Memory Bank Resident Processor for Bandwidth Bottleneck Relief
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In memory devices, controlling operations externally leads to inefficiencies due to command transfer times and bandwidth bottlenecks, particularly in IoT applications where multiple memory devices ingest data that needs to be transferred to a host, resulting in reduced computing system performance.
Innovation Solution
Implementing a system processor resident on the memory device to control and perform memory operations locally, reducing the need for external command transfers and enabling processing-in-memory (PIM) operations without external data bus usage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If memory operations are controlled externally, then device complexity is reduced, but command transfer time increases and bandwidth bottlenecks occur
Solution Approach 1:
The patent combines the memory device and processor into a single integrated unit, merging previously separate functions into one device. This integration eliminates the need for external command transfers between separate memory and processor components, directly resolving the contradiction by reducing command transfer time while maintaining controlled operation within the unified device structure
2Device complexity
If data is transferred externally between memory and processor, then device complexity is reduced, but processing performance decreases due to bandwidth bottlenecks
Solution Approach 1:
By merging the processor directly into the memory device, the patent eliminates external data transfer pathways and their associated bandwidth limitations. The integrated architecture allows data to be processed in-place within the memory device, dramatically improving processing performance while simplifying the overall data transfer structure by removing external bus dependencies
3Device complexity
If multiple memory devices transfer data to host externally, then device complexity is reduced, but system performance decreases due to transfer bottlenecks
Solution Approach 1:
The patent merges processing capabilities directly into each memory device, enabling local data processing before host transfer. This integration allows multiple memory devices to perform computations independently without creating transfer bottlenecks, as data can be pre-processed and condensed locally, reducing the volume and frequency of external transfers while maintaining simple transfer architecture
Data Source
AI summary
An example apparatus includes a memory device comprising a plurality of banks of memory cells. A particular bank of memory cells among the plurality of banks includes a system processor resident on a particular bank of the plurality of banks.