Flash Memory Data Merge via Pre-identified Valid Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing data merge processes in flash memory technology lead to increased access numbers to the memory storage device, resulting in decreased performance due to the need to access large amounts of management information during the data merge operation.
Innovation Solution
The method involves selecting source nodes for both the current and next data merge operations in advance, allowing for the identification of valid data using existing mapping information, thereby reducing the need to read the logical-to-physical mapping table and minimizing device access during the data merge process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If data merge operations access management information (mapping information) to identify valid data, then data merge accuracy is improved, but the number of memory accesses increases and overall performance deteriorates
Solution Approach 1:
The patent applies preliminary action by pre-selecting source nodes for both current and next data merge operations, and pre-identifying valid data in advance. This allows the system to have valid data information ready before the actual data merge operation begins, reducing the need for repeated memory accesses during the operation and thereby improving overall performance while maintaining data merge accuracy
2Measurement precision
If the system reads mapping information from the memory storage device during data merge operations, then valid data identification is accurate, but the number of device accesses increases
Solution Approach 1:
The system performs preliminary identification of valid data by selecting source nodes in advance and determining which data is valid before the actual data merge operation begins. This preliminary action eliminates the need to read mapping information from the memory storage device during the operation, thereby reducing device access time while maintaining accurate valid data identification
Solution Approach 2:
The patent extracts the mapping information access requirement from the critical path of the data merge operation. By pre-identifying valid data and storing this information in the memory control circuit unit, the system removes the need to access the memory storage device for mapping information during the data merge operation, thus reducing device access time while preserving identification accuracy
3Reliability
If multiple source nodes are selected for data merge operations, then data merge completeness is improved, but the complexity of managing source nodes increases
Solution Approach 1:
The patent applies self-service by enabling the memory control circuit unit to autonomously manage multiple source nodes. The circuit automatically selects source nodes for current and next data merge operations, tracks their states, and determines valid data without requiring complex external management, thereby maintaining data merge completeness while reducing management complexity
Data Source
AI summary
A data merge method for a rewritable non-volatile memory module including multiple physical units is provided. The method includes: starting a first data merge operation, and selecting at least one first physical unit for executing the first data merge operation and at least one second physical unit for executing a second data merge operation from the physical units; reading first mapping information from the rewritable non-volatile memory module, and copying first valid data collected from the at least one first physical unit to at least one third physical unit in the physical units; identifying second valid data in the at least one second physical unit according to the first mapping information in the first data merge operation; and starting the second data merge operation, and copying the second valid data collected from the at least one second physical unit to at least one fourth physical unit in the physical units.