3D Stacked IC Blocks for High-Density Interconnects

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

High-end computer systems face challenges in achieving high component density, direct access, and efficient interconnectivity while minimizing processing delays and heat dissipation in multi-processor systems due to physical separation of processors by metallic or optical paths.

Innovation Solution

A method of arranging planar components, such as integrated circuits, into local blocks and forming a three-dimensional structure with regularly repeating units, allowing for short, direct pathways and flexibility in inter-component spacings, enabling scalability and modular additions within a compact volume.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Quantity of substance

If processors are physically separated and interconnected by metallic wires or optical light paths, then interconnection bandwidth can be maintained, but processing delays increase due to limited logical states in communication paths

Engineering Contradiction:
Improvecomponent densityVSAvoidprocessing delay
Core Design Contradiction:
Quantity of substanceVSLoss of time

Solution Approach 1:

The patent transitions from two-dimensional planar arrangements to three-dimensional stacking of integrated circuit substrates. Multiple substrates are positioned at different vertical levels (z-dimension) and interconnected through vertical vias and conductive paths, enabling high component density while maintaining short communication paths. The 3D architecture allows processors to be closely spaced in three dimensions rather than spread out in two dimensions, reducing signal transmission distance and processing delays.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

Solution Approach 2:

The patent implements a hierarchical nesting structure where multiple integrated circuit substrates are stacked and interconnected. Each substrate contains multiple processors and memory elements nested within the three-dimensional volume. The nested arrangement allows processors to access both local and remote components through vertically stacked interconnection layers, achieving high density without proportionally increasing communication path lengths.

Inventive Principle:
Principle #7Nested doll (Nesting)

2Quantity of substance

If component density is increased through miniaturization, then more processors can be packed into a given volume, but accessibility and interconnectivity of individual components become more difficult

Engineering Contradiction:
Improvecomponent densityVSAvoidcomponent accessibility
Core Design Contradiction:
Quantity of substanceVSEase of operation

Solution Approach 1:

The patent divides the computing system into multiple discrete integrated circuit substrates, each containing a manageable number of processors and memory elements. This segmentation allows individual substrates to be accessed, replaced, or modified independently while maintaining high overall system density. The modular substrate architecture enables easier maintenance and upgrading compared to a monolithic high-density arrangement.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

By stacking substrates in the third dimension, the patent achieves high component density vertically while maintaining horizontal accessibility. Individual substrates can be accessed from the sides or top/bottom surfaces, and the vertical stacking allows multiple components to coexist in a compact volume without compromising individual component accessibility through established semiconductor packaging and interconnection techniques.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

3Loss of time

If processors are closely interconnected to reduce processing delays, then communication efficiency improves, but heat dissipation pathways become more constrained

Engineering Contradiction:
Improveprocessing delayVSAvoidheat dissipation
Core Design Contradiction:
Loss of timeVSTemperature

Solution Approach 1:

The patent employs different interconnection approaches for different functional requirements. Critical communication paths use direct vertical vias and short horizontal traces for speed, while thermal management utilizes dedicated heat dissipation structures such as heat sinks, thermal vias, and conductive pathways routed to external cooling systems. This local differentiation allows simultaneous optimization of communication speed and thermal management in different regions of the 3D architecture.

Inventive Principle:
Principle #3Local quality

Solution Approach 2:

By segmenting the system into multiple substrates with intermediate spacing and dedicated thermal pathways, the patent creates natural thermal management zones. Each substrate can be equipped with localized heat dissipation structures, and the vertical stacking allows heat to be conducted through multiple pathways to external cooling systems, preventing heat accumulation while maintaining close processor interconnections for low-latency communication.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentUS8214786B2Scalable, component-accessible, and highly interconnected three-dimensional component arrangement within a system
Publication Date: 2012.07.03 VALTRUS INNOVATIONS LTD
  • US8214786B2 patent drawing
  • US8214786B2 patent drawing
  • US8214786B2 patent drawing

AI summary

Embodiments of the present invention include dense, but accessible and well-interconnected component arrangements within multi-component systems, such as high-end multi-processor computer systems, and methods for constructing such arrangements. In a described embodiment, integrated-circuit-containing processing components, referred to as a “flat components,” are arranged into local blocks of intercommunicating flat components. The local flat-component blocks are arranged into interconnected, primitive multi-local-block repeating units, and the primitive local-block repeating units are layered together in a three-dimensional, regularly repeating structure that can be assembled to approximately fill any specified three-dimensional volume. The arrangement provides for relatively short, direct pathways from the surface of the specified volume to any particular local block and flat component within the three-dimensional arrangement.