Color Space Conversion Algorithm Optimization Based on Loongson Multimedia Instructions

SIMD (Single Instruction/Multiple Data) technology can significantly improve the performance of computationally intensive applications, but the performance gain ratio is not only related to the original algorithm itself, but also depends on the specific hardware environment. In order to tap the hardware potential of the processor and solve the performance bottleneck of the color space conversion link of the video playback of the Loongson platform browser, this paper first analyzes the algorithm characteristics and the optimization of other architectures; then summarizes and reconstructs the algorithm with the help of the architectural characteristics of the Loongson multimedia vector instructions, and proposes methods of setting register cache, peephole optimization and making full use of register bit width to improve parallelism; finally, a comparative experiment is designed to test the performance data of the three architectures before and after optimization. The experimental results show that the method proposed in this paper can make up for the shortcomings of the hardware to a certain extent, and can achieve a 50% performance gain even when the vector register bit width differs by one time; the relevant results are applied to the chrome browser, which can meet the smooth playback of 1080P 60 fps video sources, and the relevant work results have been applied to actual products.