Advanced modular cell placement system with functional sieve optimization technique

• US 5,963,455 A
• Filed: 06/28/1996
• Issued: 10/05/1999
• Est. Priority Date: 06/28/1996
• Status: Expired due to Term

• Current Assignee: Bell Semiconductor, LLC (Hilco Inc. (d/b/a Hilco Global))
• Original Assignee: LSI Logic Corporation (Broadcom, Inc.)
• Inventors: Scepanovic, Ranko, Andreev, Alexander E., Koford, James S.
• Primary Examiner(s): Teska, Kevin J.
• Assistant Examiner(s): Kik, Phallaka
• Application Number:
  US08/672,936
• Time in Patent Office: 1,194 Days
• Field of Search: 364/488-491, 364/578
• US Class Current: 716/119
• CPC Class Codes: G06F 30/392 Floor-planning or layout, e.g. partitioning or placement

Abstract

A system for optimizing placement of a cell on a surface of a semiconductor chip is disclosed herein. The cells may belong to nets and may belong to neighborhoods. The system initially calculates affinities based on repositioning the cell. The system then combines affinities and repositions cells based on these combined affinities. The system then computes a cost function and repeats the combining, repositioning, and computing functions a predetermined number of times.

Claims

1. A method for optimizing placement of elements on a surface, wherein said elements are associated in groups, comprising the steps of:

(a) combining a plurality of affinities for relocating said elements to alternate locations on said surface;

(b) repositioning elements based on the combined affinities;

(c) computing a cost function;

(d) repeating said combining, repositioning, and computing steps a predetermined number of times;

(e) calculating an average cost function based upon all cost functions from said iterated computing steps;

(f) subsequently repeating said combining, repositioning, computing, repeating, and calculating steps until two consecutive average cost functions satisfy a given accuracy,

wherein affinity relates to preference to place a cell in a given alternate position on a cell-by-cell basis and cost function relates to quality of a placement.

2. The method of claim 1, wherein said combining step combines affinities for elements clustered together, affinities based on region capacity, and affinities based on groups crossing divisions between regions.

3. The method of claim 1, wherein said combining step comprises computing a total affinity based on affinities for elements clustered together and region capacities, and calculating affinity based on groups crossing divisions between regions and region capacities.

4. The method of claim 2, wherein said combining step comprises computing a total affinity by first combining affinities for elements clustered together, affinities based on region capacity, and affinities based on groups crossing divisions between regions, and subsequently combining affinities based on region capacity and affinities based on groups crossing divisions between regions.

5. The method of claim 4, where in the first affinity combination comprises weighing the affinity for elements clustered together greater than the affinities based on groups crossing divisions between regions.

6. The method of claim 4, wherein the first affinity combination comprises weighing the affinity for elements clustered together approximately equal to the affinities based on groups crossing divisions between regions.

7. The method of claim 4, wherein the first affinity combination comprises weighing the affinity for elements clustered together with a relatively small value and said predetermined number of times equals one.

8. A system for optimizing placement of cells on a surface of a semiconductor chip, wherein said cells are associated in nets and belong to neighborhoods, said system comprising:

(a) calculating a plurality of affinities based on repositioning said cells;

(b) combining said affinities;

(c) repositioning cells based on combined affinities;

(d) computing a cost function; and

(e) repeating said combining, repositioning, and computing steps a predetermined number of times,

wherein affinity relates to preference to place a cell in a given alternate position on a cell-by-cell basis and cost function relates to quality of a placement.

9. The system of claim 8, further comprising the step of calculating an average cost function based upon all cost functions from said iterated computing steps in (d) and (e).

10. The system of claim 9, further comprising the step of subsequently repeating said combining, repositioning, computing, repeating, and calculating steps until two consecutive average cost functions satisfy a given accuracy.

11. The system of claim 8, wherein said plurality of affinities comprises at least two of:

(a) affinities for cell neighborhoods;

(b) affinities based on region capacities; and

(c) affinities based on nets crossing divisions between regions.

12. The system of claim 11, wherein said combining step comprises computing a first affinity based on affinities for cell neighborhoods and region capacities, and calculating a second affinity based on groups crossing divisions between regions and region capacities.

13. The system of claim 11, wherein said combining step comprises computing a total affinity by first combining affinities for cell neighborhoods, affinities based on region capacities, and affinities based on groups crossing divisions between regions, and subsequently combining affinities based on region capacity and affinities based on groups crossing divisions between regions.

14. The system of claim 13, wherein the first affinity combination comprises weighing the affinities for cell neighborhoods greater than the affinities based on groups crossing divisions between regions.

15. The method of claim 13, wherein the first affinity combination comprises weighing the affinities for cell neighborhoods approximately equal to the affinities based on groups crossing divisions between regions.

16. The method of claim 13, wherein the first affinity combination comprises weighing the affinities for cell neighborhoods with a relatively small value and said predetermined number of times equals one.

17. A method for optimizing placement of elements on a surface, comprising the steps of:

(a) calculating affinities for repositioning elements;

(b) repositioning elements based on said affinities;

(c) computing a cost associated with repositioning the elements; and

(d) repeating said calculating, repositioning, and computing steps until the difference between costs for two subsequent iterations satisfies a predetermined accuracy,

wherein affinity relates to preference to place a cell in a given alternate position on a cell-by-cell basis and cost relates to quality of a placement.

18. A system for optimizing placement of an element on a surface, comprising:

(a) computing an affinity associated with repositioning said element;

(b) repositioning said element if said affinity satisfies a specified criterion;

(c) repeating said computing and repositioning steps a predetermined number of times;

(d) calculating a cost function; and

(e) repeating steps (a), (b), (c) and (d) until cost functions satisfy a predetermined accuracy on two subsequent iterations,

wherein affinity relates to preference to place a cell in a given alternate position on a cell-by-cell basis and cost function relates to quality of a placement.

19. A computer storage device having stored thereon a computer program comprising the following steps:

(a) computing an affinity associated with repositioning said element;

(b) repositioning said element if said affinity satisfies a specified criterion;

(c) repeating said computing and repositioning steps a predetermined number of times;

(d) calculating a cost function; and

(e) repeating steps (a), (b), (c) and (d) until cost functions satisfy a predetermined accuracy on two subsequent iterations,

wherein affinity relates to preference to place a cell in a given alternate position on a cell-by-cell basis and cost function relates to quality of a placement.