Dynamic Memory Efficiency

I need a large 3D array (typical size 4000x400x3) in my program. Which is more efficient of -
(1) dynamically allocate a 1D array, then use index calculation (I am thinking of a macro) to simulate 3D array
(2) dynamically allocate 3D array and use it directly.

Will things change for a 2D array?

In my program, the array elements will be referenced like 2-4 million times in a typical execution.

If you're looking for raw speed, I'd lean toward the 1D array using index calculation, but the only way you'll really find out for sure is to try it both ways and see which one works best for you.