5.3.5p5: (754270) - Particles: Start delay... Still Have Problem

I tried the latest patch of 5.3.5p5, single particle system is ok. But it seems the particle start delay still have problem on complicated multi layer particle systems.
When particle system Start Delay param set greater than 0, and pause during playing, and then resume, some particle delay time still recount from 0, but before pause, the play time already exceed the Start Delay param.
bellow is my particle structure: