Runtime per se won't be a problem here (maybe it is about 0.4ms). The real trouble is, that this ends up in an infinite loop for values above the square closest to INT_MAX.

Well, I feel dumb for not figuring out a fast enough algorithm for this.

I didn't mean to imply or claim anything. I was just wondering, if there was a reason behind it -- like style, code length or readability etc.

There is always something important, but without asking questions you might miss it.

;-)

Thank you very much.

One thing: I don't like that conditional within the loop. One branch is only executed once, so why don't you execute that part/branch outside of the loop?

@rowcased ~ Thanks for the effort. Now it works fine (but sadly the large values are gone).

@rowcased: I disagree. These failed tests (testing large random values) only occur in the C version (well at least not in Python or JavaScript). So maybe there is something wrong with the expected results. It would help, if there was a more specific message (with n and m replaced by actual values). See also my next comment with spoilers...

Thank you as well, now I can mess things up myself.

That's half-hearted.

;-)

If you use a list of precalculated values, why not a list of the final results?

Are you kidding me? Comments on hidden solutions are not hidden?