Forcing loop iterations to execute sequentially

Hi all, I have some code that has a nested for loop in a single work item kernel like so: __local lmem[M][2]; // ping pong buffer for (uint outer = 0; outer < N; ++outer) {__privat...