Video decoding technologies have been widely used in our daily life. Higher resolutions and more advanced coding technologies may promote the capabilities of video decoding. A new multi-core digital signal processing processor, ePUMA,which stands for embedded Parallel DSP platform with Unique Memory Access, is chosen to investigate how it supports video decoding.
This project aims to benchmark the algorithms of video decoding and evaluate the performance using ePUMA in MPEG-2 standard, which is a common standard with the purpose of compressing video signals. Based on the slice-parallelism methodology on eight co-processors of ePUMA, the implementation of the algorithms consists of variable length decoding, inverse scan, inverse quantization,two-dimensional inverse discrete cosine transform, motion vector decoding, form prediction and motion compensation.
The performance of the kernels is benchmarked by ePUMA system simulator. The result shows that to decode real-time Full HD (1920*1080 pixels, 30 frames per second) video, it will require ePUMA to run at 280 MHz for I frames and at 320MHz for P frames.
Source: Linköping University
Author: Xiaoyi, Peng