#pragma GCC optimize("Ofast,inline,unroll-loops")
#include <bits/stdc++.h>
#include <immintrin.h>
#include <algorithm>
using namespace std;
template<int n>
struct foo {
template<class T>
static void F(uint* __restrict__ buc, uint* __restrict__ a, uint* __restrict__ b, T lambda) {
for (int i = 0; i < n; i += 16) {
_mm_prefetch(&a[i + 256], _MM_HINT_NTA);
#pragma GCC unroll 16
for (int j = 0; j < 16; j++)
b[buc[lambda(a[i + j])]++] = a[i + j];
}
}
uint b[1<<27];
static void sort(uint* a, int) {
uint buc[4][256] = {};
for (int i = 0; i < n; i++) {
buc[0][a[i] & 255]++;
buc[1][a[i] >> 8 & 255]++;
buc[2][a[i] >> 16 & 255]++;
buc[3][(a[i] >> 16) >> 8 & 255]++;
}
for (int k = 0; k < 4; k++) {
uint32_t offset = 0;
for (int i = 0; i < 256; i++)
swap(buc[k][i], offset), offset += buc[k][i];
}
F(buc[0], a, b, [](uint x) { return x & 255; });
F(buc[1], b, a, [](uint x) { return x >> 8 & 255; });
F(buc[2], a, b, [](uint x) { return x >> 16 & 255; });
F(buc[3], b, a, [](uint x) { return x >> 24; });
}
};
void sort(uint* a, int __n) {
// nth_element(a, a+34217728, a+134217728);
foo<34217728>::sort(a, 34217728);
foo<100000000>::sort(a+34217728,100000000);
for(int i=0; i<34217728; ++i)b[i] = a[i];
int p = 0, q = 0, r = 34217728;
while (p!=r) a[p++] = r==134217728||b[q]<a[r]?b[q++]:a[r++];
}
Compilation | N/A | N/A | Compile Error | Score: N/A | 显示更多 |