| № | Пользователь | Рейтинг |
|---|---|---|
| 1 | tourist | 4009 |
| 2 | jiangly | 3773 |
| 3 | Radewoosh | 3646 |
| 4 | ecnerwala | 3624 |
| 5 | jqdai0815 | 3620 |
| 5 | Benq | 3620 |
| 7 | orzdevinwang | 3612 |
| 8 | Geothermal | 3569 |
| 8 | cnnfls_csy | 3569 |
| 10 | Um_nik | 3396 |
| Страны | Города | Организации | Всё → |
| № | Пользователь | Вклад |
|---|---|---|
| 1 | Um_nik | 163 |
| 2 | cry | 161 |
| 3 | maomao90 | 160 |
| 4 | -is-this-fft- | 159 |
| 5 | awoo | 158 |
| 6 | atcoder_official | 157 |
| 7 | adamant | 155 |
| 8 | nor | 154 |
| 9 | maroonrk | 152 |
| 10 | Dominater069 | 148 |
I tried implementing Mo's Algorithm to solve this problem but I ran into a couple of TLE's when taking block size as sqrt(Max N) which is 448.
The submission can be viewed here-21756475 The result was TLE when using block size=448.
In the next submission-21756591 , I put block size=1000 and the result was AC.
So my question is how do I figure out what value to use as my block size?
Auto comment: topic has been updated by Vicennial (previous revision, new revision, compare).
Assuming you do not want to submit a large number of times with different block sizes, one way to do this is by creating a test case with the highest possible values for n and t (and randomly generated queries) and testing the runtimes for different block sizes.
I submitted different forms of my code for around 20 times in this question. What I saw is that time taken in this question changes abruptly even on subtle changes in code.
For example I changed ll to int in your code for some arrays and numbers except answer ones http://codeforces.me/contest/86/submission/24079499 and it got AC.
When I was doing it, I removed #include <bits/stdc++.h> from my code and time reduced to 2240 from 3630 ms. I think it takes pretty neat and to the point code at Codeforces otherwise these things shouldn't matter.