research/erasure_code/ec65536/faster_poly_utils.cpp

341 lines
10 KiB
C++
Raw Normal View History

2017-04-26 00:56:32 +00:00
#include "stdlib.h"
#include "stdio.h"
#include <iostream>
#include <vector>
#include <deque>
using namespace std;
vector<int> glogtable(65536, 0);
vector<int> gexptable(196608, 0);
const int ROOT_CUTOFF = 32;
void initialize_tables() {
int v = 1;
for (int i = 0; i < 65536; i++) {
glogtable[v] = i;
gexptable[i] = v;
gexptable[i + 65535] = v;
gexptable[i + 131070] = v;
if (v & 32768)
v = (v * 2) ^ v ^ 103425;
else
v = (v * 2) ^ v;
}
}
int eval_poly_at(vector<int> poly, int x) {
if (x == 0)
return poly[0];
int logx = glogtable[x];
int y = 0;
for (int i = 0; i < poly.size(); i++) {
if (poly[i])
y ^= gexptable[(logx * i + glogtable[poly[i]]) % 65535];
}
return y;
}
int eval_log_poly_at(vector<int> poly, int x) {
if (x == 0)
return poly[0] == 65537 ? 0 : gexptable[poly[0]];
int logx = glogtable[x];
int y = 0;
for (int i = 0; i < poly.size(); i++) {
if (poly[i] != 65537)
y ^= gexptable[(logx * i + poly[i]) % 65535];
}
return y;
}
// Compute the product of two (equal length) polynomials. Takes ~O(N ** 1.59) time.
vector<int> karatsuba_mul(vector<int> p, vector<int> q) {
int L = p.size();
if (L <= 64) {
vector<int> o(L * 2);
vector<int> logq(L);
for (int i = 0; i < L; i++) logq[i] = glogtable[q[i]];
for (int i = 0; i < L; i++) {
int log_pi = glogtable[p[i]];
for (int j = 0; j < L; j++) {
if (p[i] && q[j])
o[i + j] ^= gexptable[log_pi + logq[j]];
}
}
return o;
}
if (L % 2) {
L += 1;
p.push_back(0);
q.push_back(0);
}
int halflen = L / 2;
vector<int> low1 = vector<int>(p.begin(), p.begin() + halflen);
vector<int> low2 = vector<int>(q.begin(), q.begin() + halflen);
vector<int> high1 = vector<int>(p.begin() + halflen, p.end());
vector<int> high2 = vector<int>(q.begin() + halflen, q.end());
vector<int> sum1(halflen);
vector<int> sum2(halflen);
for (int i = 0; i < halflen; i++) {
sum1[i] = low1[i] ^ high1[i];
sum2[i] = low2[i] ^ high2[i];
}
vector<int> z0 = karatsuba_mul(low1, low2);
vector<int> z2 = karatsuba_mul(high1, high2);
vector<int> m = karatsuba_mul(sum1, sum2);
vector<int> o(L * 2);
for (int i = 0; i < L; i++) {
o[i] ^= z0[i];
o[i + halflen] ^= (m[i] ^ z0[i] ^ z2[i]);
o[i + L] ^= z2[i];
}
return o;
}
vector<int> mk_root(vector<int> xs) {
int L = xs.size();
if (L >= ROOT_CUTOFF) {
int halflen = L / 2;
vector<int> left = vector<int>(xs.begin(), xs.begin() + halflen);
vector<int> right = vector<int>(xs.begin() + halflen, xs.end());
vector<int> o = karatsuba_mul(mk_root(left), mk_root(right));
o.resize(L + 1);
return o;
}
vector<int> root(L + 1);
root[L] = 1;
for (int i = 0; i < L; i++) {
int logx = glogtable[xs[i]];
int offset = L - i - 1;
root[offset] = 0;
for (int j = offset; j < i + 1 + offset; j++) {
if (root[j + 1] and xs[i])
root[j] ^= gexptable[glogtable[root[j+1]] + logx];
}
}
return root;
}
vector<int> subroot_linear_combination(vector<int> xs, vector<int> factors) {
int L = xs.size();
/*if (L <= ROOT_CUTOFF) {
vector<int> out(L + 1);
