Block-Level RC Extraction via Full-Chip Domain Translation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current full-chip modeling approaches for integrated circuits are limited to the sign-off stage and cannot be utilized early in the design implementation flow due to the lack of full-chip information and sufficient time for long simulations, preventing accurate resistance and capacitance extraction, especially at the block level where each IP or macro block is implemented independently.

Innovation Solution

A method and system that translate full-chip design domain information to a block-level design domain, allowing model-based prediction results to be used for parasitic RC and timing extractions, enabling early design implementation flow by converting manufacturing variation models like CMP and etch models into block-level coordinates for localized analysis.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If full-chip modeling approaches are used for RC extraction, then manufacturing variation accuracy is improved, but design flow timing and computational efficiency deteriorate

Engineering Contradiction:
ImproveRC extraction accuracyVSAvoiddesign implementation time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent divides the full-chip design into independent macro blocks, allowing RC extraction to be performed at the block level rather than requiring complete full-chip information. This segmentation enables earlier execution of RC extraction during design implementation while maintaining accuracy by considering manufacturing variations specific to each block's context.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent performs RC extraction with manufacturing variation modeling at the macro block level before complete full-chip design finalization. By conducting preliminary RC extraction on individual blocks with estimated manufacturing contexts, designers can make informed decisions earlier in the design flow without waiting for complete full-chip modeling results.

Inventive Principle:
Principle #10Preliminary action

2Manufacturing precision

If full-chip information is available for modeling, then manufacturing variation prediction accuracy is improved, but design flexibility and early implementation capability deteriorate

Engineering Contradiction:
Improvemanufacturing variation prediction accuracyVSAvoidearly design implementation flexibility
Core Design Contradiction:
Manufacturing precisionVSAdaptability or versatility

Solution Approach 1:

The patent applies local quality by considering manufacturing variations specific to each macro block's local context rather than applying uniform full-chip models. Each block can be analyzed with manufacturing parameters relevant to its specific location and characteristics, enabling early design flexibility while maintaining manufacturing accuracy.

Inventive Principle:
Principle #3Local quality

3Measurement precision

If long simulations are performed for accurate RC extraction, then extraction precision is improved, but productivity and design iteration speed deteriorate

Engineering Contradiction:
ImproveRC extraction precisionVSAvoiddesign iteration speed
Core Design Contradiction:
Measurement precisionVSProductivity

Solution Approach 1:

By segmenting the design into macro blocks, the patent enables parallel processing of RC extraction for multiple blocks independently. This reduces overall computation time compared to sequential full-chip analysis, improving design iteration speed while maintaining extraction precision through block-level manufacturing variation modeling.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentUS8219944B2Method and system performing block-level RC extraction
Publication Date: 2012.07.10 CADENCE DESIGN SYST INC
  • US8219944B2 patent drawing
  • US8219944B2 patent drawing
  • US8219944B2 patent drawing

AI summary

A method, system, and computer program product are disclosed for performing RC extraction from the perspective of the block level. A translation mechanism is employed to convert from a full-chip design domain to a block-level design domain. This allows model-based prediction results to be used in the early design implementation flow when parasitic RC and timing extractions are performed, where the model-based prediction results relate to predictions of manufacturing variations such as thickness and topography.