Regarding hold violations:

[vijji.sami]

guys please can anybody rectify my problem.
in my project ,
after CTS optimization i am able to clear setup violations.but i am getting large hold violations.but for the same hold violation paths i have large positive values in setup.how can i borrow some amount of time from it.how usefulSkew helps in such situations?
is it bettter to keep "setoptMode usefulSkew true" before CTS? if we can How it is helpful before n after CTS?

Plz guys reply me soon.

Thank u in advance.

 

[morris_mano]

It helps if you analyze the violating path first, before turning on other switches like usefulskew. What is the clock speed? Are the clock frequency of launching flop and capturing flop different in the violating path? What is causing such a long data path in the violating path? Are there too many combinational logic in the path? Are there too many buffers added by the tool in that path to meet hold?

 

[vijji.sami]

Clock Period is actually 5ns n same clock is applied for launching n capturing.i have seen that there are number of added in data path.How can i control it? n one more doubt that eating my brain always is ,Is it better to do all domains optimization at a time r domain wise optimization?

 

[morris_mano]

Couple things I usually check when I have to debug timing violating paths are
1) Is the skew between the launching clock and capturing clock too large?
2) Is the path constrained correctly?
3) Are there any false path the tool is trying to fix?

Optimizing for hold before CTS may also help. But since the hold violation is very large as you have mentioned, from my experience its usually clock issue or constraint issue.