mirror of
https://github.com/f4exb/sdrangel.git
synced 2024-11-03 07:21:14 -05:00
417 lines
8.8 KiB
C++
417 lines
8.8 KiB
C++
/*
|
|
SIMD-ified LDPC algorithms
|
|
|
|
Copyright 2018 Ahmet Inan <xdsopl@gmail.com>
|
|
*/
|
|
|
|
#ifndef ALGORITHMS_HH
|
|
#define ALGORITHMS_HH
|
|
|
|
#include "generic.h"
|
|
#include "exclusive_reduce.h"
|
|
|
|
namespace ldpctool {
|
|
|
|
template <typename VALUE, int WIDTH>
|
|
struct SelfCorrectedUpdate<SIMD<VALUE, WIDTH>>
|
|
{
|
|
typedef SIMD<VALUE, WIDTH> TYPE;
|
|
static void update(TYPE *a, TYPE b)
|
|
{
|
|
*a = vreinterpret<TYPE>(vand(vmask(b), vorr(vceqz(*a), veor(vcgtz(*a), vcltz(b)))));
|
|
}
|
|
};
|
|
|
|
template <typename VALUE, int WIDTH, typename UPDATE>
|
|
struct MinSumAlgorithm<SIMD<VALUE, WIDTH>, UPDATE>
|
|
{
|
|
typedef SIMD<VALUE, WIDTH> TYPE;
|
|
static TYPE zero()
|
|
{
|
|
return vzero<TYPE>();
|
|
}
|
|
static TYPE one()
|
|
{
|
|
return vdup<TYPE>(1);
|
|
}
|
|
static TYPE min(TYPE a, TYPE b)
|
|
{
|
|
return vmin(a, b);
|
|
}
|
|
static TYPE sign(TYPE a, TYPE b)
|
|
{
|
|
return vsign(a, b);
|
|
}
|
|
static void finalp(TYPE *links, int cnt)
|
|
{
|
|
TYPE mags[cnt], mins[cnt];
|
|
for (int i = 0; i < cnt; ++i)
|
|
mags[i] = vabs(links[i]);
|
|
CODE::exclusive_reduce(mags, mins, cnt, min);
|
|
|
|
TYPE signs[cnt];
|
|
CODE::exclusive_reduce(links, signs, cnt, sign);
|
|
|
|
for (int i = 0; i < cnt; ++i)
|
|
links[i] = sign(mins[i], signs[i]);
|
|
}
|
|
static TYPE add(TYPE a, TYPE b)
|
|
{
|
|
return vadd(a, b);
|
|
}
|
|
static TYPE sub(TYPE a, TYPE b)
|
|
{
|
|
return vsub(a, b);
|
|
}
|
|
static bool bad(TYPE v, int blocks)
|
|
{
|
|
auto tmp = vcgtz(v);
|
|
for (int i = 0; i < blocks; ++i)
|
|
if (!tmp.v[i])
|
|
return true;
|
|
return false;
|
|
}
|
|
static void update(TYPE *a, TYPE b)
|
|
{
|
|
UPDATE::update(a, b);
|
|
}
|
|
};
|
|
|
|
template <int WIDTH, typename UPDATE>
|
|
struct MinSumAlgorithm<SIMD<int8_t, WIDTH>, UPDATE>
|
|
{
|
|
typedef int8_t VALUE;
|
|
typedef SIMD<VALUE, WIDTH> TYPE;
|
|
static TYPE zero()
|
|
{
|
|
return vzero<TYPE>();
|
|
}
|
|
static TYPE one()
|
|
{
|
|
return vdup<TYPE>(1);
|
|
}
|
|
static TYPE sign(TYPE a, TYPE b)
|
|
{
|
|
return vsign(a, b);
|
|
}
|
|
static TYPE eor(TYPE a, TYPE b)
|
|
{
|
|
return vreinterpret<TYPE>(veor(vmask(a), vmask(b)));
|
|
}
|
|
static TYPE orr(TYPE a, TYPE b)
|
|
{
|
|
return vreinterpret<TYPE>(vorr(vmask(a), vmask(b)));
|
|
}
|
|
static TYPE other(TYPE a, TYPE b, TYPE c)
|
|
{
|
|
return vreinterpret<TYPE>(vbsl(vceq(a, b), vmask(c), vmask(b)));
|
|
}
|
|
