文件列表:

LICENSE (18047, 2015-09-28)
src (0, 2015-09-28)
src\CPU_decode.cpp (16469, 2015-09-28)
src\CPU_decode.cpp.bak (16469, 2015-09-28)
src\CPU_decode.h (813, 2015-09-28)
src\GPU_decode_10240.cuh (161280, 2015-09-28)
src\GPU_decode_7168.cuh (131772, 2015-09-28)
src\Makefile (494, 2015-09-28)
src\cuda_helper.cuh (7360, 2015-09-28)
src\kernel.cu (17173, 2015-09-28)
src\linux.h (2010, 2015-09-28)
src\randn.h (1909, 2015-09-28)
src\totalDefine.h (2644, 2015-09-28)

# LDPC-CC_Pipeline_Decoder Copyright (c) 2014-2015 Mokyy and Haybla. All rights reserved. This is the source codes of a CUDA application named LDPC-CC_Pipeline_Decoder. Techniques of design and optimization can be found in the article "High Throughput Pipeline Decoder for LDPC Convolutional Codes on GPU" published in the IEEE Communications Letters journal (The article was accepted just now). In order to compile this project, Linux OS is recommended. Just open a terminal and go into the "src" directory. >cd src Then compile all the codes using "make". >make An executed file named "decode" will be generated if everything goes right. Just execute the "decode" as follows. >./decode And you will see some information about decoding times and throughputs printed in the terminal. You can clean all compilation using "make clean". >make clean #Besides Macro definitions are written in the file named "totalDefine.h". You can enable/disable some of them to change the mode of the decoder. Descriptions are as follows. CODE1: represents the LDPC-CC constructed by (4096, 10240) LDPC codes in CCSDS standard. CODE2: represents the LDPC-CC constructed by (4096, 7168) LDPC codes in CCSDS standard. LINUX: once defined, the project must be compiled and run on Linux OS. TEST_PERF once defined, the project can simulate BER&FER performance of the LDPC-CC.

近期下载者：

相关文件：

评论：[我要评论] [举报此文件]

收藏者：