Simultaneous placement and routing (SPAR) method for integrated circuit physical design automation system

• US 5,875,117 A
• Filed: 04/23/1996
• Issued: 02/23/1999
• Est. Priority Date: 04/19/1994
• 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, Koford, James S., Boyle, Douglas B., Rostoker, Michael D., Jones, Edwin R.
• Primary Examiner(s): Teska, Kevin J.
• Assistant Examiner(s): PHAN, THAI Q
• Application Number:
  US08/636,349
• Time in Patent Office: 1,036 Days
• Field of Search: 364/489, 364/490, 364/491, 364/468.03, 395/800
• US Class Current: 716/124
• CPC Class Codes: G06F 30/392 Floor-planning or layout, e.g. partitioning or placement

Abstract

An initial placement of cells for an integrated circuit chip is decomposed into a hierarchial order of groups of cells. The groups are routed simultaneously using parallel processors, and the results are recomposed to provide a global routing that provides a detailed mapping of cell interconnect congestion in the placement. Areas of high congestion are identified, and a congestion reduction algorithm is applied using the parallel processors to alter the placement in these areas simultaneously. The overall fitness of the placement is then computed, and if it has not attained a predetermined value, the steps of identifying congested areas and applying the congestion reduction algorithm to these areas are repeated. The cumulative error created by altering the placement without repeating the global routing is estimated, and if it exceeds a predetermined value, the global routing is also repeated.

Claims

1. A method of optimizing a cell placement for an integrated circuit chip, comprising the steps of:

(a) routing said placement, wherein said placement includes congested areas and non-congested areas;

(b) Identifying said congested areas generated by said routing;

(c) selectively applying a congestion reduction algorithm only within said congested areas to alter said placement in said congested areas, wherein said congestion reduction algorithm comprises;

(c1) assessing individual cell cost values within said areas of increased congestion; and

(c2) relocating at least one cell having a highest cost value;

(d) computing a fitness of said placement; and

(e) if said fitness is below a predetermined value, repeating steps (b) to (d).

2. A method as in claim 1, in which step (e) further comprises repeating step (a).

3. A method as in claim 1, in which:

said placement comprises a plurality of nets of cells; and

step (a) comprises routing said nets in parallel.

4. A method as in claim 1, in which step (a) comprises the substeps of:

(f) dividing said placement into a plurality of predetermined areas; and

(g) routing said predetermined areas in parallel.

5. A method as in claim 1, in which step (c) comprises applying said congestion reduction algorithm to said areas of increased congestion in parallel.

6. A method as in claim 1, in which step (c) comprises moving selected cells out of said areas of increased congestion.

7. A method as in claim 1, in which step (c) comprises relocating selected cells such that routing thereof does not pass through said areas of increased congestion.

8. A method as in claim 7, in which step (c) comprises the substeps of:

(f) relocating said selected cells in a predetermined manner;

(g) rerouting nets of cells that include said selected cells; and

(h) repeating steps (f) and (g) if said rerouting passes through said areas of increased congestion.

9. A method as in claim 8, comprising performing steps (f), (g) and (h) on said selected cells and said nets respectively in parallel.

10. A method as in claim 8, in which step (g) comprises applying a Steiner tree algorithm for rerouting said nets.

11. A method as in claim 7, in which step (c) comprises selecting said cells for relocation as having high values of interconnect congestion.

12. A method as in claim 7, in which step (c) comprises selecting said cells for relocation if routing thereof passes through said areas of increased congestion.

13. A method as in claim 7, in which step (c) comprises selecting said cells for relocation if said cells are outside perimeters that surround other cells in nets to which said selected cells are connected respectively.

14. A method as in claim 13, in which step (c) comprises relocating said selected cells inside said perimeters respectively.

15. A method as in claim 14, in which step (c) comprises relocating said selected cells to centroids of said perimeters respectively.

16. A method as in claim 1, in which:

step (c) comprises relocating selected cells without rerouting said placement; and

the method further comprises the steps of;

(f) calculating an amount of error in said routing created by step (c); and

(g) repeating step (a) if said amount of error exceeds a predetermined value.

17. A method as in claim 16, in which steps (f) and (g) are performed after step (c).

18. A method as in claim 1, in which step (c) comprises the substeps of:

(f) dividing said placement into a plurality of predetermined areas; and

(g) applying said congestion reduction algorithm to said predetermined areas in parallel.

19. A method as in claim 18, in which step (g) comprises the substeps of:

(h) relocating selected cells without rerouting said placement;

(i) calculating amounts of error in said routing created by step (h) in each of said predetermined areas respectively; and

(j) repeating step (a) if said amount of error in any of said predetermined areas exceeds a predetermined value.

20. A method as in claim 1, further comprising the step, performed prior to step (a), of:

(f) decomposing said placement into a hierarchial order of groups of cells; and

step (a) comprises routing said groups of cells in parallel.

21. A physical design automation system for optimizing a cell placement for an integrated circuit chip, comprising:

a router for routing said placement, wherein said placement includes congested and non-congested areas;

a computer for performing operations of;

identifying said congested areas generated by said router;

selectively applying a congestion reduction algorithm only within said congested areas to alter said placement in said congested areas, wherein said congestion reduction algorithm comprises;

assessing individual cell cost values within said areas of increased congestion; and

relocating at least one cell having a highest cost value; and

computing a fitness of said placement; and

a controller for controlling the computer to repeat performing said operations until said fitness attains a predetermined value.

22. A system as in claim 21, in which the controller further controls the router to reroute said placement if said fitness is below said predetermined value.

23. A system as in claim 21, in which the router comprises a global router for globally routing said placement.

24. A system as in claim 21, in which:

said placement comprises a plurality of nets of cells; and

the router comprises a plurality of parallel processors for routing said nets in parallel.

25. A system as in claim 21, in which the router divides said placement into a plurality of predetermined areas, and comprises a plurality of parallel processors for routing said predetermined areas in parallel.

26. A system as in claim 21, in which the computer comprises a plurality of parallel processors for applying said congestion reduction algorithm to alter said placement in said areas of increased congestion in parallel.

27. A system as in claim 21, in which:

said congestion reduction algorithm comprises relocating selected cells such that routing thereof does not pass through said areas of increased congestion; and

the computer comprises a plurality of parallel processors for relocating said selected cells in parallel.

28. A method for optimizing cell placement over a global area on an integrated circuit chip, comprising the steps of:

fixing an initial cell placement within said global area;

routing wiring between cells placed by said fixing step;

determining wiring congestion for isolated areas within said global area;

identifying congested isolated areas having greater wiring congestion than other non-congested isolated areas;

selectively applying a congestion reduction algorithm only within said congested isolated areas, wherein said congestion reduction algorithm comprises;

accessing individual cell cost values within said areas of increased congestion; and

relocating at least one cell having a highest cost value;

computing a fitness of said cell placement; and

repeating said determining, identifying, selectively applying, and computing steps until said fitness improves to within a predetermined range.

29. A method as in claim 28, in which said repeating step further comprises said routing step.

30. A method as in claim 28, in which said routing step comprises the substeps of dividing said placement into a plurality of predetermined areas and routing said predetermined areas in parallel.

31. A method as in claim 28, in which said selectively applying step comprises relocating selected cells such that routing thereof does not pass through said congested isolated areas.

32. A method as in claim 31, in which said selectively applying step comprises the substeps of:

relocating said selected cells in a predetermined manner;

rerouting nets of cells that include said selected cells; and

repeating said relocating and rerouting steps if said rerouting passes through said congested isolated areas.