Wafer-on-Wafer Memory-Logic Bus for High-Bandwidth 3D Integration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current manufacturing processes for 3D integrated circuits, such as wafer-on-wafer bonding, face challenges like expensive and time-consuming alignment and bonding operations, which hinder the development of high-bandwidth memory devices that are energy-efficient and cost-effective.
Innovation Solution
The implementation of a wafer-on-wafer bonding process that forms a wide bus between memory and logic dies, allowing for high bandwidth and flexible operation with a standardized input/output interface, enabling efficient power and thermal management in constrained environments like drones and data centers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If wafer-on-wafer bonding is used to form 3D integrated circuits, then bandwidth between memory and logic is improved, but manufacturing cost and time increase
Solution Approach 1:
The patent applies preliminary action by performing alignment marks formation and wafer bonding operations before dicing the wafers into individual devices. This allows the bonding interface to be established at the wafer level with proper alignment, and then the bonded wafers are cut into individual memory and logic devices. This approach resolves the contradiction by enabling high-bandwidth bonded connections to be created efficiently at the wafer level rather than requiring expensive and time-consuming individual device alignment and bonding operations.
2Device complexity
If separate memory and accelerator devices are used, then device complexity is reduced, but power consumption increases
Solution Approach 1:
The patent applies merging by integrating memory devices and logic devices into a single bonded 3D integrated circuit package. The memory wafer and logic wafer are bonded together to form a unified system where data can be transferred directly between memory and logic through the bonded interface. This resolves the contradiction by combining separate devices into an integrated system that reduces power consumption through direct high-bandwidth communication, while the overall device complexity remains manageable due to the modular wafer-level integration approach.
Data Source
AI summary
A memory device includes an array of memory cells configured on a die or chip and coupled to sense lines and access lines of the die or chip and a respective sense amplifier configured on the die or chip coupled to each of the sense lines. Each of a plurality of subsets of the sense lines is coupled to a respective local input/output (I/O) line on the die or chip for communication of data on the die or chip and a respective transceiver associated with the respective local I/O line, the respective transceiver configured to enable communication of the data to one or more device off the die or chip.


