Split Address Mapping in Memory Systems for Lower DRAM Cost
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The increasing capacity of memory systems leads to a rise in the amount of mapping information, which can increase manufacturing costs when stored in non-volatile memory, and existing methods do not effectively manage this data efficiently.
Innovation Solution
A memory system is designed with a non-volatile memory device storing first mapping information and a volatile memory device storing second mapping information, where the memory controller determines a target logical address and obtains a target physical address using both types of mapping information, reducing the amount of data stored in volatile memory.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Duration of action of stationary object
If mapping information is stored in non-volatile memory, then data retention is improved, but manufacturing cost increases due to larger capacity requirements
Solution Approach 1:
The patent divides mapping information into two segments: frequently accessed mapping information stored in volatile memory and less frequently accessed mapping information stored in non-volatile memory. This segmentation allows the system to maintain data retention for critical mappings while reducing the capacity requirements and manufacturing cost of non-volatile memory by only storing a portion of the total mapping information.
2Speed
If mapping information is stored in volatile memory, then data access speed is improved, but manufacturing cost increases due to larger capacity requirements
Solution Approach 1:
The patent segments mapping information storage between volatile and non-volatile memory types. Volatile memory stores only the most frequently accessed mapping information, reducing its capacity requirements and manufacturing cost, while maintaining fast access speeds for critical data. Less frequently accessed mapping information is stored in non-volatile memory.
Solution Approach 2:
The patent extracts less frequently accessed mapping information from volatile memory and stores it in non-volatile memory. This extraction reduces the capacity burden on volatile memory, lowering manufacturing costs while maintaining fast access performance for the most critical mapping information that remains in volatile memory.
3Quantity of substance
If all mapping information is stored in non-volatile memory, then volatile memory capacity is reduced, but data access efficiency deteriorates
Solution Approach 1:
The patent segments mapping information based on access frequency, storing frequently accessed mappings in volatile memory for fast access and less frequently accessed mappings in non-volatile memory. This segmentation reduces volatile memory capacity requirements while maintaining data access efficiency for critical operations.
Solution Approach 2:
The patent performs preliminary actions by pre-loading frequently accessed mapping information into volatile memory before it is needed for data access operations. This ensures that critical mapping information is readily available in fast memory, maintaining access efficiency while reducing the overall capacity burden on volatile memory.
Data Source
AI summary
In some embodiments, the memory system for communicating with a host includes a non-volatile memory device storing first mapping information, a volatile memory device storing second mapping information, and a memory controller. The first mapping information indicates a relationship between a logical address and a portion of a first physical address. The first physical address indicates a location where user data is stored. The second mapping information indicates a second relationship between the logical address and a second physical address that corresponds to a remaining portion of the first physical address. The memory controller is configured to obtain a target logical address that has been received from the host, and determine, based on the second mapping information, a target second physical address mapped to the target logical address. The non-volatile memory device is configured to obtain a target first physical address by using the first mapping information.


