Method for partitioning of connected circuit components before placement in one or more integrated circuits

• US 5,202,840 A
• Filed: 12/19/1990
• Issued: 04/13/1993
• Est. Priority Date: 12/19/1990
• Status: Expired due to Term

• Current Assignee: NXP B.V. (NXP Semiconductors NV)
• Original Assignee: VLSI Technology, Inc. (Koninklijke Philips N.V.)
• Inventors: Wong, Dale M.
• Primary Examiner(s): Trans, Vincent N.
• Application Number:
  US07/630,284
• Time in Patent Office: 846 Days
• Field of Search: 364/488, 364/489, 364/490, 364/491
• US Class Current: 716/123
• CPC Class Codes: G06F 30/392 Floor-planning or layout, e.g. partitioning or placement

Abstract

In a method for partitioning components of an integrated circuit into clusters in preparation for placement of the clusters onto an integrated circuit, a connectivity cost is assigned to every pair of connected components. Then, beginning with a pair of connected components with a lowest assigned connectivity cost and proceeding to a pair of connected components with a highest assigned connectivity cost, the components are placed in a cluster tree so that when merging cluster tree nodes, a cluster tree node with a higher connectivity cost is made a parent cluster tree node of a cluster tree node with a lower connectivity cost. The cluster tree may then traversed to partition the components into clusters. The clusters can be optimized by moving individual components to clusters when such a move results in a lower incremental connectivity cost. Once the components have been partitioned into clusters, the clusters of components may be placed on an integrated circuit.

Claims

1. A computer implemented method for partitioning components of an integrated circuit into clusters in preparation for placement of the clusters onto an integrated circuit, the method implemented by a computer and comprising the steps of:

(a) assigning, by the computer, a connectivity cost to every pair of connected components;

(b) beginning with a pair of connected components with a lowest assigned connectivity cost and proceeding to a pair of connected components with a highest assigned connectivity cost, placing, by the computer, the components in a cluster tree so that when merging cluster tree nodes, a cluster tree node with a higher connectivity cost is made a parent cluster tree node of a cluster tree node with a lower connectivity cost; and

,

(c) traversing, by the computer, the cluster tree to partition the components into clusters.

2. A computer implemented method as in claim 1 wherein the connectivity cost in step (a) is calculated using the formula:

space="preserve" listing-type="equation">cost=(total.sub.-- connections--total.sub.-- interconnections)/total.sub.-- connections.

3. A computer implemented method as in claim 1 wherein step (b) includes the following substeps:

(b.1) sorting the pairs of connected components based on the connectivity cost associated with each pair; and

,

(b.2) for each pair of connected components beginning with the pair of connected components with the lowest assigned connectivity cost and proceeding to the pair of connected components with the highest assigned connectivity cost, if both components in the pair of connected components are not in a same class of connected components, performing the following substep;

(b.2.1) merging a first class containing a first component of the pair of connected components into a second class containing a second component of the pair of connected components to form a cluster tree node and making the first class a child cluster tree node of the cluster tree node when a connectivity cost of the first class is not greater than a connectivity cost of the second class.

4. A computer implemented method as in claim 3 wherein in substep (b.1) the sorting is done using a heap sort algorithm.

5. A computer implemented method as in claim 3 wherein the connectivity cost in step (a) is calculated using the formula:

space="preserve" listing-type="equation">cost=(total.sub.-- connections--total.sub.-- interconnections)/total.sub.-- connections.

6. A computer implemented method as in claim 3 additionally comprising the step of:

(d) optimizing, by the computer, the clusters by moving individual components to clusters when such a move results in a lower incremental connectivity cost.

7. A computer implemented method as in claim 6 additionally comprising the step of:

(e) placing the clusters of components on an integrated circuit.

8. A computer implemented method as in claim 3 wherein in step (c) a criterion used in partitioning components is size of cluster.

9. A computer implemented method as in claim 3 wherein in step (c) a criterion used in partitioning components is a maximum number of external connections from each cluster.

10. A computer implemented method as in claim 3 wherein in step (c) a criterion used in partitioning components is a packaging cost.

11. A computer implemented method as in claim 3 wherein the substep (b.2) the connectivity cost for each class is a maximum of connectivity costs associated with a pair of connected components in the class.

12. A computer implemented method as in claim 3 wherein in substep (b.2) the connectivity cost for each class is calculated by summing connectivity costs for all component pairs in the class.

13. A computer implemented method as in claim 3 wherein in substep (b.2) the connectivity cost for a class is calculated by subtracting a number of connections which are connected to at least two components in the class from a number of connections which are connected to at least one component in the class and dividing a resulting difference by the number of connections which are connected to at least one component in the class.

14. A computer implemented method as in claim 1 additionally comprising the step of:

(d) optimizing, by the computer, the clusters by moving individual components to clusters when such a move results in a lower incremental connectivity cost.

15. A computer implemented method as in claim 14 additionally comprising the step of:

(e) placing the clusters of components on an integrated circuit.

16. A computer implemented method as in claim 1 wherein the step (c) a criterion used in partitioning components is size of cluster.

17. A computer implemented method as in claim 1 wherein in step (c) a criterion used in partitioning components is a maximum number of external connections from each cluster.

18. A computer implemented method as in claim 1 wherein in step (c) a criterion used in partitioning components is a packaging cost.