Volatile Memory Metadata Allocation via Wordline Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current volatile memory devices face challenges in efficiently allocating metadata and normal data, leading to inefficiencies in chip size overhead and operational modes, particularly in achieving optimal ratios of metadata to normal data access without additional resources.
Innovation Solution
The proposed solution involves allocating every r-th wordline in a memory device for metadata, where r is a positive integer, allowing for simultaneous output of normal data and metadata without additional Global Input/Output (GIO) lines, by intersecting the allocation relationship between subbanks, thus optimizing the normal/meta data ratio without increasing chip size.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If every r-th wordline is allocated for metadata in a column direction, then the ratio of metadata to normal data can be controlled, but the chip size overhead increases due to subbank division structure
Solution Approach 1:
The memory bank is divided into multiple subbanks, with specific subbanks allocated for metadata storage. By segmenting the bank structure and assigning every r-th wordline to metadata in specific subbanks, the patent achieves controlled metadata-to-normal data ratios while managing chip size overhead through structured organization.
Solution Approach 2:
Different regions of the memory bank are assigned different functions: some subbanks are dedicated to metadata storage while others handle normal data. This local differentiation allows optimal allocation of resources, enabling the system to achieve desired metadata ratios without uniformly increasing chip size across the entire bank structure.
2Productivity
If subbanks are divided to allocate metadata, then metadata access can be optimized, but the operational complexity increases compared to single bank structure
Solution Approach 1:
The bank is segmented into multiple subbanks with specialized functions. Metadata is stored in specific subbanks (e.g., even-numbered subbanks), while normal data is stored in other subbanks. This segmentation enables optimized metadata access paths while maintaining manageable operational complexity through clear separation of duties.
Solution Approach 2:
The subbank structure is designed to handle both metadata and normal data operations through a unified control mechanism. The same wordline and column select line infrastructure serves both metadata and normal data, allowing the system to maintain operational simplicity while achieving specialized metadata access optimization.
3Productivity
If additional GIO lines are added for simultaneous output, then normal data and metadata can be output simultaneously, but the chip size increases
Solution Approach 1:
The GIO lines are designed to serve dual purposes: they can output both normal data and metadata simultaneously by leveraging the subbank structure. Instead of adding dedicated GIO lines for metadata, the system reuses the existing GIO infrastructure through intelligent subbank allocation, achieving simultaneous output without increasing chip size.
Solution Approach 2:
The patent resolves the simultaneous output requirement by adding a dimensional aspect to the data organization - using the subbank dimension rather than requiring additional GIO line dimensions. By organizing data across multiple subbanks and selecting appropriate subbanks for output, the system achieves parallel output capability without expanding the GIO interface physically.
Data Source
AI summary
A memory device includes a bank connected to a plurality of wordlines and a plurality of column select lines (CSLs). Among the plurality of wordlines, every r-th wordline in a column direction may be allocated for metadata where r is a positive integer.