OCV methodology : blocks vs top

[chris06]

Hello,


I have a question regarding OCV methodology.
I work on a top level that contains a block (that I'm designing also) and I'm looking for the better methodology to apply OCV.
At top-level the block is seen as a blak-box (LIB/LEF/GDSII)
I use a flat OCV approach (different derating factors for setup/hold data/clock-launch/clock-capture cell/net )

I'm generating the (LIB/LEF/GDSII) models of the block to be used at top-level.

1/ Do I have to apply OCV derating factors when generating the LIB files of the block or is it better not to apply the derating factors at block level?

2/ When using the LIB files of the blocks how do I have to apply the derating factors at top-level?

3/ How is handled OCV for the following LIB arc timings of the block at top-level?
- combinational
- setup_rising/falling
- hold_rising/falling

Do I have to use the -cell_check option of "set_timing_derate" command?



Thanks in advance for your help.
Best Regards
Chris