static void finalp(TYPE *links, int cnt)
|
|
{
|
|
TYPE mags[cnt];
|
|
for (int i = 0; i < cnt; ++i)
|
|
mags[i] = vqabs(links[i]);
|
|
|
|
TYPE mins[2];
|
|
mins[0] = vmin(mags[0], mags[1]);
|
|
mins[1] = vmax(mags[0], mags[1]);
|
|
for (int i = 2; i < cnt; ++i) {
|
|
mins[1] = vmin(mins[1], vmax(mins[0], mags[i]));
|
|
mins[0] = vmin(mins[0], mags[i]);
|
|
}
|
|
|
|
TYPE signs = links[0];
|
|
for (int i = 1; i < cnt; ++i)
|
|
signs = eor(signs, links[i]);
|
|
|
|
for (int i = 0; i < cnt; ++i)
|
|
links[i] = sign(other(mags[i], mins[0], mins[1]), orr(eor(signs, links[i]), vdup<TYPE>(127)));
|
|
}
|
|
static TYPE add(TYPE a, TYPE b)
|
|
{
|
|
return vqadd(a, b);
|
|
}
|
|
static TYPE sub(TYPE a, TYPE b)
|
|
{
|
|
return vqsub(a, b);
|
|
}
|
|
static bool bad(TYPE v, int blocks)
|
|
{
|
|
auto tmp = vcgtz(v);
|
|
for (int i = 0; i < blocks; ++i)
|
|
if (!tmp.v[i])
|
|
return true;
|
|
return false;
|
|
}
|
|
static void update(TYPE *a, TYPE b)
|
|
{
|
|
UPDATE::update(a, vmin(vmax(b, vdup<TYPE>(-32)), vdup<TYPE>(31)));
|
|
}
|
|
};
|
|
|
|
template <typename VALUE, int WIDTH, typename UPDATE, int FACTOR>
|
|
struct OffsetMinSumAlgorithm<SIMD<VALUE, WIDTH>, UPDATE, FACTOR>
|
|
{
|
|
typedef SIMD<VALUE, WIDTH> TYPE;
|
|
static TYPE zero()
|
|
{
|
|
return vzero<TYPE>();
|
|
}
|
|
static TYPE one()
|
|
{
|
|
return vdup<TYPE>(1);
|
|
}
|
|
static TYPE min(TYPE a, TYPE b)
|
|
{
|
|
return vmin(a, b);
|
|
}
|
|
static TYPE sign(TYPE a, TYPE b)
|
|
{
|
|
return vsign(a, b);
|
|
}
|
|
static void finalp(TYPE *links, int cnt)
|
|
{
|
|
TYPE beta = vdup<TYPE>(0.5 * FACTOR);
|
|
TYPE mags[cnt], mins[cnt];
|
|
for (int i = 0; i < cnt; ++i)
|
|
mags[i] = vmax(vsub(vabs(links[i]), beta), vzero<TYPE>());
|
|
CODE::exclusive_reduce(mags, mins, cnt, min);
|
|
|
|
TYPE signs[cnt];
|
|
CODE::exclusive_reduce(links, signs, cnt, sign);
|
|
|
|
for (int i = 0; i < cnt; ++i)
|
|
links[i] = sign(mins[i], signs[i]);
|
|
}
|
|
static TYPE add(TYPE a, TYPE b)
|
|
{
|
|
return vadd(a, b);
|
|
}
|
|
static TYPE sub(TYPE a, TYPE b)
|
|
{
|
|
return vsub(a, b);
|
|
}
|
|
static bool bad(TYPE v, int blocks)
|
|
{
|
|
auto tmp = vcgtz(v);
|
|
for (int i = 0; i < blocks; ++i)
|
|
if (!tmp.v[i])
|
|
return true;
|
|
return false;
|
|
}
|
|
static void update(TYPE *a, TYPE b)
|
|
{
|
|
UPDATE::update(a, b);
|
|
}
|
|
};
|
|
|
|
template <int WIDTH, typename UPDATE, int FACTOR>
|
|
struct OffsetMinSumAlgorithm<SIMD<int8_t, WIDTH>, UPDATE, FACTOR>
|
|
{
|
|
typedef int8_t VALUE;
|
|
typedef SIMD<VALUE, WIDTH> TYPE;
|
|
static TYPE zero()
|
|
{
|
|
