Method for providing performance-driven logic optimization in an integrated circuit layout design

Claim 3

A method as in claim 1, further comprising: dividing each cluster into a plurality of subclusters; and applying (b) through (f) on the subclusters.

Claim 4

A method as in claim 2, wherein said dividing step and said applying step are recursively applied to said subclusters until said subcluster include only primitive circuit elements.

Claim 5

A method as in claim 1, further comprising a placement optimization operation which reassigns circuit elements in a selected cluster to a second cluster mapped to a portion of said layout design adjacent the portion of said layout design mapped to said selected cluster.

Claim 6

A method as in claim 1, wherein said logic optimization operation inserts or deletes buffers between circuit elements.

Claim 7

A method as in claim 1, wherein said logic optimization operation provides one of several logic circuits performing substantially the same logic function.

Claim 8

A method as in claim 1, wherein said plurality of clusters are grouped into a plurality of cluster groups and wherein each of said mapping step, said performing a timing analysis step, said performing a logic optimization operation step and said routing step is applied concurrently to each cluster group using multiple processing units.

Claim 9

A method as in claim 1, wherein said clusters are divided so that connectivity amongst elements within a cluster are higher than connectivity amongst elements of different clusters.

Claim 10

A method as in claim 1, wherein said circuit elements are provided in a logic net list.

Claim 11

A method as in claim 1, further providing an initialization step which includes in said layout design test structures.

Claim 14

A method as in claim 11, wherein (e) comprises: (a1) partitioning circuit elements of said layout design into a plurality of clusters; (b1) mapping each cluster to a portion of said layout so as to obtain a placement; (c1) based on said placement, providing routing or routing estimates for circuit elements within each cluster and between circuit elements of different clusters; (d1) performing a timing analysis to provide estimates of interconnect delay between circuit elements within each cluster and between circuit elements of different clusters; (e1) performing a logic optimization operation on a selected one of said clusters to obtain a second gate-level logic design based on a cost function, said cost function being a function of said estimates of interconnect delays; (f1) based on said second gate-level logic design, providing an improved plurality of clusters; and (g1) repeating (b1)-(f1) until said cost function is less than a predetermined value.

Claim 15

A method as in claim 12, further comprising: dividing each cluster into a plurality of subclusters; and applying (b1) through (g1) on the subclusters.

Claim 16

A method as in claim 13, wherein said dividing step and applying step are recursively applied to said subclusters until said subcluster include only primitive circuit elements.

Claim 17

A method as in claim 12, further comprising a placement optimization operation which reassigns circuit elements in a selected cluster to a second cluster mapped to a portion of said layout design adjacent the portion of said layout design mapped to said selected cluster.

Claim 18

A method as in claim 12, wherein said logic optimization operation inserts or deletes buffers between circuit elements.

Claim 19

A method as in claim 12, wherein said logic optimization operation provides one of a plurality of logic circuits performing substantially the same logic function.

Claim 20

A method as in claim 12, wherein said plurality of clusters are grouped into a plurality of cluster groups and wherein each of said mapping step, said performing a timing analysis step, said performing a logic optimization operation step and routing step is applied concurrently to each cluster group using multiple processing units.

Claim 21

A method as in claim 12, wherein said clusters are divided so that connectivity amongst elements within a cluster are higher than connectivity amongst elements of different clusters.

Claim 22

A method as in claim 12, wherein said circuit elements are provided in a logic net list.

Claim 23

A method as in claim 12, further providing an initialization step which includes in said layout design test structures.

Claim 24

A method as in claim 12, wherein said logic optimization operation reassigns circuit elements in a selected cluster to a second cluster mapped to a portion of said layout design adjacent the portion of said layout design mapped to said selected cluster.

Claim 25

A parallel processing design automation system for optimizing an integrated circuit design, comprising: a plurality of central processing units; a memory for holding said integrated circuit design, said memory being accessible by each central processing units; a control program managing a task list, said task list specifying tasks for optimizing said integrated circuit design, said control program allocating each of said tasks in said task list to said central processing units said, control program supporting multithread execution; wherein each said task is one of placement, performance analysis, logic optimization and routing; wherein said control program provides a locking mechanism for data structures stored in said memory to support concurrent operations by said central processing units each accessing said integrated circuit design represented by said data structures.

Claim 26

A parallel processing design automation system for optimizing an integrated circuit design, comprising: a plurality of central processing units; a memory for holding said integrated circuit design, said memory being accessible by each central processing units; a control program managing a task list, said task list specifying tasks for optimizing said integrated circuit design, said control program allocating each of said tasks in said task list to said central processing units said, control program supporting multithread execution; wherein each said task is one of placement, performance analysis, logic optimization and routing; wherein said control program provides a locking mechanism for data structures stored in said memory to support concurrent operations by said central processing units each accessing said integrated circuit design represented by said data structures.

Claim 27

A parallel processing design automation system as in claim 23, further configured to include a placement program based upon an algorithm capable of parallel execution by said central processing units accessing said data structures.

Claim 28

A parallel processing design automation system as in claim 23, further configured to include a timing analysis program based upon an algorithm capable of parallel execution by said central processing units accessing said data structures.

Claim 29

A parallel processing design automation system as in claim 23, further configured to include a logic optimization program based upon an algorithm capable of parallel execution by said central processing units accessing said data structures.

Claim 30

A parallel processing design automation system as in claim 23, further configured to include a routing or routing estimation program based upon an algorithm capable of parallel execution by said central processing units accessing said data structures.

