This question was originally only asking for performance measurement tools. It has since been modified somewhat to ask for something closer to a review, but neither of the answers address this.
So:
As the moderator who responded to your flag, it was correct. The original question did seem to ask for information on how to profile the application (off topic). However, I saw a good review question - performance is explicitly on topic. So I edited the title to ask for a performance review rather than a "how do I.." question.
Now, concerning the answers: I agree that the answers provided thus far don't really answer the question. That is cause for downvotes. My criteria for "Not an answer" is that the answer does not provide a solution, i.e., asks a question which should be in a comment or is merely a useless linky answer. Both of these I will convert to comments if I see them or respond to a flag. If I judge the link is irrelevant (not necessarily spam, but completely irrelevant to the question) I may delete it outright. Generally I err toward comment conversion.
tl;dr: You are correct; the current answers don't really answer the question. They should be downvoted. If more attention to the question is needed, a bounty could be posted or the question could be edited to be more specific, which would bump it to the top of the front page.
I don't think the OP was asking about how to use profiling tools. The question was How do i I find the bottlenecks in my code? My read on it is that the OP wanted help finding his bottlenecks which would in fact be on-topic here.
Flagging doesn't do anything besides have me or another moderator have to decide what to do with the question. So flagging really isn't action to take, its a request for moderators to take action. What do you think the moderators should do to those answers?
IMO, the underlying problem here is that programming contest questions don't lend themselves to real code reviews. The kind of stuff I'd normally complain about in code simply doesn't apply to a programming contest problem.
