mirror of
https://github.com/aicodix/code.git
synced 2026-04-27 14:30:36 +00:00
98 lines
3.5 KiB
C++
98 lines
3.5 KiB
C++
/*
|
|
Regression Test for the Bose Chaudhuri Hocquenghem Encoder and Decoder
|
|
|
|
Copyright 2018 Ahmet Inan <inan@aicodix.de>
|
|
*/
|
|
|
|
#include <cassert>
|
|
#include <random>
|
|
#include <iostream>
|
|
#include "galois_field.hh"
|
|
#include "bose_chaudhuri_hocquenghem_encoder.hh"
|
|
#include "bose_chaudhuri_hocquenghem_decoder.hh"
|
|
|
|
template <typename ENC, typename DEC>
|
|
void bch_reference_test(ENC *encode, DEC *decode, int trials)
|
|
{
|
|
std::random_device rd;
|
|
std::default_random_engine generator(rd());
|
|
typedef std::uniform_int_distribution<typename ENC::value_type> distribution;
|
|
auto rnd_cnt = std::bind(distribution(0, ENC::NR), generator);
|
|
auto rnd_len = std::bind(distribution(1, ENC::K), generator);
|
|
auto rnd_val = std::bind(distribution(0, 1), generator);
|
|
while (--trials) {
|
|
int data_len = rnd_len();
|
|
auto rnd_pos = std::bind(distribution(0, data_len + ENC::NP - 1), generator);
|
|
typename ENC::value_type data[data_len], orig_data[data_len];
|
|
for (int i = 0; i < data_len; ++i)
|
|
data[i] = orig_data[i] = rnd_val();
|
|
typename ENC::value_type parity[ENC::NP];
|
|
(*encode)(reinterpret_cast<typename ENC::ValueType *>(data), reinterpret_cast<typename ENC::ValueType *>(parity), data_len);
|
|
for (int i = 0; i < data_len; ++i)
|
|
assert(data[i] == orig_data[i]);
|
|
typename ENC::value_type orig_parity[ENC::NP];
|
|
for (int i = 0; i < ENC::NP; ++i)
|
|
orig_parity[i] = parity[i];
|
|
int error_count = rnd_cnt() / 2;
|
|
typename ENC::value_type errors[error_count];
|
|
for (int i = 0; i < error_count; ++i) {
|
|
int pos = rnd_pos();
|
|
for (int j = 0; j < i; ++j) {
|
|
if (errors[j] == pos) {
|
|
pos = rnd_pos();
|
|
j = -1;
|
|
}
|
|
}
|
|
errors[i] = pos;
|
|
if (pos < data_len)
|
|
data[pos] ^= 1;
|
|
else
|
|
parity[pos-data_len] ^= 1;
|
|
}
|
|
int erasures_count = ENC::NR - 2 * error_count;
|
|
typename ENC::value_type erasures[erasures_count];
|
|
for (int i = 0; i < erasures_count; ++i) {
|
|
int pos = rnd_pos();
|
|
for (int j = 0; j < i; ++j) {
|
|
if (erasures[j] == pos) {
|
|
pos = rnd_pos();
|
|
j = -1;
|
|
}
|
|
}
|
|
erasures[i] = pos;
|
|
if (pos < data_len)
|
|
data[pos] ^= 1;
|
|
else
|
|
parity[pos-data_len] ^= 1;
|
|
}
|
|
int ret = (*decode)(reinterpret_cast<typename ENC::ValueType *>(data), reinterpret_cast<typename ENC::ValueType *>(parity), reinterpret_cast<typename ENC::IndexType *>(erasures), erasures_count, data_len);
|
|
assert(ret >= 0);
|
|
for (int i = 0; i < data_len; ++i)
|
|
assert(data[i] == orig_data[i]);
|
|
for (int i = 0; i < ENC::NP; ++i)
|
|
assert(parity[i] == orig_parity[i]);
|
|
}
|
|
}
|
|
|
|
int main()
|
|
{
|
|
if (1) {
|
|
// NASA INTRO BCH(15, 5) T=3
|
|
typedef CODE::GaloisField<4, 0b10011, uint8_t> GF;
|
|
GF instance;
|
|
CODE::BoseChaudhuriHocquenghemEncoderReference<6, 1, 5, GF> encoder({0b10011, 0b11111, 0b00111});
|
|
CODE::BoseChaudhuriHocquenghemDecoderReference<6, 1, 5, GF> decoder;
|
|
bch_reference_test(&encoder, &decoder, 1000000);
|
|
}
|
|
if (1) {
|
|
// DVB-S2 FULL BCH(65535, 65343) T=12
|
|
typedef CODE::GaloisField<16, 0b10000000000101101, uint16_t> GF;
|
|
GF instance;
|
|
CODE::BoseChaudhuriHocquenghemEncoderReference<24, 1, 65343, GF> encoder({0b10000000000101101, 0b10000000101110011, 0b10000111110111101, 0b10101101001010101, 0b10001111100101111, 0b11111011110110101, 0b11010111101100101, 0b10111001101100111, 0b10000111010100001, 0b10111010110100111, 0b10011101000101101, 0b10001101011100011});
|
|
CODE::BoseChaudhuriHocquenghemDecoderReference<24, 1, 65343, GF> decoder;
|
|
bch_reference_test(&encoder, &decoder, 100);
|
|
}
|
|
std::cerr << "Bose Chaudhuri Hocquenghem regression test passed!" << std::endl;
|
|
return 0;
|
|
}
|
|
|