Raptor Code coding method, decoding method and systems based on CUDA (Compute Unified Device Architecture)

Abstract

The invention discloses a Raptor Code coding method, decoding method and systems based on CUDA (Compute Unified Device Architecture). According to the parallel characteristic of Raptor Code coding and decoding operation, a decomposition mode and an iterative data storage structure for a coding and decoding task in a parallel program and a specific operating process of each thread are designed. Thus, the executing speed of Raptor Code coding and decoding is improved, and the technical problems of long test cycle and low test efficiency when a Raptor Code coding and decoding algorithm is implemented at present are solved.

Description

technical field [0001] The invention belongs to the technical field of computer coding and coding, and more specifically, relates to a CUDA-based Raptor Code coding method, decoding method and system. Background technique [0002] Raptor Code (Rapid tornado Code), as a rateless code with excellent performance, has been applied more and more in communication systems. Raptor Code is a concatenated code composed of precoding and LT Code (Luby Transform Code), where precoding uses a high bit rate Low Density Parity Check Code (Low Density Parity Check Code, LDPC). The LDPC encoded data is generated according to the generator matrix, and its decoding uses the Belief Propagation (BP) algorithm; the LT Code encoding is performed on the LDPC encoded data, and its decoding also uses the BP algorithm. The BP decoding of LDPC is performed on the Tanner graph generated according to the check matrix, and the decoding process can be completed after 20 iterations, while the BP decoding of...

AI Technical Summary

Problems solved by technology

[0004] For the above defects or improvement needs of the prior art, the present invention provides a CUDA-based RaptorCode encoding method, decoding method and system, and according to the parallel characteristics of Raptor Code encoding and decoding operations, the decomposition of encoding and decoding tasks in parallel programs is designed method and iterative data storage structure, as well as the specific operation process of each thread, thereby improving the execution speed of Raptor Code encoding and decoding, thereby solving the problem of long test period and low test efficiency when implementing Raptor Code encoding and decoding algorithm in the prior art technical problem

Images