Selective Memory Page Sampling Based on Program-Erase Execution Time
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current memory sub-systems face performance penalties and increased latency due to the need for comprehensive scanning of all pages during program/erase cycles to detect defects, which is time and bandwidth-consuming, and can lead to reduced overall performance.
Innovation Solution
Implementing selective sampling based on program/erase execution times, where data units with prolonged program or erase times are scanned, and only subsets of pages are checked for defects, allowing for the release of parity data and potential retirement of defective units, thereby reducing unnecessary scanning and resource allocation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If comprehensive scanning of all pages is performed during program/erase cycles to detect defects, then defect detection reliability is improved, but scanning overhead and latency increase
Solution Approach 1:
The patent applies partial action by scanning only a subset of pages rather than all pages during program/erase cycles. Specifically, it scans pages that are likely to have defects based on program/erase execution time thresholds, while skipping pages that are unlikely to have defects. This partial scanning approach maintains adequate defect detection reliability while significantly reducing scanning overhead and latency compared to comprehensive scanning of all pages.
Solution Approach 2:
The patent changes the scanning parameter from a fixed comprehensive scan to a dynamic selective scan based on program/erase execution time. By monitoring execution time as a parameter and comparing it against thresholds, the system adapts the scanning behavior - scanning pages with prolonged execution times and skipping pages with normal execution times. This parameter-based approach optimizes the balance between defect detection reliability and scanning efficiency.
2Reliability
If comprehensive scanning of all pages is performed during program/erase cycles, then defect detection coverage is improved, but bandwidth consumption increases
Solution Approach 1:
The patent reduces bandwidth consumption by performing partial scans of only a subset of pages rather than scanning all pages. By selecting only pages with prolonged program/erase execution times for scanning, the system minimizes the amount of data transferred and processed during defect detection, thereby reducing bandwidth consumption while maintaining adequate defect detection coverage for pages that are most likely to have defects.
Solution Approach 2:
The patent dynamically changes the scanning scope parameter based on program/erase execution time measurements. Pages with normal execution times are excluded from scanning, reducing bandwidth consumption, while pages with prolonged execution times are included to maintain defect detection coverage. This parameter-driven selective scanning optimizes the trade-off between detection coverage and bandwidth usage.
3Reliability
If comprehensive scanning of all pages is performed during program/erase cycles, then overall system reliability is improved, but overall performance decreases
Solution Approach 1:
The patent improves overall performance by performing partial scans of only a subset of pages rather than scanning all pages during program/erase cycles. This selective scanning approach reduces the time and resources spent on scanning operations, thereby improving throughput and overall system performance. At the same time, it maintains adequate system reliability by focusing scan resources on pages with prolonged execution times that are most likely to have defects.
Solution Approach 2:
The patent dynamically adjusts the scanning parameter based on program/erase execution time, changing from a static comprehensive scan to a dynamic selective scan. This parameter-based approach optimizes the balance between reliability and performance by scanning only when necessary (when execution time exceeds thresholds) and skipping scans when not necessary (when execution time is normal), thereby improving overall system performance while maintaining adequate reliability.
Data Source
AI summary
A method includes obtaining a first operation execution time corresponding to an operation performed on a page of a first data unit of a memory device, determining whether the first operation execution time satisfies a condition that is based on a second operation execution time, wherein the second operation execution time is indicative of lack of defect in at least a second data unit of the memory device, and responsive to determining that the first operation execution time satisfies the condition that is based on the second operation execution time, initiating a defect scan operation of at least a subset of pages of the first data unit.