Top-Down Physical Design of Soft FPGA Fabrics for Global Timing

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing synthesizable FPGA fabrics face challenges in ensuring global timing views across multiple tiles, leading to combinational loops and long false paths, which complicates integration with ASIC blocks and requires manual buffering and floorplanning, making them incompatible with typical SoC design schedules.

Innovation Solution

A top-down design methodology is employed to provide a global timing view of the FPGA fabric to EDA tools, restructuring switch-box multiplexers to enable static timing analysis exceptions, eliminating the need for manual buffering and floorplanning, and allowing for a push-button ASIC design flow.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of manufacture

If bottom-up methodology is used to synthesize individual tiles in isolation, then ease of manufacture is improved, but device complexity increases due to lack of global timing view across multiple tiles

Engineering Contradiction:
Improveease of synthesizing individual tilesVSAvoidcomplexity of ensuring fabric-level performance targets
Core Design Contradiction:
Ease of manufactureVSDevice complexity

Solution Approach 1:

The patent divides the FPGA fabric into modular tiles that can be synthesized independently using bottom-up methodology, yet maintains global timing control through a hierarchical approach where each tile is a self-contained unit with standardized interfaces

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces an intermediary layer of abstraction that connects individual tiles to the global timing system, allowing local tile synthesis while maintaining fabric-level performance through coordinated timing constraints and interface standardization

Inventive Principle:
Principle #24Intermediary (Mediator)

2Manufacturing precision

If floorplanning and manual buffer insertion are used to address timing issues, then manufacturing precision is improved, but device complexity increases and ease of operation deteriorates

Engineering Contradiction:
Improveprecision of timing controlVSAvoidcomplexity of floorplanning and manual buffering
Core Design Contradiction:
Manufacturing precisionVSDevice complexity

Solution Approach 1:

The patent enables the synthesis tool to automatically handle timing constraints and buffer insertion through standardized interfaces, eliminating the need for manual floorplanning and buffer insertion while maintaining precise timing control through self-service mechanisms

Inventive Principle:
Principle #25Self-service

3Ease of manufacture

If soft eFPGA fabrics are implemented using RTL design abstraction, then ease of manufacture is improved, but device complexity increases due to difficulty in providing global timing view

Engineering Contradiction:
Improveease of RTL design and process portabilityVSAvoidcomplexity of providing global timing view
Core Design Contradiction:
Ease of manufactureVSDevice complexity

Solution Approach 1:

The patent introduces an intermediary timing constraint system that bridges RTL design abstraction with global timing requirements, allowing soft eFPGA fabrics to maintain both RTL portability and fabric-level timing control through standardized timing interfaces

Inventive Principle:
Principle #24Intermediary (Mediator)

Data Source

PatentUS12393757B2Top down physical design of soft embedded field programmable gate array (FPGA) fabrics
Publication Date: 2025.08.19 CARNEGIE MELLON UNIV
  • US12393757B2 patent drawing
  • US12393757B2 patent drawing
  • US12393757B2 patent drawing

AI summary

Disclosed herein is a method for performing top-down design for optimizing synthesizable FPGA fabrics which eliminates the need for floorplanning and manual buffering by providing a global timing view of the FPGA fabric to the electronic design automation (EDA) tools. This is accomplished by providing the EDA tools with a global timing view spanning multiple tiles. The method restructures the switch-box muliplexers in a unique way to enable the use of powerful static timing analysis (STA) exceptions supported by modern EDA tools to work around the problem of combinational loops and long false paths created by the interconnect mesh in the FPGA fabric.