| # | User | Rating |
|---|---|---|
| 1 | Benq | 3676 |
| 2 | Kevin114514 | 3659 |
| 3 | jiangly | 3640 |
| 4 | turmax | 3450 |
| 5 | tourist | 3408 |
| 6 | strapple | 3360 |
| 7 | Um_nik | 3354 |
| 8 | squareOf105 | 3348 |
| 9 | maroonrk | 3312 |
| 10 | idontwannadothis…e | 3191 |
| # | User | Contrib. |
|---|---|---|
| 1 | Qingyu | 153 |
| 2 | nik_exists | 151 |
| 3 | maroonrk | 150 |
| 4 | maspy | 149 |
| 5 | AmShZ | 146 |
| 6 | Um_nik | 143 |
| 7 | Errichto | 139 |
| 8 | adamant | 135 |
| 9 | BledDest | 134 |
| 10 | cry | 130 |
| Name |
|---|



your unordered_map got an anti-hash test(unordered_map worst case look time O(n) due to all hashes colliding with each other). And because MS C++ 2017 is a different compiler, so unordered map hashing works a bit different. In whatever way, the three should get timelimit anyways. I recommend you not to use unordered_map in hackable contests or in sites that have system testing afterwards like codeforces.
At least it could have given TLE at the time of contest so i would have changed the unordered_map to map.
Sadly, authors usually don't usually think of anti-hashing tests but hackers do. Here is a really cool blog which helps if you want to use unordered_map.
But here all keys are ints and not int64_t/long longs, so their hashes do not (should not?) collide, so what's the real reason?
Numbers till $$$10^9$$$. At least not all hashes collide, but a lot of them. If numbers were only till $$$10^6$$$ or lower, I guess it was gonna pass.
Yeah, I did not realize that the hash function includes % bucket_count, thank you