IPC분류정보
국가/구분 |
United States(US) Patent
등록
|
국제특허분류(IPC7판) |
|
출원번호 |
US-0802669
(2010-06-11)
|
등록번호 |
US-9230047
(2016-01-05)
|
발명자
/ 주소 |
- Van Antwerpen, Babette
- Baeckler, Gregg William
|
출원인 / 주소 |
|
대리인 / 주소 |
|
인용정보 |
피인용 횟수 :
0 인용 특허 :
16 |
초록
▼
A method for designing a system on a target device is disclosed. A partition in the system with a plurality of instances from an extraction netlist is identified. Synthesis optimizations are performed on the partition to generate a synthesis optimization solution. The synthesis optimization solution
A method for designing a system on a target device is disclosed. A partition in the system with a plurality of instances from an extraction netlist is identified. Synthesis optimizations are performed on the partition to generate a synthesis optimization solution. The synthesis optimization solution is applied to the plurality of instances in the system.
대표청구항
▼
1. A method for designing a system on a target device, the method comprising: identifying a partition having a plurality of instances in the system from an extraction netlist, wherein the partition implements a component from the extraction netlist and each instance of the partition implements an id
1. A method for designing a system on a target device, the method comprising: identifying a partition having a plurality of instances in the system from an extraction netlist, wherein the partition implements a component from the extraction netlist and each instance of the partition implements an identical component having a same functionality as the component, wherein identifying the partition comprises identifying an initial partition, and expanding the initial partition, wherein expanding the initial partition comprises building a merge set, identifying a merge from the merge set having a highest savings in cost relative to other merges from the merge set, and applying the merge identified with the highest savings in cost;performing synthesis optimizations on the partition to generate a synthesis optimization solution; andapplying the synthesis optimization solution for the partition selectively to the plurality of instances of the partition in the system, wherein at least one of the identifying, performing, and applying is performed by a processor. 2. The method of claim 1, wherein identifying the initial partition comprises identifying the component from the extraction netlist and its connection to other initial partitions. 3. The method of claim 1, wherein the cost is a function of a number of components in a partition and a number of instances of the partition that exists. 4. The method of claim 1, wherein the cost is a function of a number of connections in a partition. 5. The method of claim 1, wherein the cost is a function of a number of inputs in a partition. 6. The method of claim 1, further comprising: identifying another merge with a partition with a next highest cost savings; andapplying the another merge with the next highest cost savings relative to the other merges in the merge set. 7. The method of claim 1, further comprising applying a merge resulting in a highest number of inputs and outputs saved when more than one merge is identified with a partition with a highest cost savings. 8. The method of claim 1, further comprising applying a merge resulting in a highest number of connections absorbed when more than one merge is identified with a partition with a highest cost savings. 9. The method of claim 1, wherein performing synthesis optimizations on the partition to generate a synthesis optimization solution comprises: identifying a number of instances of the partition in the extraction netlist; andapplying synthesis optimization procedures in response to the number of instances. 10. The method of claim 9, wherein when the number of instances of the partition equals or exceeds a threshold number, a first class of optimization procedures is applied to aggressively optimize the partition that require a larger amount of time to perform than a second class of optimization procedures applied to partitions that have fewer instances than the threshold number. 11. The method of claim 1, further comprising: creating a hierarchical netlist after identifying the partition with a top-level netlist that includes the extraction netlist with each instance of the partition represented as a generic component, and a bottom-level netlist that includes the partition. 12. The method of claim 11, wherein applying the synthesis optimization solution to the plurality of instances in the system comprises merging the optimized bottom-level netlist with the optimized top-level netlist. 13. A non-transitory computer-readable medium including a sequence of instructions stored thereon for causing a processor to execute a method comprising: identifying a plurality of partitions in a system and a number of instances of each partition of the plurality of partitions in an extraction netlist, wherein each partition of the plurality of partitions implements a component in the system; andperforming synthesis optimizations on the plurality of partitions to generate optimization solutions, wherein a class of optimization procedures performed on one partition of the plurality of partitions is determined in response to the number of instances of the one partition of the plurality of partitions which occurs in the extraction netlist. 14. The non-transitory computer-readable medium of claim 13, wherein a first class of optimization procedures is performed on the one partition of the plurality partitions if the one partition of the plurality of partitions has greater than or equal to a threshold number of instances, and a second class of optimization procedures is performed on the one partition of the plurality of partitions if the one partition of the plurality of partitions has fewer than the threshold number of instances. 15. The non-transitory computer-readable medium of claim 14, wherein the first class of optimization procedures requires a larger amount of time to complete than the second class of optimization procedures. 16. The non-transitory computer-readable medium of claim 13, wherein the number of instances is in a magnitude of 100. 17. The non-transitory computer-readable medium of claim 14, wherein the threshold number of instances is in a magnitude of 1000. 18. The non-transitory computer-readable medium of claim 13, wherein the method further comprises applying the optimization solutions to the instances of each partition of the plurality of partitions in the system. 19. The non-transitory computer-readable medium of claim 13, wherein identifying the plurality of partitions comprises: identifying initial partitions; andexpanding the initial partitions. 20. The non-transitory computer-readable medium of claim 19, wherein identifying the initial partitions comprises identifying a component from the extraction netlist and its connection to other initial partitions. 21. The non-transitory computer-readable medium of claim 19, wherein expanding the initial partition comprises: building a merge set;identifying a merge from the merge set having a highest savings in cost relative to other merges from the merge set; andapplying the merge identified with the highest savings in cost. 22. The non-transitory computer-readable medium of claim 21, wherein cost is a function of a number of gates in a partition and a number of instances of the partition that exists. 23. The non-transitory computer-readable medium of claim 13, wherein the synthesis optimizations comprises factoring. 24. A system designer comprising: a synthesis unit having a partition identification unit that identifies a partitions in a system with a plurality of instances in an extraction netlist, and an optimization unit that performs synthesis optimizations on the partitions to generate a synthesis optimization solution and to apply the synthesis optimization solutions only to the plurality of instances in the system, wherein the partition implements a component in the system, and wherein each instance of the partition implements an identical component in the system, wherein the optimization unit applies a first class of optimization procedures on a first partition having a number of instances that exceed a redetermined value and a second class of optimization procedures on a second partition having a number of instances that does not exceed the predetermined value. 25. The system designer of claim 24, wherein the partition identification unit identifies an initial partition by identifying a component from the extraction netlist, and expands the initial partition by building a potential merge set, identifies a merge with a partition with a highest cost savings, and applies the merge with the highest cost savings. 26. The system designer of claim 24, further comprising: a placement unit that places the system on a target device; anda routing unit that routes the system on the target device.
※ AI-Helper는 부적절한 답변을 할 수 있습니다.