return vzero<TYPE>();
|
|
}
|
|
static TYPE one()
|
|
{
|
|
return vdup<TYPE>(1);
|
|
}
|
|
static TYPE sign(TYPE a, TYPE b)
|
|
{
|
|
return vsign(a, b);
|
|
}
|
|
static TYPE eor(TYPE a, TYPE b)
|
|
{
|
|
return vreinterpret<TYPE>(veor(vmask(a), vmask(b)));
|
|
}
|
|
static TYPE orr(TYPE a, TYPE b)
|
|
{
|
|
return vreinterpret<TYPE>(vorr(vmask(a), vmask(b)));
|
|
}
|
|
static TYPE other(TYPE a, TYPE b, TYPE c)
|
|
{
|
|
return vreinterpret<TYPE>(vbsl(vceq(a, b), vmask(c), vmask(b)));
|
|
}
|
|
static void finalp(TYPE *links, int cnt)
|
|
{
|
|
auto beta = vunsigned(vdup<TYPE>(std::nearbyint(0.5 * FACTOR)));
|
|
TYPE mags[cnt];
|
|
for (int i = 0; i < cnt; ++i)
|
|
mags[i] = vsigned(vqsub(vunsigned(vqabs(links[i])), beta));
|
|
|
|
TYPE mins[2];
|
|
mins[0] = vmin(mags[0], mags[1]);
|
|
mins[1] = vmax(mags[0], mags[1]);
|
|
for (int i = 2; i < cnt; ++i) {
|
|
mins[1] = vmin(mins[1], vmax(mins[0], mags[i]));
|
|
mins[0] = vmin(mins[0], mags[i]);
|
|
}
|
|
|
|
TYPE signs = links[0];
|
|
for (int i = 1; i < cnt; ++i)
|
|
signs = eor(signs, links[i]);
|
|
|
|
for (int i = 0; i < cnt; ++i)
|
|
links[i] = sign(other(mags[i], mins[0], mins[1]), orr(eor(signs, links[i]), vdup<TYPE>(127)));
|
|
}
|
|
static TYPE add(TYPE a, TYPE b)
|
|
{
|
|
return vqadd(a, b);
|
|
}
|
|
static TYPE sub(TYPE a, TYPE b)
|
|
{
|
|
return vqsub(a, b);
|
|
}
|
|
static bool bad(TYPE v, int blocks)
|
|
{
|
|
auto tmp = vcgtz(v);
|
|
for (int i = 0; i < blocks; ++i)
|
|
if (!tmp.v[i])
|
|
return true;
|
|
return false;
|
|
}
|
|
static void update(TYPE *a, TYPE b)
|
|
{
|
|
UPDATE::update(a, vmin(vmax(b, vdup<TYPE>(-32)), vdup<TYPE>(31)));
|
|
}
|
|
};
|
|
|
|
|
|
template <typename VALUE, int WIDTH, typename UPDATE, int FACTOR>
|
|
struct MinSumCAlgorithm<SIMD<VALUE, WIDTH>, UPDATE, FACTOR>
|
|
{
|
|
typedef SIMD<VALUE, WIDTH> TYPE;
|
|
static TYPE zero()
|
|
{
|
|
return vzero<TYPE>();
|
|
}
|
|
static TYPE one()
|
|
{
|
|
return vdup<TYPE>(1);
|
|
}
|
|
static TYPE sign(TYPE a, TYPE b)
|
|
{
|
|
return vsign(a, b);
|
|
}
|
|
static TYPE correction_factor(TYPE a, TYPE b)
|
|
{
|
|
TYPE apb = vabs(vadd(a, b));
|
|
TYPE apb2 = vadd(apb, apb);
|
|
TYPE amb = vabs(vsub(a, b));
|
|
TYPE amb2 = vadd(amb, amb);
|
|
TYPE factor2 = vdup<TYPE>(FACTOR * 2);
|
|
auto pc = vmask(vdup<TYPE>(VALUE(FACTOR) / VALUE(2)));
|
|
auto nc = vmask(vdup<TYPE>(-VALUE(FACTOR) / VALUE(2)));
|
|
pc = vand(pc, vand(vcgt(factor2, apb), vcgt(amb, apb2)));
|
|
nc = vand(nc, vand(vcgt(factor2, amb), vcgt(apb, amb2)));
|
|
return vreinterpret<TYPE>(vorr(pc, nc));
|
|
}
|
|
