Silicon Interconnect Structure for Stacked MTDRAM Random Access
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current DRAM systems face challenges in achieving high random access bandwidth, reduced access latency, low operating/standby power, improved random access capability, increased memory capacity, and higher memory density, while also requiring an efficient refresh scheme to meet the demands of complex applications like machine learning.
Innovation Solution
An arrayed processor system with stacked multi-threaded DRAM processor systems, interconnected by a silicon substrate with patterned metal interconnect layers, includes communication control chips, power management chips, and high-speed communication links, facilitating enhanced connectivity and data transmission between processor blocks and flash memory systems.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional HBM architectures increase data bandwidth for large data block accesses, then data bandwidth is improved, but power consumption increases significantly and random access capability deteriorates
Solution Approach 1:
The memory system is divided into multiple independently accessible DRAM unit cells (first, second, third, and fourth unit cells) organized in a stacked configuration. Each unit cell can be accessed independently through separate bit line pairs and sense amplifiers, enabling parallel random access operations. This segmentation allows the system to achieve high bandwidth for large data blocks by simultaneously accessing multiple unit cells while maintaining low power consumption by activating only the specific unit cells needed for each operation, rather than requiring high-power analog circuits for all accesses.
2Speed
If conventional DRAM systems use wide-interface architecture for high-speed operation, then data transmission speed is improved, but access latency increases and random access capability deteriorates
Solution Approach 1:
The patent transitions from a planar wide-interface architecture to a three-dimensional stacked architecture with multiple DRAM unit cells arranged vertically. Each unit cell has its own dedicated bit line pairs and sense amplifiers, enabling simultaneous independent access to multiple memory locations. This vertical stacking in the third dimension allows high-speed data transmission through parallel access paths while reducing access latency by eliminating the need for complex multiplexing and demultiplexing operations required in wide-interface architectures.
3Quantity of substance
If conventional DRAM systems increase memory capacity through larger arrays, then memory capacity is improved, but memory density and refresh efficiency deteriorate
Solution Approach 1:
The memory system is segmented into multiple independent DRAM unit cells, each with its own bit line pairs and sense amplifiers. This segmentation allows the total memory capacity to be distributed across multiple smaller, independently manageable units. Each unit cell can be refreshed independently, enabling efficient refresh operations that do not require activating the entire memory array. This approach maintains high memory capacity while reducing refresh complexity and improving overall refresh efficiency compared to conventional large-array architectures.
Data Source
AI summary
An arrayed processor system having an array of stacked MTDRAM processor systems. Each stacked MTDRAM processor system includes a controller chip having a plurality of processor blocks arranged in an array, and a plurality of DRAM chips. Each DRAM chip includes a plurality of independent DRAM unit cells arranged in an array, wherein each of the processor blocks of the controller chip is coupled to a corresponding DRAM unit cell in each of the DRAM chips. The arrayed processor system further includes communication control chips coupled to the stacked MTDRAM processor systems, power management chips coupled to the communication control chips and the stacked MTDRAM processor systems, and high-speed communication links coupled to the communication control chips. The various elements of the arrayed processor system are mounted on, and are interconnected by, an interconnect structure that includes a silicon substrate with a plurality of patterned metal interconnect layers formed thereon.


