Data Transfer Settings Between Concurrent Tasks

I don't understand the difference between Ensure deterministic transfer (maximum delay) and Ensure deterministic transfer (minimum delay).

One with maximum delay is said to 'Create opportunity for parallelism', whereas minimum delay one is said to enforce 'Enforce data dependency.'.

Can someone elaborate on this?

In my application, I'm defining 4 tasks which I hope to be assigned to different cores on my 4 core target machine.

task #1 is for hardware abstraction and runs at 4 khz.

task #2 is for signal processing/estimation and runs at 1 khz.

task #3 is for what I call behavior and generates desired trajectories for the controls (task #4) at 500 hz.

task #4 is for controls running at 1 khz.