Star simulator is an important aerospace simulation application, which produces real-time star imaging under any time and any attitude from the ground. The star image can be used in many devices, such as star sensor,an important instrument of attitude determination on satellite
that primarily uses star image for real-time attitude adjustment. In addition, a star simulator can also produce the celestial star background, widely used in space environment simulation systems.

Intensity model with blur effects are widely employed to accurately simulate the imaging process of a star simulator used for attitude determination and guiding feedback. The model is computationally intensive and the time requirements are proportional to the number of stars in the simulation, imposing great demands of computing power for realistic uses.

Our work includes: a) parallelization of intensity simulation on GPUs and firtsly a CUDA version of the star map simulator; b) adapting parallel simulator to defined problem characteristic and GPU architectural idiosyncrasies with on-chip memory redesign; c). strategies in achieving high performance of our simulators. d). a performance balance analysis to direct the choice of the best from two GPU simulators according to the simulation parameter.