Method and device for turbo decoder storage conflict resolution

Abstract

The invention relates to a method and a device for solving memory conflicts of a Turbo decoder. A device such as a turbo decoding device is proposed in which an intermediate buffer device including an address buffer device and an element buffer device is communicatively coupled to a plurality of processing devices and memory devices. During a cycle of the parallel decoding process, the intermediate buffer device receives first and second address information corresponding respectively to first and second elements of the code sequence stored in the memory device from two different processing devices. During the period, the intermediate buffer device sends a request for the first element to the memory device based on the first address information and stores the second address information in the address buffer device. Subsequently, during the period, the intermediate buffer device receives first element information corresponding to the first element from the memory device and stores the received first element information in the element buffer device.

Description

Background technique [0001] In a usual turbo decoder, a parallel decoding process can be used, where a non-permuted code block containing non-permuted symbols and a permuted code block containing permuted symbols can be divided into sub-blocks, stored in different memory banks and decoded in parallel using multiple processors. Non-permuted code blocks can be decoded in this way because the non-permuted symbols are stored in the memory banks in a linear order. [0002] However, since permutation symbols are generated on-the-fly so that corresponding non-permutation symbols are read from memory banks and placed into permutation locations as needed, it is possible to generate question. For example, when two or more symbols must be read from the same bank at the same time, collisions may arise, causing delays and reducing throughput. The number of collisions and the resulting reduction in throughput is usually high enough that the throughput requirements of the turbo decoder ca...

AI Technical Summary

Problems solved by technology

[0002] However, since permutation symbols are generated on-the-fly so that corresponding non-permutation symbols are read from memory banks and placed into permutation locations as needed, it is possible to generate question

For example, when two or more symbols must be read from the same bank at the same time, conflicts can arise, causing delays and reducing throughput

The number of collisions and the resulting reduction in throughput is usually high enough that the throughput requirements of the turbo decoder cannot be met

Common solutions include the use of specially designed hardware (e.g., a contention-free interleaver that may be part of the standard) or larger amounts of memory and processors, both of which have the advantage of improving the performance of turbo decoders. Undesirable Effects of Complexity and Cost

Images