Hi avinash,
3% and 8% of the core physical area is required for the decoupling capacitors. This is referred to as the “dcap density”. For worst cases decoupling capacitance comes entirely from dedicated Cox (gate oxide capacitance) sources.
Designs using DSM technologies now require the power grid voltage drop to be much less than 10% of Vdd. To achieve this goal, decoupling capacitors are added to minimize switching noise. The charge held by dcap is used to ensure voltage stability during the high-speed switching events that charge and discharge.
Dcap cells can be added incrementally as needed. The number of dcaps can be estimated and then placed based on intermediate power density analysis results and IR drop analysis results. This will most likely be done at the ECO cell and filler insertion stage.
Decoupling capacitors are inserted in a two step approach. The first step consists of inserting dcaps before the standard cells are placed. This will ensure proper distribution of dcaps in the design before timing driven placement cause clumping that can prevent even distribution of dcaps. The second step is done post-route to add decoupling capacitors where necessary based on power rail analysis.