Method for improving area in reduced programmable logic devices

• US 6,526,563 B1
• Filed: 07/13/2000
• Issued: 02/25/2003
• Est. Priority Date: 07/13/2000
• Status: Expired due to Term

• Current Assignee: Xilinx, Inc. (Advanced Micro Devices, Inc.)
• Original Assignee: Xilinx, Inc. (Advanced Micro Devices, Inc.)
• Inventors: Baxter, Glenn A.
• Primary Examiner(s): Siek, Vuthe
• Assistant Examiner(s): LEVIN, NAUM B
• Application Number:
  US09/615,108
• Time in Patent Office: 957 Days
• Field of Search: 716/1-21, 326/37-41
• US Class Current: 716/103
• CPC Class Codes: G06F 30/39 Circuit design at the physical level physical level design for reconfigurable circuits G06F30/347
• CPC Class Codes: H01L 27/11803 using field effect technology

Abstract

Method and circuits to create reduced field programmable gate arrays (RFPGA) from the configuration data of field programmable gate arrays (FPGA) are disclosed. The configurable elements of the FPGA are replaced with standard cell circuits that reproduce the functionality of the configured FPGA. Specifically, reduced logic blocks are derived from the configuration data of configurable logic blocks. Similarly, reduced input/output blocks and reduced matrices are derived from the configuration data for input/output blocks and programmable switch matrices of the FPGA, respectively. The reduced logic blocks are arranged in a similar layout to the original CLBs so that timing relationships remain similar in the RFPGA and FPGA. The actual timing of the RFPGA can be modified by increasing or decreasing the timing delay on various signal paths based on the FPGA design or additional timing constraints. To reduce the time required to generate RFPGAs, a database can be used to contain configurable logic block models and the corresponding reduced logic block models. The database can be expanded as new reduced logic block models are created for configurable logic block models that were not in the database. Similarly, a database can be used for the input/output blocks and programmable switch matrices of an FPGA.

Claims

1. A method for creating a reduced logic block model from a configurable logic block model, the method comprising:

defining wired paths in place of a plurality of configured selector circuits in the configurable logic block model;

defining a non-configurable logic circuit in place of configured function generators of the configurable logic block model.

2. The method of claim 1, wherein the non-configurable logic circuit is a multiplexer.

3. The method of claim 2, wherein the multiplexer is sized to decode the number of input signals.

4. The method of claim 3, wherein the number of input signals four.

5. The method of claim 2, wherein the multiplexer has one or more input signals coupled to logic low.

6. The method of claim 2, wherein the multiplexer has one or more input signals coupled to logic high.

7. The method of claim 2, wherein each input signal of the multiplexer are coupled to logic low or logic high.

8. The method of claim 1, further comprising placing the wired paths and non-configurable logic circuits to minimize horizontal area.

9. The method of claim 1, further comprising placing the wired paths and non-configurable logic circuits to minimize vertical area.

10. The method of claim 1, further comprising placing the wired paths and non-configurable logic circuits to minimize area.

11. The method of claim 1, further comprising placing the wired paths and non-configurable logic circuits to control timing on one or more signal paths.

12. A method for creating a reduced input/output block model from a configurable input/output block model, the method comprising:

defining wired paths in place of a plurality of configured selector circuits in the configurable input/output block model;

defining a non-configurable logic circuit in place of configured circuits of the configurable input/output block model.

13. The method of claim 12, further comprising placing the wired paths and non-configurable logic circuits to minimize horizontal area.

14. The method of claim 12, further comprising placing the wired paths and non-configurable logic circuits to minimize vertical area.

15. The method of claim 12, further comprising placing the wired paths and non-configurable logic circuits to minimize area.

16. The method of claim 12, further comprising placing the wired paths and non-configurable logic circuits to control timing on one or more signal paths.

17. A method for creating a reduced matrix model from a programmable switch matrix model having a plurality of programmable interconnect points, wherein each programmable interconnect point can be configured to interconnect a plurality of wires, the method comprising:

determining the configuration of each programmable interconnect point;

defining a set of one or more wired paths in place of each programmable interconnect point configured to interconnect any of said wires to create a plurality of sets of wired paths; and

removing unused programmable interconnect points.

18. The method of claim 17, wherein the plurality of sets of wired paths comprises vias.

19. The method of claim 17, further comprising moving the sets of wired paths to reduce the size of the reduced matrix.

20. The method of claim 17, wherein the programmable switch matrix model uses a plurality of vertical wire channels and a plurality of horizontal wire channels, and wherein each programmable interconnect point is in a distinct horizontal wire channel and a distinct vertical wire channel.

21. The method of claim 20, wherein each set of wired paths is in the same vertical wire channel and horizontal wire channel as the replaced programmable interconnect point.

22. The method of claim 21, further comprising moving a set of wired paths to a vacated vertical wire channel and a vacated horizontal wire channel previously used by a removed programmable interconnect point.

23. The method of claim 21, further comprising moving a first set of wired paths to share a vertical wire channel with a second set of wired paths.

24. The method of claim 21, further comprising moving a first set of wired paths to share a horizontal wire channel with a second set of wired paths.

25. The method of claim 17, further comprising defining diagonal wires to connect the reduced matrix model to a second reduced matrix model.

26. A method for creating a reduced logic block model from a configurable logic block model, the method comprising the steps of:

extracting configuration data for input/output blocks (IOBs), configurable logic blocks (CLBs), and programmable switch interconnect including programmable switch matrices (PSMs);

generating IOB and CLB models;

estimating an area required for reduced input/output blocks (RIOBs), reduced logic blocks (RLBs), reduced interconnect, and reduced matrices (RMs) by approximating an area for each RIOB, RLB, and reduced interconnect;

building a routing grid for RLBs inside a routing ring and a routing grid for RIOBs outside the routing ring building RLB and RIOB models using the IOB and CLB models generated and the area estimated for the RIOBs, RLBS, and reduced interconnect;

deriving a draw list for the reduced interconnect;

interconnecting RLBs using a route reduced interconnect;

building the routing ring;

routing connections between RLB and the routing ring;

and placing the RIOBs around the routing ring.

27. The method of claim 26, wherein the steps of extracting configuration data for IOBs, CLBs, and programmable switch interconnect are done in parallel.

28. The method of claim 26, wherein the steps of extracting configuration data for IOBs, CLBS, and programmable switch interconnect are done in series.

29. The method of claim 26, wherein the step of estimating an area further comprises determining an optimal grid pitch for a non-uniform array of RLBs.

30. The method of claim 26, wherein the step of estimating an area further comprises determining an optimal grid pitch for accommodating the largest row and column respectively.

31. The method of claim 26, wherein the step of estimating an area further comprises determining an optimal grid pitch by determining the actual size of each row and column using a variable grid pitch.

32. The method of claim 26, wherein the step of deriving a draw list for the reduced interconnect further comprises the step of deriving draw lists for reduced matrices.

33. The method of claim 26, wherein each wire in the routing ring has a first endpoint on an internal routing grid and a second endpoint on the external routing grid.

34. The method of claim 26, wherein the method further comprises the step of adding demarcation lines to enable placement of scribe lines on a silicon wafer during manufacture of an integrated circuit die.

35. The method of claim 26, wherein the method further comprises the step of extracting FPGA timing information from an FPGA design and placing and routing timing buffers based on additional timing constraints.

36. The method of claim 35, wherein the method further comprises the step of extracting reduced FPGA timing information.

37. The method of claim 36, wherein the method further comprises the step of comparing the FPGA timing information with the reduced FPGA timing information to adjust at least one selected from the group of timing constraints, placement of timing buffers, and routing of timing buffers.