GPU-based Polar code high-speed parallel decoding method

Abstract

The invention discloses a GPU-based Polar code high-speed parallel decoding method, the whole decoding process can be divided into three stages: an initialization stage, a decoding stage and a resultreturn stage, and the method specifically comprises the following steps: step 1, initializing a host; step 2, carrying out GPU initialization; 3, enabling the decoding kernel function to be subjectedto loop iteration multiple times, wherein the maximum number of loops is preset by a program; 4, for the No.0 threads of all the thread blocks of the factor graph p _good, performing inverse permutation on Local _ L [] [0] + Local _ R [] [0] in a shared memory of the No.0 threads to be as decoding results; and step 5, enabling the host to transmit a decoding result back to the host from the GPU. The method comprises three levels of parallelism, namely parallelism among multiple sub-graphs, parallelism among multiple thread blocks and parallelism among multiple threads. In addition, according to the method, the starting expenditure of the kernel function is reduced to the maximum extent; and the memory access efficiency and the running speed are improved.

Description

technical field [0001] The invention belongs to the technical field of communication, and relates to a high-speed parallel decoding method of Polar codes based on a GPU (Graphics Processing Unit, graphics processor). Background technique [0002] Polar code was proposed by Erdal Arikan in 2008 (Reference [1]: Erdal Arikan, "ChannelPolarization: A Method for Constructing Capacity-Achieving Codes", IEEEISIT2008), is currently the only channel coding method that can be strictly proven to reach the Shannon limit . Polar codes have been officially adopted by the 5G standardization organization. The decoding methods of Polar codes can be divided into two categories: methods based on serial cancellation and methods based on belief propagation. The calculation amount of the method based on serial cancellation is small, but the algorithm is serial in nature, so the decoding delay is relatively large; for the method based on belief propagation, in order to ensure the error correctio...

AI Technical Summary

Problems solved by technology

The calculation amount of the method based on serial cancellation is small, but the algorithm is serial in nature, so the decoding delay is relatively large; for the method based on belief propagation, in order to ensure the error correction performance of Polar code decoding, confidence Propagation list algorithm, that is, an iterative algorithm based on multiple replacement factor graphs, so this decoding method is computationally intensive, but the belief propagation list algorithm has the potential for parallel implementation

Images