vector<int> root = mk_root(xs);
for (int i = 0; i < L; i++) {
vector<int> output(L + 1);
output[root.size() - 2] = 1;
int logx = glogtable[xs[i]];
if (factors[i]) {
int log_fac = glogtable[factors[i]];
for (int j = root.size() - 2; j > 0; j--) {
if (output[j] and xs[i])
output[j - 1] = root[j] ^ gexptable[glogtable[output[j]] + logx];
else
output[j - 1] = root[j];
out[j] ^= gexptable[glogtable[output[j]] + log_fac];
}
out[0] ^= gexptable[glogtable[output[0]] + log_fac];
}
}
return out;
}*/
if (L == 1) {
vector<int> o(2);
o[0] = factors[0];
return o;
}
int halflen = L / 2;
vector<int> xs_left = vector<int>(xs.begin(), xs.begin() + halflen);
vector<int> xs_right = vector<int>(xs.begin() + halflen, xs.end());
vector<int> factors_left = vector<int>(factors.begin(), factors.begin() + halflen);
vector<int> factors_right = vector<int>(factors.begin() + halflen, factors.end());
vector<int> R1 = mk_root(xs_left);
vector<int> R2 = mk_root(xs_right);
vector<int> o1 = karatsuba_mul(R1, subroot_linear_combination(xs_right, factors_right));
vector<int> o2 = karatsuba_mul(R2, subroot_linear_combination(xs_left, factors_left));
vector<int> o(L + 1);
for (int i = 0; i < L; i++) {
o[i] = o1[i] ^ o2[i];
}
return o;
}
vector<int> derivative_and_square_base(vector<int> p) {
vector<int> o((p.size() - 1) / 2);
for (int i = 0; i < o.size(); i+= 1) {
o[i] = p[i * 2 + 1];
}
return o;
}
vector<int> poly_to_logs(vector<int> p) {
vector<int> o(p.size());
for (int i = 0; i < p.size(); i++) {
if (p[i])
o[i] = glogtable[p[i]];
else
o[i] = 65537;
}
return o;
}
vector<int> xn_mod_poly(vector<int> inp) {
if (inp.size() == 1) {
vector<int> o(1);
o[0] = gexptable[65535 - glogtable[inp[0]]];
return o;
}
int halflen = inp.size() / 2;
int highlen = inp.size() - (inp.size() / 2);
vector<int> low(inp.begin(), inp.begin() + halflen);
vector<int> high(inp.begin() + halflen, inp.end());
vector<int> lowinv = xn_mod_poly(low);
vector<int> submod = karatsuba_mul(lowinv, low);
vector<int> submod_high(submod.begin() + halflen, submod.end());
lowinv.resize(highlen);
vector<int> med = karatsuba_mul(high, lowinv);
vector<int> med_plus_high(halflen);
for (int i = 0; i < halflen; i++) {
med_plus_high[i] = med[i] ^ submod_high[i];
}
vector<int> highinv = karatsuba_mul(med_plus_high, lowinv);
vector<int> o(inp.size());
for (int i = 0; i < halflen; i++) {
o[i] = lowinv[i];
o[i + halflen] = highinv[i];
}
return o;
}
vector<int> reverse(vector<int> inp) {
vector<int> o(inp.size());
for (int i = 0; i < inp.size(); i++) {
o[inp.size() - 1 - i] = inp[i];
}
return o;
}
vector<int> mod(vector<int> a, vector<int> b) {
int L = b.size();
vector<int> rev_b = reverse(b);
rev_b.resize((L - 1) * 2);
vector<int> inv_rev_b = xn_mod_poly(reverse(b));
inv_rev_b.resize(L);
vector<int> rev_a = reverse(a);
rev_a.resize(L);
vector<int> rev_quotient = karatsuba_mul(inv_rev_b, rev_a);
rev_quotient.resize(L - 1);
vector<int> quotient = reverse(rev_quotient);
quotient.resize(L);
vector<int> diff = karatsuba_mul(b, quotient);
vector<int> o(L-1);
