#line 2 "cpplib/util/template.hpp"
#pragma GCC optimize("Ofast")
#pragma GCC optimize("unroll-loops")
#pragma GCC target("avx2")
#include<bits/stdc++.h>
using namespace std;
struct __INIT__{__INIT__(){cin.tie(0);ios::sync_with_stdio(false);cout<<fixed<<setprecision(15);}}__INIT__;
typedef long long lint;
#define INF (1LL<<60)
#define IINF (1<<30)
#define EPS (1e-10)
#define endl ('\n')
typedef vector<lint> vec;
typedef vector<vector<lint>> mat;
typedef vector<vector<vector<lint>>> mat3;
typedef vector<string> svec;
typedef vector<vector<string>> smat;
template<typename T>using V=vector<T>;
template<typename T>using VV=V<V<T>>;
template<typename T>inline void output(T t){bool f=0;for(auto i:t){cout<<(f?" ":"")<<i;f=1;}cout<<endl;}
template<typename T>inline void output2(T t){for(auto i:t)output(i);}
template<typename T>inline void debug(T t){bool f=0;for(auto i:t){cerr<<(f?" ":"")<<i;f=1;}cerr<<endl;}
template<typename T>inline void debug2(T t){for(auto i:t)debug(i);}
#define loop(n) for(long long _=0;_<(long long)(n);++_)
#define _overload4(_1,_2,_3,_4,name,...) name
#define __rep(i,a) repi(i,0,a,1)
#define _rep(i,a,b) repi(i,a,b,1)
#define repi(i,a,b,c) for(long long i=(long long)(a);i<(long long)(b);i+=c)
#define rep(...) _overload4(__VA_ARGS__,repi,_rep,__rep)(__VA_ARGS__)
#define _overload3_rev(_1,_2,_3,name,...) name
#define _rep_rev(i,a) repi_rev(i,0,a)
#define repi_rev(i,a,b) for(long long i=(long long)(b)-1;i>=(long long)(a);--i)
#define rrep(...) _overload3_rev(__VA_ARGS__,repi_rev,_rep_rev)(__VA_ARGS__)

// #define rep(i,...) for(auto i:range(__VA_ARGS__)) 
// #define rrep(i,...) for(auto i:reversed(range(__VA_ARGS__)))
// #define repi(i,a,b) for(lint i=lint(a);i<(lint)(b);++i)
// #define rrepi(i,a,b) for(lint i=lint(b)-1;i>=lint(a);--i)
// #define irep(i) for(lint i=0;;++i)
// inline vector<long long> range(long long n){if(n<=0)return vector<long long>();vector<long long>v(n);iota(v.begin(),v.end(),0LL);return v;}
// inline vector<long long> range(long long a,long long b){if(b<=a)return vector<long long>();vector<long long>v(b-a);iota(v.begin(),v.end(),a);return v;}
// inline vector<long long> range(long long a,long long b,long long c){if((b-a+c-1)/c<=0)return vector<long long>();vector<long long>v((b-a+c-1)/c);for(int i=0;i<(int)v.size();++i)v[i]=i?v[i-1]+c:a;return v;}
// template<typename T>inline T reversed(T v){reverse(v.begin(),v.end());return v;}
#define all(n) begin(n),end(n)
template<typename T,typename E>bool chmin(T& s,const E& t){bool res=s>t;s=min<T>(s,t);return res;}
template<typename T,typename E>bool chmax(T& s,const E& t){bool res=s<t;s=max<T>(s,t);return res;}
const vector<lint> dx={1,0,-1,0,1,1,-1,-1};
const vector<lint> dy={0,1,0,-1,1,-1,1,-1};
#define SUM(v) accumulate(all(v),0LL)
#if __cplusplus>=201703L
    template<typename T,typename ...Args>auto make_vector(T x,int arg,Args ...args){if constexpr(sizeof...(args)==0)return vector<T>(arg,x);else return vector(arg,make_vector<T>(x,args...));}
#endif
#define extrep(v,...) for(auto v:__MAKE_MAT__({__VA_ARGS__}))
#define bit(n,a) ((n>>a)&1)
vector<vector<long long>> __MAKE_MAT__(vector<long long> v){if(v.empty())return vector<vector<long long>>(1,vector<long long>());long long n=v.back();v.pop_back();vector<vector<long long>> ret;vector<vector<long long>> tmp=__MAKE_MAT__(v);for(auto e:tmp)for(long long i=0;i<n;++i){ret.push_back(e);ret.back().push_back(i);}return ret;}
using graph=vector<vector<int>>;
template<typename T>using graph_w=vector<vector<pair<int,T>>>;
template<typename T,typename E>ostream& operator<<(ostream& out,pair<T,E>v){out<<"("<<v.first<<","<<v.second<<")";return out;}
#if __cplusplus>=201703L
    constexpr inline long long powll(long long a,long long b){long long res=1;while(b--)res*=a;return res;}
#endif

