I think you meant 2N+2M vs 2N+M, but yes, that’s the point: not reading twice is cheaper because unlike in traditional big-O analysis compute is cheap but data

I think you meant 2N+2M vs 2N+M, but yes, that’s the point: not reading twice is cheaper because unlike in traditional big-O analysis compute is cheap but data access is very expensive.