Is there an efficient way to calculate the sum of lcm(x,y) for every value x and y such that 1<=x<=n && 1<=y<=n with a given n ?
| # | User | Rating |
|---|---|---|
| 1 | jiangly | 3810 |
| 2 | Benq | 3676 |
| 3 | Kevin114514 | 3655 |
| 4 | maroonrk | 3463 |
| 5 | strapple | 3390 |
| 6 | Um_nik | 3387 |
| 7 | tourist | 3384 |
| 8 | heuristica | 3322 |
| 9 | turmax | 3319 |
| 10 | jiangbowen | 3291 |
| # | User | Contrib. |
|---|---|---|
| 1 | Qingyu | 156 |
| 2 | nik_exists | 150 |
| 2 | maspy | 150 |
| 4 | Um_nik | 143 |
| 4 | AmShZ | 143 |
| 6 | Errichto | 139 |
| 7 | adamant | 137 |
| 8 | maroonrk | 134 |
| 9 | BledDest | 132 |
| 10 | qwexd | 129 |
Is there an efficient way to calculate the sum of lcm(x,y) for every value x and y such that 1<=x<=n && 1<=y<=n with a given n ?
| Name |
|---|



Mobius Inversion
You can also do dp where you dp[i] is sum of all x * y numbers with gcd(x, y) equal to i, and to transition u find with basic math sum of all x * y with gcd(x, y) >= i by taking sum of all multiples of i and squaring it like sieve then in sieve like fashion subtract out dp values that are multiples of i as that gets rid of overcounting. Obviously you then just take sum of dp[i] / i for all dp values. I'm p sure this dp method and mobius inversion are almost always interchangeable but not sure...