C: The reference solution incorrectly expects 0 for all 1-element sequences.

And I like how I was stuck this whole time until I read your comment. My code kept getting timed out because of these summations.
Thanks mate !

specs are unclear : what should we return if end == start ? and if step == 0 ?

not a kata issue

not a kata issue

-5 is not greater than -1. look at the step (third argument)

this does not constitute a kata issue. kata issues are only bugs in the code, don't raise them lightly

I get the following in C :

Testing for -1, -5 and -3
Expected: -5
Actual: 0

Since when is -5 greater than -1 ?

Why is this faster? Are lots of cases in an if statement slow? Would a switch be faster?

Your method works but why would you waste your memory to store all these square values if you won't need them later?

Please use spoiler flag next time.