Question about clock reconvergence pessimism

[promach]

Could anyone explain how is the pessimism being removed ?

and the calculation in https://www.physicaldesign4u.com/2020/07/ocv-on-chip-variation-and-crpr-clock.html confused me

 

[promach]

should it be Setup slack = (arrival time) min - (required time) max instead ?

--- Updated Nov 30, 2021 ---

Besides, why the static CRP is equal to 2 (5 minus 3) ?

 

[ThisIsNotSam]

here is all you need to know about CRPR: if a cell is present on the clock launch path and also on the clock arrival path, it should not be considered to be slow and fast at the same time. this has no meaning, the cell is a single cell and process variation will make it either faster or slower, but never both. CRPR is the calculation you do to remove this absurd scenario from the clock pessimism.

 

[promach]

I have understood how CRPP mechanism works for the same-type transition, but I am still not sure about One rising, one falling scenario

--- Updated Nov 30, 2021 ---

I am also still studying about clock-to-data pessimism

 

[promach]

Besides, why The early source latency affects only the minimum values of the arrival window ?