static TYPE minc(TYPE a, TYPE b)
|
|
{
|
|
TYPE m = vmin(vabs(a), vabs(b));
|
|
TYPE x = vsign(vsign(m, a), b);
|
|
x = vadd(x, correction_factor(a, b));
|
|
return x;
|
|
}
|
|
static void finalp(TYPE *links, int cnt)
|
|
{
|
|
TYPE tmp[cnt];
|
|
CODE::exclusive_reduce(links, tmp, cnt, minc);
|
|
for (int i = 0; i < cnt; ++i)
|
|
links[i] = tmp[i];
|
|
}
|
|
static TYPE add(TYPE a, TYPE b)
|
|
{
|
|
return vadd(a, b);
|
|
}
|
|
static TYPE sub(TYPE a, TYPE b)
|
|
{
|
|
return vsub(a, b);
|
|
}
|
|
static bool bad(TYPE v, int blocks)
|
|
{
|
|
auto tmp = vcgtz(v);
|
|
for (int i = 0; i < blocks; ++i)
|
|
if (!tmp.v[i])
|
|
return true;
|
|
return false;
|
|
}
|
|
static void update(TYPE *a, TYPE b)
|
|
{
|
|
UPDATE::update(a, b);
|
|
}
|
|
};
|
|
|
|
template <int WIDTH, typename UPDATE, int FACTOR>
|
|
struct MinSumCAlgorithm<SIMD<int8_t, WIDTH>, UPDATE, FACTOR>
|
|
{
|
|
typedef int8_t VALUE;
|
|
typedef SIMD<VALUE, WIDTH> TYPE;
|
|
static TYPE zero()
|
|
{
|
|
return vzero<TYPE>();
|
|
}
|
|
static TYPE one()
|
|
{
|
|
return vdup<TYPE>(1);
|
|
}
|
|
static TYPE sign(TYPE a, TYPE b)
|
|
{
|
|
return vsign(a, b);
|
|
}
|
|
static TYPE correction_factor(TYPE a, TYPE b)
|
|
{
|
|
TYPE apb = vqabs(vqadd(a, b));
|
|
TYPE apb2 = vqadd(apb, apb);
|
|
TYPE amb = vqabs(vqsub(a, b));
|
|
TYPE amb2 = vqadd(amb, amb);
|
|
TYPE factor2 = vdup<TYPE>(FACTOR * 2);
|
|
auto pc = vmask(vdup<TYPE>(VALUE(FACTOR) / VALUE(2)));
|
|
auto nc = vmask(vdup<TYPE>(-VALUE(FACTOR) / VALUE(2)));
|
|
pc = vand(pc, vand(vcgt(factor2, apb), vcgt(amb, apb2)));
|
|
nc = vand(nc, vand(vcgt(factor2, amb), vcgt(apb, amb2)));
|
|
return vreinterpret<TYPE>(vorr(pc, nc));
|
|
}
|
|
static TYPE minc(TYPE a, TYPE b)
|
|
{
|
|
TYPE m = vmin(vqabs(a), vqabs(b));
|
|
TYPE x = vsign(vsign(m, a), b);
|
|
x = vqadd(x, correction_factor(a, b));
|
|
return x;
|
|
}
|
|
static void finalp(TYPE *links, int cnt)
|
|
{
|
|
TYPE *tmp = new TYPE[cnt];
|
|
CODE::exclusive_reduce(links, tmp, cnt, minc);
|
|
for (int i = 0; i < cnt; ++i)
|
|
links[i] = tmp[i];
|
|
delete[] tmp;
|
|
}
|
|
static TYPE add(TYPE a, TYPE b)
|
|
{
|
|
return vqadd(a, b);
|
|
}
|
|
static TYPE sub(TYPE a, TYPE b)
|
|
{
|
|
return vqsub(a, b);
|
|
}
|
|
static bool bad(TYPE v, int blocks)
|
|
{
|
|
auto tmp = vcgtz(v);
|
|
for (int i = 0; i < blocks; ++i)
|
|
if (!tmp.v[i])
|
|
return true;
|
|
return false;
|
|
}
|
|
static void update(TYPE *a, TYPE b)
|
|
{
|
|
UPDATE::update(a, vmin(vmax(b, vdup<TYPE>(-32)), vdup<TYPE>(31)));
|
|
}
|
|
};
|
|
|
|
} // namespace ldpctool
|
|
|
|
#endif
|