for (int i = 0; i < L-1; i++) {
o[i] = a[i] ^ diff[i];
}
return o;
}
vector<int> multi_eval(vector<int> poly, vector<int> xs) {
int L = xs.size();
if (L <= 1024) {
vector<int> o(L);
vector<int> logz = poly_to_logs(poly);
for (int i = 0; i < L; i++) {
o[i] = eval_log_poly_at(logz, xs[i]);
}
return o;
}
int halflen = L / 2;
vector<int> left(xs.begin(), xs.begin() + halflen);
vector<int> right(xs.begin() + halflen, xs.end());
vector<int> o1;
vector<int> o2;
if (poly.size() < xs.size()) {
o1 = multi_eval(poly, left);
o2 = multi_eval(poly, right);
}
else {
o1 = multi_eval(mod(poly, mk_root(left)), left);
o2 = multi_eval(mod(poly, mk_root(right)), right);
}
o1.resize(L);
for (int i = 0; i < halflen; i++) {
o1[halflen + i] = o2[i];
}
return o1;
}
vector<int> lagrange_interp(vector<int> ys, vector<int> xs) {
int xs_size = xs.size();
vector<int> root = mk_root(xs);
vector<int> rootprime = derivative_and_square_base(root);
vector<int> xsquares(xs_size);
for (int i = 0; i < xs_size; i++)
xsquares[i] = xs[i] ? gexptable[glogtable[xs[i]] * 2] : 0;
vector<int> denoms = multi_eval(rootprime, xsquares);
vector<int> factors(xs_size);
for (int i = 0; i < xs_size; i++) {
if (ys[i])
factors[i] = gexptable[glogtable[ys[i]] + 65535 - glogtable[denoms[i]]];
}
vector<int> o = subroot_linear_combination(xs, factors);
o.resize(xs_size);
return o;
}
const int SIZE = 4096;
int main() {
initialize_tables();
/*int myxs[] = {1, 2, 3, 4, 5};
std::vector<int> test (myxs, myxs + sizeof(myxs) / sizeof(int) );
int myxs2[] = {6, 7, 8, 9, 10, 11, 12, 13};
std::vector<int> test2 (myxs2, myxs2 + sizeof(myxs2) / sizeof(int) );*/
/*std::vector<int> test(257);
for (int i = 0; i < 257; i++) test[i] = i;
std::vector<int> test2(512);
for (int i = 0; i < 512; i++) test2[i] = 1000 + i;
vector<int> moose = mod(test2, test);
for (int i = 0; i < 256; i++) cout << moose[i] << " ";
cout << "\n";*/
vector<int> xs(SIZE);
vector<int> ys(SIZE);
for (int v = 0; v < SIZE; v++) {
ys[v] = v * 3;
xs[v] = 1000 + v * 7;
}
//vector<int> d = derivative(mk_root(xs));
//for (int i = 0; i < d.size(); i++) cout << d[i] << " ";
//cout << "\n";
/*vector<int> prod = mk_root(xs);
vector<int> prod = karatsuba_mul(xs, ys);
for (int i = 0; i < SIZE + 1; i++)
cout << prod[i] << " ";
cout << "\n";
cout << eval_poly_at(prod, 189) << " " << gexptable[glogtable[eval_poly_at(xs, 189)] + glogtable[eval_poly_at(ys, 189)]] << "\n";*/
for (int a = 0; a < 10; a++) {
ys[0] = a;
vector<int> poly = lagrange_interp(ys, xs);
vector<int> new_xs(SIZE);
for (int i = 0; i < SIZE; i++) new_xs[i] = SIZE + i;
vector<int> results = multi_eval(poly, new_xs);
cout << eval_poly_at(poly, 1700) << "\n";
unsigned int o = 0;
for (int i = 0; i < SIZE; i++) {
o += results[i];
}
cout << o << "\n";
}
//cout << eval_poly_at(poly, 0) << " " << ys[0] << "\n";
//cout << eval_poly_at(poly, 134) << " " << ys[134] << "\n";
//cout << eval_poly_at(poly, 375) << " " << ys[375] << "\n";
//int o;
//for (int i = 0; i < 524288; i ++)
// o += eval_poly_at(poly, i % 65536);
//std::cout << o;
}