MaxxxCoder's blog

Question about system testing
By MaxxxCoder, history, 2 hours ago, In English

Why system testing on Div2 rounds is fast, while system testing on Div.3 and Div.4 are so long? Is it Div.2 rounds specific or the reason is the number of submissions (on Div2 it's probably smaller than on Div3-4 ones)?

  • Vote: I like it
  • +3
  • Vote: I do not like it

»
118 minutes ago, # |
  Vote: I like it 0 Vote: I do not like it

It's because there's about a billion submissions on the earlier problems from newbies and unrated users. Codeforces is falling.

→ Reply
  • »
    »
    108 minutes ago, # ^ |
      Vote: I like it 0 Vote: I do not like it

    Thank you.

    → Reply
»
12 minutes ago, # |
← Rev. 2   Vote: I like it +1 Vote: I do not like it

12hr hacking phase causes one layer of delay. Extra preparation (filtering hack tests, then adding them into the final testset) adds the next layer. System testing itself, with the amount of test data usually doubling/tripling the original, adds the final layer.

→ Reply
»
7 minutes ago, # |
  Vote: I like it 0 Vote: I do not like it

Honestly this open hacking idea seems to fall apart especially in the last month or so. In my opinion the test cases shouldn't be that loose so you require 100 additional user submitted cases. And they should do away with the cliche hash TLE stuff. Only gives inconvenience to those who wanna upsolve.

Look at yesterday's Div 3 even. It has been in testing for hours. T_T

→ Reply
  • »
    »
    5 minutes ago, # ^ |
    ← Rev. 2   Vote: I like it 0 Vote: I do not like it

    "cliche hash TLE stuff" will always exist as long as hacks exist. to avoid TLE during upsolving, use regular map/set instead of hashmap/set. If you're doing manual hashing fiddle around with different modulus.

    also, the reason that a lot of hack cases exist is that there are probably lots of different possible hacks to a WA/TLE solution, which all get added to systests after.

    → Reply