Method and apparatus for optimizing body bias connections in CMOS circuits using a deep n-well grid structure
View Patent ↗A method and apparatus for optimizing body bias connections to NFETs and PFETs using a deep n-well grid structure. A deep n-well is formed below the surface of a CMOS substrate supporting a plurality of NFETs and PFETs having a nominal gate length of less than 0.2 microns. The deep n-well is a grid structure with a regular array of apertures providing electrical continuity between the bottom of the substrate and the NFETs. The PFETs reside in surface n-wells that are continuous with the buried n-well grid structure. The grid and n-well layout is performed on the basis of the functionality of the PFETs contained in the n-wells.
1. A computer-implemented method for optimizing body bias connections in a CMOS circuit substrate using a deep n-well grid structure, said method comprising:
in a computer system:
a) performing a circuit layout for said substrate;
b) prioritizing a plurality of PFETs for placement in a plurality of body biased n-wells;
c) selecting a set of dimensional parameters for said deep n-well grid structure from a set of available parameters;
d) selecting an alignment between said deep n-well grid structure and said circuit layout for a set of available alignments;
e) determining a score for said alignment;
f) repeating d) through e) until said set of available alignments is exhausted;
g) repeating c) through f) until said set of available parameters is exhausted; and
h) performing a layout for said deep n-well grid structure using the dimensional parameters and alignment having a score which exceeds a predetermined threshold.
2. The method of claim 1 , wherein the circuit layout is performed independently from the grid layout.
3. The method of claim 1 , wherein the prioritizing is performed on the basis of inclusion on a critical path.
4. The method of claim 1 , wherein the prioritizing is performed on the basis of a power dissipation.
5. The method of claim 1 , wherein the prioritizing is performed on the basis of the effects of leakage current anticipated by modeling of the PFETs.
6. The method of claim 1 , wherein said layout comprises placing a subset of said plurality of PFETs in said body biased n-wells according to said prioritizing.
7. A computer-implemented method for optimizing body bias connections in a CMOS circuit substrate using a deep n-well grid structure, said method comprising:
in a computer system:
a) accessing a pre-existing circuit layout for said substrate;
b) prioritizing a plurality of PFETs for placement in a plurality of body biased n-wells;
c) selecting a set of dimensional parameters for said deep n-well grid structure from a set of available parameters;
d) selecting an alignment between said deep n-well grid structure and said circuit layout from a set of available alignments; and
e) performing a layout for said deep n-well grid structure using the selected dimensional parameters and the selected alignment.
8. The method of claim 7 further comprising:
f) determining a score for said selected alignment;
g) repeating d) and f) until said set of available alignments is exhausted; and
h) repeating c), d) and f) until said set of available parameters is exhausted, wherein said selected alignment results from said h).
9. The method of claim 7 , wherein the prioritizing is performed on the basis of inclusion on a critical path.
10. The method of claim 7 , wherein the prioritizing is performed on the basis of a power dissipation.
11. The method of claim 7 wherein the prioritizing is performed on the basis of the effects of leakage current anticipated by modeling of the PFETs.
12. The method of claim 7 , wherein said layout comprises placing a subset of said plurality of PFETs in said body biased n-wells according to said prioritizing.
13. A computer-implemented method for optimizing body bias connections in a CMOS circuit substrate using a deep well grid structure, said method comprising:
in a computer system:
accessing a first circuit layout for said substrate;
selecting a set of dimensional parameters for said deep well grid structure from a set of available parameters;
selecting an alignment between said deep well grid structure and said first circuit layout for a set of available alignments;
determining a score for said alignment;
first repeating said selecting and said determining until said set of available alignments is exhausted;
second repeating said selecting, said determining and said first repeating until said set of available parameters is exhausted; and
creating a second layout for said deep well grid structure using the dimensional parameters and alignment having a score which exceeds a predetermined threshold.
14. The method of claim 13 , wherein the circuit layout is performed independently from the grid layout.
15. The method of claim 13 , wherein the prioritizing is performed on the basis of inclusion on a critical path.
16. The method of claim 13 , wherein the prioritizing is performed on the basis of a power dissipation.
17. The method of claim 13 , wherein the prioritizing is performed on the basis of the effects of leakage current anticipated by modeling of a plurality of NFETs included in said circuit.
18. The method of claim 13 , wherein said creating a second layout comprises placing a subset of said plurality of NFETs in body biased p-wells.