Chapter 38. Parallel LDPC Decoding
Gabriel Falcao, Vitor Silva and Leonel Sousa
In this chapter we present a pair of kernels to decode a class of powerful error-correcting codes, known as low-density parity-check (LDPC) codes, on graphics processing units (GPU). The proposed parallel implementations adopt a compact data structures representation to access data, while the processing on the GPU grid efficiently exploits a balanced distribution of the computational workload. Moreover, we have analyzed different levels of thread coarsening and propose an efficient one, which balances computation and memory accesses. This case study shows that by adopting these practical techniques, it is possible to use the GPU's computational power to perform this ...
Get GPU Computing Gems Emerald Edition now with the O’Reilly learning platform.
O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.
