I have been asked this question in the interview of a start-up\n\nWe have an array with millions of elements. We are given a quad core machine. We need to add these elements in the most efficient way. \n\nMy approach\n\nSince we are given a quad core machine, let us fork out four threads. Let us divide our input array into chunks of equal size. The last chunk may have less elements, depending on the size of the array.\n\nStart 4 threads and to each of these threads pass the chunk. This can be done by passing the individual array along with its start index and the end index. \n\nAlso declare a result array of size 4 in which we will save the sum obtained by summing the elements of each thread.\nAfter the threads return, then do a summation over this array of size 4 and return the sum.\n\nDiscussion points:\n1. Corner cases - Make sure the code runs fine over smaller arrays as well. Array of size 4, array of size 5, array of size 6 and so on\n2. How to handle overflow as the sum of millions of elements may overflow.

I have been asked this question in the interview of a start-up\n\nWe have an array with millions of elements. We are given a quad core machine. We need to add t