Claim 31

An integrated circuit having a layout design created by a process comprising: (a) partitioning circuit elements of a first gate-level logic design into a plurality of clusters; (b) mapping each cluster to a portion of said layout design so as to obtain a placement; (c) based on said placement, providing routing or routing estimates between circuit elements of each cluster and between circuit elements of different clusters; (d) based on said routing or routing estimates, performing a timing analysis to provide estimates of interconnect delays between circuit elements within each cluster and between circuit elements of different clusters; (e) performing a logic optimization operation on a selected one of said clusters to obtain a second gate-level logic design based on a cost function, said cost function being a function of said estimates of interconnect delays; (f) partitioning said second gate-level logic design to obtain a second plurality of clusters; and (g) repeating (b)-(f) until said cost function is less than a predetermined value.

Claim 32

An integrated circuit having a layout design created by a process comprising: (a) partitioning circuit elements of a first gate-level logic design into a plurality of clusters; (b) mapping each cluster to a portion of said layout design so as to obtain a placement; (c) based on said placement, providing routing or routing estimates between circuit elements of each cluster and between circuit elements of different clusters; (d) based on said routing or routing estimates, performing a timing analysis to provide estimates of interconnect delays between circuit elements within each cluster and between circuit elements of different clusters; (e) performing a logic optimization operation on a selected one of said clusters to obtain a second gate-level logic design based on a cost function, said cost function being a function of said estimates of interconnect delays; (f) partitioning said second gate-level logic design to obtain a second plurality of clusters; and (g) repeating (b)-(f) until said cost function is less than a predetermined value.

Claim 33

An integrated circuit as in claim 28, said layout design being created by applying further: dividing each cluster into a plurality of subclusters; and applying (b) through (f) on the subclusters.

Claim 34

An integrated circuit as in claim 29, wherein said dividing step and applying step are recursively applied to said subclusters until said subcluster include only primitive circuit elements.

Claim 35

An integrated circuit as in claim 28, further comprising a placement optimization operation which reassigns circuit elements in a selected cluster to a second cluster mapped to a portion of said layout design adjacent the portion of said layout design mapped to said selected cluster.

Claim 36

An integrated circuit as in claim 28, wherein said logic optimization operation inserts or deletes buffers between circuit elements.

Claim 37

An integrated circuit as in claim 28, wherein said logic optimization operation provides one of a plurality of logic circuits performing substantially the same logic function.

Claim 38

An integrated circuit as in claim 28, wherein said plurality of clusters are grouped into a plurality of cluster groups and wherein each of said step mapping step, said performing a timing analysis step, said performing a logic optimization operation step and said routing step is applied concurrently to each cluster group using multiple processing units.

Claim 39

An integrated circuit as in claim 28, wherein said clusters are divided so that connectivity amongst elements within a cluster are higher than connectivity amongst elements of different clusters.

Claim 40

An integrated circuit as in claim 28, wherein said logic optimization operation reassigns circuit elements in a selected cluster to a second cluster mapped to a portion of said layout design adjacent the portion of said layout design mapped to said selected cluster.

Claim 41

A computer system coupled to a shared memory accessible to a plurality of computer systems, comprising: a central processing unit; an operation system which provides services for creating and executing multiple threads; an interface to said shared memory, said interface reading tasks from and writing tasks to a task queue residing in said shared memory, and reading from and writing to design data of an integrated circuit residing in said shared memory; a memory containing application programs for processing said design data read from said shared memory and writing back said processed design data to said shared memory through said interface; wherein said applications programs include programs for performing in a method for optimizing a layout design based on a gate-level design: (a) partitioning circuit elements of said layout design into a plurality of clusters; (b) mapping each cluster to a portion of said layout design so as to obtain a placement; (c) based on said placement, providing routing or routing estimates between circuit elements of each cluster and between circuit elements of different clusters; (d) based on said routing or routing estimates, performing a timing analysis to provide estimates of interconnect delays between circuit elements within each cluster and between circuit elements of different clusters; (e) performing a logic optimization operation on a selected one of said clusters to obtain a second gate-level logic design based on a cost function, said cost function being a function of said estimates of interconnect delays; (f) based on said second gate-level logic design, providing an improved plurality of clusters; and (g) repeating (b)-(f) until said cost function is less than a predetermined value.

Claim 42

A computer system coupled to a shared memory accessible to a plurality of computer systems, comprising: a central processing unit; an operation system which provides services for creating and executing multiple threads; an interface to said shared memory, said interface reading tasks from and writing tasks to a task queue residing in said shared memory, and reading from and writing to design data of an integrated circuit residing in said shared memory; a memory containing application programs for processing said design data read from said shared memory and writing back said processed design data to said shared memory through said interface; wherein said applications programs include programs for performing in a method for optimizing a layout design based on a gate-level design: (a) partitioning circuit elements of said layout design into a plurality of clusters; (b) mapping each cluster to a portion of said layout design so as to obtain a placement; (c) based on said placement, providing routing or routing estimates between circuit elements of each cluster and between circuit elements of different clusters; (d) based on said routing or routing estimates, performing a timing analysis to provide estimates of interconnect delays between circuit elements within each cluster and between circuit elements of different clusters; (e) performing a logic optimization operation on a selected one of said clusters to obtain a second gate-level logic design based on a cost function, said cost function being a function of said estimates of interconnect delays; (f) based on said second gate-level logic design, providing an improved plurality of clusters; and (g) repeating (b)-(f) until said cost function is less than a predetermined value.

Claim 43

A computer system as in claim 37, said method further comprising: dividing each cluster into a plurality of subclusters; and applying (b) through (f) on the subclusters.

Claim 44

A computer system as in claim 38, wherein said dividing step and applying step are recursively applied to said subclusters until said subcluster include only primitive circuit elements.