LDPCStorage.jl

Reads and writes file formats for storing sparse matrices containing only zeros and ones. Intended for use with low density parity check (LDPC) matrices. Also supports efficient storage for quasi-cyclic LDPC codes.

Installation

Run Julia, enter ] to bring up Julia's package manager, and add the package:

julia> ]
(v1.9) pkg> add LDPCStorage

Supported File Formats

• alist (by David MacKay et al., see http://www.inference.org.uk/mackay/codes/alist.html)
• cscmat (our custom format) DEPRECATED
• bincsc.json (Based on compressed sparse column (CSC). Valid json.)
• qccsc.json (Based on compressed sparse column (CSC). Valid json. Store exponents of quasi-cyclic LDPC matrices)
• hpp (C++ header) CSC of matrix as static data (write-only, reading not supported!)

How to use

using SparseArrays
using LDPCStorage

H = sparse(Int8[
        0 0 1 1 0 0 0 0 1 0 0 1 1 0
        1 0 0 1 1 0 0 0 0 0 1 0 0 1
        0 1 0 1 0 1 1 0 1 0 0 1 1 0
        1 0 0 1 0 0 0 1 0 1 0 1 0 1
    ])

save_to_alist("./ldpc.alist", H)
H_alist = load_alist("./ldpc.alist")
H == H_alist || warn("Failure")

save_to_bincscjson("./ldpc.bincsc.json", H)
H_csc = load_ldpc_from_json("./ldpc.bincsc.json")
H == H_csc || warn("Failure")

open("./autogen_ldpc.hpp", "w+") do io
    print_cpp_header(io, H)
end

There also are methods accepting an IO object: print_alist, print_bincscjson, etc.

Some methods support outputting quasi-cyclic exponents directly, e.g., print_cpp_header_QC outputs a C++ header.

Contributing

Contributions, feature requests and suggestions are welcome. Open an issue or contact us directly.