model and the algorithm are capable of conducting high-quality modeling of realistic three-dimensional images of various objects with a significant reduction in the requirements for the productive capacity and computer resources of the student.
Databáze: OpenAIRE