# US 10,320,393 B2 · Dynamic multicycles for core-periphery timing closure

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 10,320,393 B2
- Document ID: US10320393B2
- Status: Granted
- Kind: B2
- Application number: 15/719,194
- Filed: 2017-09-28
- Published: 2019-03-28
- Granted: 2019-06-11
- Earliest claimed: 2017-09-28
- Assignees: Intel Corporation
- Applicants: Intel Corporation
- Inventors: Navid Azizi; Aditi Kumaraswamy; Emily Alexandra Ng
- Web page: https://patentgrep.com/patent/US10320393B2

## CPC

- H03K 19/17764: for reliability
- G06F 30/34: for reconfigurable circuits, e.g. field programmable gate arrays [FPGA] or programmable logic devices [PLD]
- H03K 5/15006: {with two programmable outputs}
- G06F 2119/12: Timing analysis or timing optimisation
- H03K 19/1735: {by wiring, e.g. uncommitted logic arrays}
- H03K 19/1737: {using multiplexers (H03K19/1738 takes precedence)}
- H03K 19/17728: Reconfigurable logic blocks, e.g. lookup tables
- H03K: PULSE TECHNIQUE (measuring pulse characteristics G01R; modulating sinusoidal oscillations with pulses H03C; transmission of digital information H04L; discriminator circuits detecting phase difference between two signals by counting or integrating cycles of oscillation H03D3/04; automatic control, starting, synchronisation or stabilisation of generators of electronic oscillations or pulses where the type of generator is irrelevant or unspecified H03L; coding, decoding or code conversion, in general H03M)
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

Methods and systems for timing analysis and closure during logic synthesis of synchronous digital circuitry are provided, which may be used to prevent timing conflicts in logic designs that may have data transfers between regions with substantial clock skew. In programmable logic devices having hardened circuitry and programmable fabric, data transfers between memory elements in hardened circuitry and programmable fabric may be subject to substantial clock skews and unknown latencies. Embodiments may employ pre-calculated latencies that may be stored in a file and/or a database, and dynamically retrieved during timing synthesis to determine multicycle constraints to mitigate latencies. Embodiments may employ destination multicycle constraints, which use as reference the clock waveforms delayed due to latency.

## Claims

### Claim 1

An article of manufacture comprising one or more non-transitory machine-readable media comprising instructions to cause a processor to synthesize logic circuitry to be implemented in a programmable logic device, the instructions comprising:
receiving a logic design;
identifying a data transfer of the logic design, wherein the data transfer comprises a source memory element and a destination memory element, and wherein the source memory element and the destination memory element receive a clock signal from a common clock;
calculating a clock skew between the source memory element and the destination memory element;
adjusting the data transfer to assign a launch edge of the clock signal as received by the source memory element and a setup edge and a hold edge of the clock signal as received by the destination memory element based on the clock skew; and
programming the programmable logic device to perform at least the data transfer.

### Claim 2 (depends on 1)

The article of manufacture of claim 1, wherein the source memory element and the destination memory element comprise a look-up table, a register, a latch, a flip flop, or any combination therein.

### Claim 3 (depends on 1)

The article of manufacture of claim 1, wherein the programmable logic device comprises a programmable fabric that comprises the source memory element and a hardened circuitry region that comprises the destination memory element.

### Claim 4 (depends on 3)

The article of manufacture of claim 3, wherein calculating the clock skew comprises retrieving latency information associated with the hardened circuitry region from a data source, wherein the data source comprises a file or a database.

### Claim 5 (depends on 3)

The article of manufacture of claim 3, wherein the hardened circuitry region comprises circuitry that implements a communication protocol.

### Claim 6 (depends on 1)

The article of manufacture of claim 1, wherein calculating the clock skew comprises identifying a latency of the source memory element, identifying a latency of the destination memory element, and calculating a phase shift between the latency of the source memory element and the latency of the destination memory element.

### Claim 7 (depends on 6)

The article of manufacture of claim 6, wherein adjusting the data transfer comprises:
assigning the launch edge of the clock signal as received by the source memory element based on the latency of the source memory element; and
assigning the setup edge and the hold edge of the clock signal as received by the destination memory element based on the clock skew.

### Claim 8 (depends on 1)

The article of manufacture of claim 1, wherein adjusting the data transfer comprises calculating a slack of the data transfer.

### Claim 9

A method to synthesize logic circuitry to be implemented in a programmable logic device comprising:
receiving a logic design;
identifying a plurality of data transfers of the logic design, wherein each data transfer comprises a respective source memory element and a respective destination memory element;
calculating, for each data transfer of the logic design, a respective clock skew based on a respective source latency of a clock signal received by the respective source memory element and a respective destination latency of a clock signal received by the respective destination memory element;
assigning, for each data transfer, a respective launch edge of a clock signal received by the respective source memory element and a respective hold edge of a clock signal received by the respective destination memory element based on the respective clock skew; and
programming the programmable logic device to perform at least the data transfer.

### Claim 10 (depends on 9)

The method of claim 9, wherein the respective source latency, the respective destination latency, or both, are retrieved from a file or a database comprising a pre-calculated latency.

### Claim 11 (depends on 9)

The method of claim 9, wherein assigning the respective hold edge comprises determining an edge of the clock signal received by the respective destination memory element that follows the respective launch edge based on the respective clock skew.

### Claim 12 (depends on 9)

The method of claim 9, comprising:
calculating, for each data transfer, a data transfer slack comprising a difference between the respective launch edge and the respective hold edge; and
iteratively determining a figure of merit based on the plurality of the calculated data transfer slacks and reassigning, for each data transfer, the respective launch edge and the respective hold edge based on the figure of merit until the figure of merit crosses a threshold.

### Claim 13 (depends on 12)

The method of claim 12, wherein the figure of merit comprises a number of transfers with positive slack, a number of transfers with negative slack, a maximum absolute negative slack, or any combination thereof.

### Claim 14 (depends on 9)

The method of claim 9, wherein the programmable logic device comprises a programmable fabric region and a hardened circuitry region, and wherein at least one data transfer of the logic design comprises a respective destination memory element disposed in the programmable logic region and a respective source memory element disposed in the hardened circuitry region.

### Claim 15 (depends on 14)

The method of claim 14, wherein the hardened circuitry region comprises low-voltage differential signaling.

### Claim 16 (depends on 9)

The method of claim 9, comprising determining a clock tree that provides clock signals to a plurality of logic elements based on the logic design.

### Claim 17

An article of manufacture, comprising instructions to cause a processor to:
communicate with a programmable logic device that comprises a programmable fabric and fixed circuitry;
receive a logic design, the logic design comprising at least one data transfer between a memory element in the programmable fabric and a memory element in the fixed circuitry;
retrieve a latency information associated with the fixed circuitry from a file, a database, or any combination thereof;
determine a multicycle constraint based on the latency information and the logic design;
assign a launch edge and a hold edge of the data transfer based on the multicycle constraint; and
program the programmable logic device.

### Claim 18 (depends on 17)

The article of manufacture of claim 17, wherein retrieving the latency information comprises performing a procedural call to the file, the database, or any combination thereof.

### Claim 19 (depends on 17)

The article of manufacture of claim 17, wherein the fixed circuitry that comprises a communication protocol.

### Claim 20 (depends on 17)

The article of manufacture of claim 17, wherein the memory element in the programmable fabric and the memory element in the fixed circuitry comprises a register.
