Parallel processing in compute unfied device architecture (CUDA) for energy saving glass (E-Glass)

Khoo, Wen Xin (2014) Parallel processing in compute unfied device architecture (CUDA) for energy saving glass (E-Glass). Project Report. UTeM. (Submitted)

Text PARALLEL PROCESSING IN COMPUTE UNFIED DEVICE ARCHITECTURE (CUDA) FOR ENERGY SAVING GLASS (E-GLASS).pdf Download (91kB)

Abstract

Energy saving glass is used to keep the warmth and temperature if the building instead of using thermal radiator or machine to generate the heat all the time during winter. Yet the coating structure of the glass is mostly in regular shape such as tripoles and circular with very small size that limits the useful signal such as wireless signal and radio frequency to pass through. In order to develop a complex coating structure genetic algorithm technique is used. Yet genetic algorithm require a fast processing speed in order to cope with the process of creating new chromosome from the population and undergoes the selection, crossover and mutation operation processes. Hence parallel processing is need to overcome this problem with the use of both CPU and GPU to eliminate the need of purchasing high performance CPU and the needs of adding additional repeaters to increase the wireless signals. The coating structure will be presented in binary bits in a text file that shows best chromosome. The result will then be analyzed in a simulation tool that uses to check for the signal transmission efficiency and rate loss. Besides, speed test will be done to determine whether parallel processing can increase the speed in running the program code.

Actions (login required)

View Item