Multilayer Routing Order Prediction via Deep Learning

Solution Overview

Problem

Current routing methods in multilayer semiconductor structures are inefficient due to the lack of consideration for the routing order between node groups, leading to suboptimal path generation and increased computational costs.

Innovation Solution

A deep learning-based method that predicts an optimal routing order for node groups in a multilayer structure by generating routing order examples using heuristic or meta-heuristic algorithms, training a deep learning model with a predefined evaluation function, and extracting features from compressed grid maps to reduce computational costs and enhance prediction accuracy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Manufacturing precision

If routing order is not considered in prior routing methods, then routing can be performed with simpler processes, but routing outcomes vary greatly and optimality cannot be achieved

Engineering Contradiction:

Improverouting outcome optimalityVSAvoidrouting process complexity

Core Design Contradiction:

Manufacturing precisionVSDevice complexity

Solution Approach 1:

The patent applies preliminary action by determining the routing order of multiple node groups before actually performing the routing. The system calculates and establishes an optimal sequence for routing different node groups, then executes routing operations following this predetermined sequence. This preliminary determination of routing order ensures optimal routing outcomes while maintaining a structured process that doesn't excessively increase complexity.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent implements dynamics by making the routing process adaptive and flexible through multiple evaluation functions. Different evaluation functions (such as total wire length, number of vias, congestion metrics) can be applied to assess routing quality, and the system can dynamically adjust routing strategies based on these evaluations. This dynamic approach allows the system to optimize routing outcomes while managing process complexity through flexible, performance-based adjustments.

Inventive Principle:

Principle #15Dynamics

2Manufacturing precision

If multiple routing algorithms are executed to find optimal paths, then routing quality improves, but computational cost and time increase significantly

Engineering Contradiction:

Improverouting path qualityVSAvoidcomputational time

Core Design Contradiction:

Manufacturing precisionVSLoss of time

Solution Approach 1:

The patent applies preliminary action by determining the routing order of multiple node groups before actually performing the routing. The system calculates and establishes an optimal sequence for routing different node groups, then executes routing operations following this predetermined sequence. This preliminary determination of routing order ensures optimal routing outcomes while maintaining a structured process that doesn't excessively increase complexity.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent implements dynamics by making the routing process adaptive and flexible through multiple evaluation functions. Different evaluation functions (such as total wire length, number of vias, congestion metrics) can be applied to assess routing quality, and the system can dynamically adjust routing strategies based on these evaluations. This dynamic approach allows the system to optimize routing outcomes while managing process complexity through flexible, performance-based adjustments.

Inventive Principle:

Principle #15Dynamics

3Productivity

If routing is performed without considering node group sequencing, then the routing process is faster and simpler, but the resulting path sets are suboptimal

Engineering Contradiction:

Improverouting process efficiencyVSAvoidpath set optimality

Core Design Contradiction:

ProductivityVSManufacturing precision

Solution Approach 1:

The patent applies preliminary action by determining the routing order of multiple node groups before actually performing the routing. The system calculates and establishes an optimal sequence for routing different node groups, then executes routing operations following this predetermined sequence. This preliminary determination of routing order ensures optimal routing outcomes while maintaining a structured process that doesn't excessively increase complexity.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent implements feedback by using evaluation functions to assess the quality of routing results and using this information to improve subsequent routing decisions. The system evaluates routing outcomes based on metrics such as wire length, via count, and congestion, then uses this feedback to adjust routing strategies and determine optimal routing orders. This feedback mechanism ensures high-quality path sets while maintaining efficiency through data-driven decision making.

Inventive Principle:

Principle #23Feedback

Data Source

PatentUS20240386182A1Routing method and system in multilayer structure

Publication Date: 2024.11.21 SAMSUNG SDS CO LTD

AI summary

A routing method in a multilayer structure is provided. The routing method may include: acquiring a routing problem, wherein the routing problem is a problem of generating a path set that includes respective paths for multiple node groups arranged in a multilayer structure, generating a routing order example for the multiple node groups, generating a path set for the multiple node groups by executing a routing algorithm based on the routing order example, establishing a training set by obtaining a cost of the generated path set based on a predefined evaluation function, and training a deep learning model to predict a routing order using the training set.