Hello, Codeforces!
We are excited to invite you to the mirror of COMPFEST 18 Final Round! 
COMPFEST is an annual national-level IT event and competition organized by the Faculty of Computer Science, Universitas Indonesia (Fasilkom UI). This mirror is based on the problemset of the COMPFEST 18 Senior Competitive Programming Contest Final Round, and is open for everyone to participate!
This mirror features a subset of the problems from the original contest.
Key Details:
- Rated for all
- Contest link: TLX
- Time: Sunday, September 27, 2026 at 13:00 UTC+7
- Duration: 150 minutes
- Problems: 8
- Allowed languages: C, C++, Pascal, Java, Python 3, Go, Rust
- Contest style: ICPC
Penalty per wrong submission: 10 minutes
- Available languages: English, Bahasa Indonesia
- Authors: RizkWar, robertoeugenio, vinren, worthytacos
We Would Like to Thank:
- COMPFEST 18 Scientific Committee: Wuskevo, DYTOBYTE, GaLz, Nalkq, Nicyue, kukuk, RizkWar, robertoeugenio, vinren, and worthytacos
- Mirror coordinators: Yoshiyuki and prabowo
- Testers: aufannn, bananade1, benedictzir, ByeWorld512, codingfennel, gansixeneh, and Zanite
For more information about the contest rules and regulations, please refer to the contest page.
Please register for the contest, and we encourage everyone to try all the problems! We hope you enjoy the mirror and have fun solving!








Auto comment: topic has been updated by worthytacos (previous revision, new revision, compare).
As a (not) tester, i (didn't) test
As the author, i hope you guys keep persevering and always carry hope inside your heart!
bump
Is it rated?
Have fun with the problems guys.. :D
letsgo
omaga, is this real chat?
bump, contest is starting in 15 hour
Friendly reminder, the contest is starting in ~2 hours, happy solving :D
Regarding Problem E:
Did you test the problem on TOKI?
It seems unreasonable to set the time limit to 2 seconds for a problem that only requires finding the height of a tree, when even the fastest submissions take around 1.3 seconds for N = 10^6.
I was able to get rid of the TLE by replacing O(N) repeated multiplications by 2 with binary exponentiation or rewriting the recursive DFS as BFS. lol
It seems like all of our testers used fast exponentiation, and the official solution also uses it. The official solution passes comfortably under 1 second on TLX, so it seems like the 2-second time limit is not too tight when using the intended approach.
However, we weren't aware that the O(N) repeated multiplications approach could be this close to the limit. We really apologize for not catching this during testing, and for any inconvenience or frustration this may have caused while solving the problem.
Oops I'm late, but as a tester, time limit exceeded spam :snake: