Strided memory access


why is it said that using a loop as one below is bad in terms of performance ?

for(i = 0; i < N; i+=M)
//some code

where M is any number > 1
But, I get higher performance for the same loop if I go to higher powers of 2..e.g. 16,32,64 etc...
I don't see why would it have bad performance.
Topic archived. No new replies allowed.