Scalable architecture for IP block integration

• US 9,847,783 B1
• Filed: 10/13/2015
• Issued: 12/19/2017
• Est. Priority Date: 10/13/2015
• Status: Active Grant

• Current Assignee: Altera Corporation (Intel Corporation)
• Original Assignee: Altera Corporation (Intel Corporation)
• Inventors: Teh, Chee Hak, Rahman, Arifur, Grenier, Richard Arthur
• Primary Examiner(s): Tran, Anh
• Application Number:
  US14/882,065
• Time in Patent Office: 798 Days
• Field of Search: None
• CPC Class Codes: H03K 19/17736 Structural details of routing resources
• CPC Class Codes: H03K 19/1776 for memories

Abstract

A scalable circuit architecture for programmable circuitry is provided. Intellectual property (IP) blocks may be integrated into a circuit design and may be formed next to programmable logic sectors on which user logic functions are implemented. IP blocks may receive configuration data from sub-system managers (SSMs) that serve as a local configuration source for the IP blocks. Configurable endpoints in the IP blocks may be represented by memory mapped addresses that may be decoded by pipeline decoders having delay elements that prevent read data collision. A reroute layer may serve as an interface between IP blocks and one or more programmable logic sectors. The reroute layer may have a higher number of connections at a logic sector interface compared to the number of connections at an IP block interface. An IP block may route clock signals having different frequencies to respective different rows or regions in the programmable logic sectors.

Claims

1. An integrated circuit, comprising:

a plurality of programmable logic blocks;

an intellectual property (IP) circuit block that is adjacent to at least one of the plurality of programmable logic blocks and that comprises a plurality of endpoints;

a local configuration source that is adjacent to the IP circuit block and that is used to address the plurality of endpoints in the IP circuit block; and

a pipeline stage that is coupled to the plurality of endpoints and that routes access commands from the local configuration source to at least one of the plurality of endpoints.

2. The integrated circuit defined in claim 1, wherein the plurality of endpoints are assigned an address based on a memory mapping, and wherein the pipeline stage comprises:

a pipeline decoder circuit that receives memory mapped address for an endpoint from the local configuration source and that addresses the endpoint by decoding the memory mapped address.

3. The integrated circuit defined in claim 2, wherein the pipeline decoder circuit comprises:

a programmable delay element that exhibits a delay that is based on the distance of the pipeline decoder circuit from the local configuration source.

4. The integrated circuit defined in claim 3, wherein the pipeline decoder circuit comprises a first pipeline decoder circuit, wherein the programmable delay element comprises a first programmable delay element, and wherein the integrated circuit further comprises:

a second pipeline decoder circuit that is interposed between the first pipeline decoder circuit and the local configuration source, wherein the second pipeline decoder circuit comprises;

a second programmable delay element that exhibits a delay that is different than the delay exhibited by the first programmable delay element.

5. The integrated circuit defined in claim 2, wherein the local configuration source reads from the plurality of endpoints during a calibration mode of the integrated circuit.

6. The integrated circuit defined in claim 1, wherein the local configuration source addresses a subset of the plurality of endpoints using a single address.

7. Programmable circuitry, comprising:

logic fabric having programmable blocks arranged in rows and columns;

an intellectual property (IP) circuit block having multiple endpoints and a clock generator that provides at least one clock signal to the logic fabric;

a local configuration source having a configuration clock generator that generates a configuration clock signal for configuring the IP circuit block, wherein the local configuration source addresses the multiple endpoints in the IP circuit block; and

a pipeline stage that is coupled to the multiple endpoints and that routes access commands from the local configuration source to at least one of the plurality of endpoints.

8. The programmable circuitry defined in claim 7, wherein the at least one clock signal comprises a full-rate clock signal, wherein the programmable circuitry comprises:

an interconnect path that routes the full-rate clock signal to a region of the logic fabric that is adjacent to the IP circuit block, wherein the region of the logic fabric is bounded by a row of programmable blocks in the logic fabric.

9. The programmable circuitry defined in claim 7, wherein the region of the logic fabric comprises:

user logic at a given row of the logic fabric that is clocked by the full-rate clock signal; and

a drop-back signal path that extends to the given row and that feeds back into the IP circuit block.

10. The programmable circuitry defined in claim 9, wherein the IP circuit block comprises:

clock phase alignment circuitry that receives a drop-back signal from the drop-back signal path and that adjusts the full-rate clock signal based on the drop-back signal.

11. The programmable circuitry defined in claim 7, wherein the at least one clock signal further comprises a divide-by-N clock signal that is based on the full-rate clock signal, wherein the interconnect path comprises a first interconnect path, wherein the region of the logic fabric comprises a first region, and wherein the programmable circuitry further comprises:

a second interconnect path that routes the divide-by-N clock signal to a second region of the logic fabric that is adjacent to the first region of the logic fabric.

12. The programmable circuitry defined in claim 11, wherein the second region of the logic fabric comprises:

user logic at a given row of the logic fabric that is clocked by the divide-by-N clock signal; and

a drop-back signal path that extends to the given row and that feeds back into the IP circuit block.

13. The programmable circuitry defined in claim 12, wherein the IP circuit block comprises:

clock phase alignment circuitry that receives a drop-back signal from the drop-back signal path and that adjusts the divide-by-N clock signal based on the drop-back signal.

14. Circuitry, comprising:

an array of programmable logic blocks;

an intellectual property (IP) block having multiple endpoints that is adjacent to the array of logic blocks;

a pipelined reroute layer having a first set of connections between a first interface of the reroute layer and the multiple endpoints in the IP block, and further having a second set of connections between a second interface of the reroute layer and the array logic blocks, wherein the first set of connections has a first connection density that is higher than a second connection density of the second set of connections; and

a local configuration source that addresses the multiple endpoints in the IP block and that configures the pipelined reroute layer to map a first connection at the first interface to a second connection at the second interface, and wherein the pipelined reroute layer routes signals between the first and second connections.

15. The circuitry defined in claim 14, wherein the IP block has a plurality of connections, and wherein the first interface of the reroute layer comprises:

connection terminals that are aligned with the plurality of connections in the IP block.

16. The circuitry defined in claim 14, wherein the second connection is the closest connection at the second interface to the first connection at the first interface.

17. The circuitry defined in claim 14, wherein a portion of the second interface is obstructed from directly interfacing with the array of logic blocks, and wherein the first connection is formed opposite the obstructed portion of the second interface.