Предложена информационная технология мониторинга процессов загрязнения атмосферы химически взаимодействующими примесями с учетом аномальных процессов самоочищения при достижении предельной концентрации на базе гибридных высокопроизводительных вычислительных систем с использованием графических ускорителей архитектуры Nvidia CUDA.
Запропоновано інформаційну технологію моніторингу процесів забруднення атмосфери хімічно взаємодіючими домішками з урахуванням аномальних процесів самоочищення при досягненні граничної концентрації на базі гібридних високопродуктивних обчислювальних систем з використанням графічних прискорювачів архітектури Nvidia CUDA.
Introduction. The tasks of atmospheric monitoring are of great practical importance in the conditions of the constant increase of the intensity of industrial pollution and in emergency situations of technogenic origin. The processing of large volumes of data of various nature, the need to apply distribution models and the computational complexity of modeling in decision support systems determines the importance of the new information technologies creation. Purpose. The purpose of this research is to develop an information technology for monitoring the processes of atmospheric pollution by systems of chemically interacting contaminants, taking into account anomalous self-purification processes, based on hybrid high-performance computing systems with Nvidia CUDA accelerators. Methods. System analysis, methods of object-oriented design, parallel processing and parallel algorithms theory are used for information technology development. Results. The created information technology allows reduction by 30% of memory usage, 12 times reduction of modeling time, simple integration with the existing decision support systems and geo information systems (GIS) because of the standard OGC WMS, WFS interfaces usage. For computational infrastructure, it is possible to use the dedicated hybrid cluster, multi-user HPC resource with the dedicated resource manager, or the grid systems based on ARC, gLite or UMD middleware. For remote sensor implementation it is proposed to use low-cost internet-of-things solution based on ESP8266 microcontroller, which allows future reduction of total system cost (beyond lower FLOPS per watt value of GPGPU computing). Conclusion. The presented information technology can be used for real-time monitoring of atmosphere pollution processes, allows the integration of the existing decision support systems and GIS.
