3D Die Partitioning for Logic, SRAM, and Analog Node Optimization

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Traditional semiconductor process nodes compromise on performance and power efficiency to accommodate both high-speed logic and other devices like SRAM and analog devices, leading to suboptimal performance across all components.

Innovation Solution

Implementing a die pair device partitioning approach where two separate process nodes are connected via 3D hybrid bonding, with one node optimized for logic transistors and the other for SRAM and analog devices, allowing for independent optimization of each without compromising the other.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If a single planar silicon process node is used to accommodate both high-speed logic and other devices (SRAM, analog), then all device types can be integrated on one chip, but the performance and power efficiency of logic devices are compromised

Engineering Contradiction:
Improvedevice integration capabilityVSAvoidlogic device performance
Core Design Contradiction:
Adaptability or versatilityVSReliability

Solution Approach 1:

The patent divides the traditional single process node into two separate process nodes: a first process node optimized for high-speed logic devices and a second process node optimized for other devices (SRAM, analog). These separate nodes are then integrated through 3D hybrid bonding to form a complete system, resolving the contradiction by allowing each node to be optimized for its specific function without compromise

Inventive Principle:
Principle #1Segmentation

2Reliability

If a single process node is optimized for logic devices, then logic performance improves, but the capability to support SRAM and analog devices is reduced

Engineering Contradiction:
Improvelogic device performanceVSAvoiddevice type capability
Core Design Contradiction:
ReliabilityVSAdaptability or versatility

Solution Approach 1:

The patent segments the device types into two groups handled by separate process nodes: logic-optimized devices on the first node and other devices (SRAM, analog, IO) on the second node. This segmentation allows each node to be specialized without sacrificing overall system capability

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent combines the two separate process nodes through 3D hybrid bonding technology, merging their capabilities into a single integrated system. This merging restores the full device type capability while maintaining the performance benefits of specialization

Inventive Principle:
Principle #5Merging (Combining)

3Ease of manufacture

If traditional planar process nodes are used, then manufacturing is simpler, but performance and power efficiency improvements are limited

Engineering Contradiction:
Improveprocess simplicityVSAvoidperformance efficiency
Core Design Contradiction:
Ease of manufactureVSReliability

Solution Approach 1:

The patent transitions from traditional 2D planar integration to 3D vertical integration using hybrid bonding. This dimensional change allows multiple specialized process nodes to be stacked and interconnected, achieving superior performance while managing manufacturing complexity through established 3D packaging techniques

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

Data Source

PatentUS20240324247A1Die pair device partitioning
Publication Date: 2024.09.26 ADVANCED MICRO DEVICES INC
  • US20240324247A1 patent drawing
  • US20240324247A1 patent drawing
  • US20240324247A1 patent drawing

AI summary

A method for die pair partitioning can include providing a circuit die that has a metal stack and that includes a majority of logic transistors of an integrated circuit. The method can also include providing one or more additional circuit die that have one or more additional metal stacks of which at least one is connected to the metal stack of the circuit die and a majority of static random access memory and analog devices of the integrated circuit. The method can further include connecting at least one of the one or more additional metal stacks to the metal stack of the circuit die. Various other methods, systems, and computer-readable media are also disclosed.