Fuzzy Logic Machine Learning for IC Place and Route Optimization

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

The challenge in integrated circuit design is optimizing the physical area of System-on-Chip (SoC) products while ensuring functional requirements are met, as existing methods require manual effort and lack consistency, and achieving the smallest possible physical area is desirable but difficult to guarantee.

Innovation Solution

Employing fuzzy logic machine learning algorithms to automatically determine the area and shape of standard cells for physical design, utilizing a partition with a netlist and proposed floorplan to minimize the required area through place and route (PnR) optimization.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If manual methods are used to optimize physical area, then design flexibility is maintained, but efficiency and consistency deteriorate

Engineering Contradiction:
Improveoptimization efficiencyVSAvoidmanual effort required
Core Design Contradiction:
ProductivityVSExtent of automation

Solution Approach 1:

The system enables self-service optimization by allowing the design tool to automatically evaluate and optimize floorplan layouts without requiring manual intervention. The automated place-and-route algorithm autonomously determines optimal cell placements and routing configurations, making the system self-optimizing based on predefined design rules and constraints.

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The patent replaces manual mechanical optimization processes with automated computational algorithms. Instead of designers manually adjusting layouts, the system uses computer-based place-and-route algorithms that automatically optimize the physical arrangement of cells and interconnects based on mathematical models and design rules.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

2Productivity

If automated EDA processes are used, then productivity is improved, but manufacturing precision and design accuracy deteriorate

Engineering Contradiction:
Improvedesign automation levelVSAvoidarea optimization accuracy
Core Design Contradiction:
ProductivityVSManufacturing precision

Solution Approach 1:

The system incorporates feedback mechanisms where the automated EDA tool continuously evaluates design metrics (such as area, timing, and signal integrity) and adjusts the floorplan layout accordingly. The place-and-route algorithm uses feedback from design rule violations and performance metrics to iteratively improve the optimization accuracy.

Inventive Principle:
Principle #23Feedback

Solution Approach 2:

The patent employs parameter changes by dynamically adjusting optimization parameters such as weightings for different design rules, cell placement constraints, and routing preferences. The system can modify these parameters based on the specific design requirements and technology node to maintain high precision while achieving automation.

Inventive Principle:
Principle #35Parameter changes

3Area of stationary object

If the physical area is minimized, then device size is reduced, but functional requirements may be compromised

Engineering Contradiction:
ImproveSoC physical areaVSAvoidfunctional requirement satisfaction
Core Design Contradiction:
Area of stationary objectVSReliability

Solution Approach 1:

The system applies local quality optimization by allowing different regions of the floorplan to have different optimization characteristics. Critical areas with timing constraints or signal integrity requirements can be prioritized over areas where area minimization is the primary goal, ensuring that functional requirements are met in critical regions while achieving overall area reduction.

Inventive Principle:
Principle #3Local quality

Solution Approach 2:

The patent implements dynamic optimization where the place-and-route algorithm can adapt its behavior based on the specific design requirements. The system dynamically adjusts the balance between area minimization and functional requirement satisfaction by weighing different design rules and constraints differently based on the project requirements, allowing flexible trade-off management.

Inventive Principle:
Principle #15Dynamics

Data Source

PatentUS11392748B2Integrated circuit design using fuzzy machine learning
Publication Date: 2022.07.19 TAIWAN SEMICONDUCTOR MANUFACTURING CO LTD
  • US11392748B2 patent drawing
  • US11392748B2 patent drawing
  • US11392748B2 patent drawing

AI summary

Systems and methods include receiving a functional integrated circuit design and generating a plurality of place and route (PnR) layouts based on the received functional integrated circuit design and one or more integrated circuit floorplans may be generated. One or more fuzzy logic rules may be applied to analyze attributes associated with each of the generated PnR layouts, and a PnR layout of the plurality of PnR layouts having an area utilization complying with the one or more fuzzy logic rules may be generated.