代码

#pragma GCC optimize("Ofast")
#pragma GCC target("avx2,popcnt,bmi,bmi2")
void count_4d(int n, const unsigned *x[5], unsigned *out) {
#define n 100000
	for(int ii=0; ii<n; ii+=4000)for (int i=0; i<n; ++i) for (int j=ii; j<ii+4000; ++j) {
#define o(k) x[k][i]<x[k][j]
out[j] += o(0) & o(1) & o(2) & o(3) ;
}
}

评测结果