template<typename T,typename E>pair<T,E>& operator+=(pair<T,E>&s,const pair<T,E>&t){s.first+=t.first;s.second+=t.second;return s;}
template<typename T,typename E>pair<T,E>& operator-=(pair<T,E>&s,const pair<T,E>&t){s.first-=t.first;s.second-=t.second;return s;}
template<typename T,typename E>pair<T,E> operator+(const pair<T,E>&s,const pair<T,E>&t){auto res=s;return res+=t;}
template<typename T,typename E>pair<T,E> operator-(const pair<T,E>&s,const pair<T,E>&t){auto res=s;return res-=t;}
#define BEGIN_STACK_EXTEND(size) void * stack_extend_memory_ = malloc(size);void * stack_extend_origin_memory_;char * stack_extend_dummy_memory_ = (char*)alloca((1+(int)(((long long)stack_extend_memory_)&127))*16);*stack_extend_dummy_memory_ = 0;asm volatile("mov %%rsp, %%rbx\nmov %%rax, %%rsp":"=b"(stack_extend_origin_memory_):"a"((char*)stack_extend_memory_+(size)-1024));
#define END_STACK_EXTEND asm volatile("mov %%rax, %%rsp"::"a"(stack_extend_origin_memory_));free(stack_extend_memory_);
#line 2 "main.cpp"
using i64 = long long;

int isqrt(i64 n) {
  return sqrtl(n);
}

__attribute__((target("avx2"), optimize("O3", "unroll-loops")))
i64 prime_pi(const i64 N) {
  if (N <= 1) return 0;
  if (N == 2) return 1;
  const int v = isqrt(N);
  int s = (v + 1) / 2;
  vector<int> smalls(s); for (int i = 1; i < s; ++i) smalls[i] = i;
  vector<int> roughs(s); for (int i = 0; i < s; ++i) roughs[i] = 2 * i + 1;
  vector<i64> larges(s); for (int i = 0; i < s; ++i) larges[i] = (N / (2 * i + 1) - 1) / 2;
  vector<bool> skip(v + 1);
  const auto divide = [] (i64 n, i64 d) -> int { return double(n) / d; };
  const auto half = [] (int n) -> int { return (n - 1) >> 1; };
  int pc = 0;
  for (int p = 3; p <= v; p += 2) if (!skip[p]) {
    int q = p * p;
    if (i64(q) * q > N) break;
    skip[p] = true;
    for (int i = q; i <= v; i += 2 * p) skip[i] = true;
    int ns = 0;
    for (int k = 0; k < s; ++k) {
      int i = roughs[k];
      if (skip[i]) continue;
      i64 d = i64(i) * p;
      larges[ns] = larges[k] - (d <= v ? larges[smalls[d >> 1] - pc] : smalls[half(divide(N, d))]) + pc;
      roughs[ns++] = i;
    }
    s = ns;
    for (int i = half(v), j = ((v / p) - 1) | 1; j >= p; j -= 2) {
      int c = smalls[j >> 1] - pc;
      for (int e = (j * p) >> 1; i >= e; --i) smalls[i] -= c;
    }
    ++pc;
  }
  larges[0] += i64(s + 2 * (pc - 1)) * (s - 1) / 2;
  for (int k = 1; k < s; ++k) larges[0] -= larges[k];
  for (int l = 1; l < s; ++l) {
    int q = roughs[l];
    i64 M = N / q;
    int e = smalls[half(M / q)] - pc;
    if (e < l + 1) break;
    i64 t = 0;
    for (int k = l + 1; k <= e; ++k) t += smalls[half(divide(M, roughs[k]))];
    larges[0] += t - i64(e - l) * (pc + l - 1);
  }
  return larges[0] + 1;
}

int main(){
    lint l,r;
    cin>>l>>r;
    if(l==r){
        cout<<prime_pi(r)-prime_pi(l-1)<<endl;
    }else{
        cout<<prime_pi(r)-prime_pi(l-1)+prime_pi(r*2-1)-prime_pi(l*2)<<endl;
    }
}