Many programmers always argue that the problems in Competitive Programming always end up with TLE(Time Limit Exceed). The Main problem of this error is that it will not allow you to know that your solution would reach to correct solution or not!
Why TLE comes?
- Online Judge Restrictions: TLE comes because the Online judge have some restriction that it will not allow to process the instruction after certain Time limit given by Problem setter usually (1 sec).
- Server Configuration: The exact time taken by the code depends on the speed of server,architecture of server,OS and certainly on the complexity of the algorithm. So different servers like practice, codechef, SPOJ etc may have different execution speed. By estimating the maximum value of N (N is total number of instructions of your whole code), you can roughy estimate the TLE would occur or not in 1 sec.
MAX value of N Time complexity 10^8 O(N) Border case 10^7 O(N) Might be accepted 10^6 O(N) Perfect 10^5 O(N * logN) 10^3 O(N ^ 2) 10^2 O(N ^ 3) 10^9 O(logN) or Sqrt(N)
So after analyzing this chart you can roughly estimate your Time complexity and make your code within the upper bound limit.
- Method of reading input and writing output is too slow: Sometimes methods used by a programmer for input output may cause TLE.
OverCome Time Limit Errors
- Change methods of Input-Output: You must choose proper input-output functions and data structure which would help you in optimization.
- In C++, do not use cin/cout – use scanf and printf instead.
- In Java, do not use a Scanner – use a BufferedReader instead.
- In Python, you could try speeding up your solutions by adding the following two lines to the start of your file:
- Bounds of loops may be reduced : Read the bounds in the input carefully before writing your program, and try to figure out which inputs will cause your program to run the slowest. For example if a problem tells you that N <= 100000 or N<=1000000, and your program has nested loops each which go up to N, your program will never be fast enough.
- Optimize your Algorithm: If nothing works after all this, then you should try changing the algorithm or the approach you are using to solve your problem. Generally , the internal test cases are designed in such a way that you will be able to clear all of them only if you choose the best possible algorithm.
- Look for Suggestions given: Although this should be the last step but you must look at the comments given below any problem in which other programmers might have given a hint on how the problem can be soved in a better and more efficient way. And even when you overcome TLE try more exhaustive and corner test cases against your program to check the performance.
Ultimately, with experience you’ll surely come to know what to do and what not to avoid TLEs. The more you code the more you get to know about how to compete TLE.
This article is contributed by Shubham Bansal. If you like GeeksforGeeks and would like to contribute, you can also write an article using contribute.geeksforgeeks.org or mail your article to firstname.lastname@example.org. See your article appearing on the GeeksforGeeks main page and help other Geeks.
Please write comments if you find anything incorrect, or you want to share more information about the topic discussed above.
- Cybersecurity Challenges In Digital Marketing - Take These Steps To Overcome
- Find the time which is palindromic and comes after the given time
- Top Programmers in the World of All Time
- Time of Flight (ToF) Sensors
- What to do at the time of Wrong Answer (WA)?
- time.h localtime() function in C with Examples
- Find the arrangement of queue at given time
- Some time-saving tips for Linux Users
- Extended Mo's Algorithm with ≈ O(1) time complexity
- Find the minimum time after which one can exchange notes
- How to find time taken by a command/program on Linux Shell?
- Game Theory in Balanced Ternary Numeral System | (Moving 3k steps at a time)
- Dynamic Programming | Wildcard Pattern Matching | Linear Time and Constant Space
- Sqrt (or Square Root) Decomposition | Set 2 (LCA of Tree in O(sqrt(height)) time)
- Number of ways to arrange K different objects taking N objects at a time