Neural Network Flash Memory Wear Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
It is difficult to consistently predict the number of physical blocks erased from flash memory during a write operation, which hampers the ability to effectively manage and extend the lifespan of flash memory devices.
Innovation Solution
A computing device equipped with a neural network inference engine uses a predictive model generated by a neural network training engine to infer the predicted number of physical blocks erased from flash memory based on inputs such as the total number of previously erased blocks and the amount of data to be written, along with operational conditions like temperature.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional EEPROM erases data on a bit-by-bit level, then data can be written efficiently, but the flash memory wears out quickly due to frequent erase operations
Solution Approach 1:
The patent changes the erase operation parameter from bit-by-bit (conventional EEPROM) to block-by-block (flash memory), fundamentally altering how data is erased to match the physical capabilities of flash memory and reduce wear
Solution Approach 2:
The system performs preliminary actions by predicting the number of physical blocks that will be erased before actual write operations occur, enabling proactive wear management and lifespan extension
2Duration of action of stationary object
If flash memory erases data on a block-by-block level, then memory wear is reduced, but predicting the number of physical blocks erased becomes difficult
Solution Approach 1:
The patent replaces conventional mechanical counting methods with a neural network-based predictive model that processes multiple input parameters (previous erase count, data amount, temperature) to accurately predict the number of physical blocks erased
Solution Approach 2:
The system uses feedback from actual erase operations to continuously improve the neural network model's prediction accuracy, creating a closed-loop system that learns from real-world behavior
3Measurement precision
If a neural network model is used to predict physical blocks erased, then prediction accuracy improves, but device complexity increases
Solution Approach 1:
The patent creates a simplified copy of the complex neural network model that can be embedded in the computing device, allowing the device to make predictions without requiring the full complexity of the training engine
Solution Approach 2:
The system extracts only the essential predictive capabilities from the complex neural network, separating the prediction function from the training process and enabling deployment in resource-constrained environments
Data Source
AI summary
Computing device and method for inferring a predicted number of physical blocks erased from a flash memory. The computing device stores a predictive model generated by a neural network training engine. A processing unit of the computing device executes a neural network inference engine, using the predictive model for inferring the predicted number of physical blocks erased from the flash memory based on inputs. The inputs comprise a total number of physical blocks previously erased from the flash memory, an amount of data to be written on the flash memory, and optionally an operating temperature of the flash memory. In a particular aspect, the flash memory is comprised in the computing device, and an action may be taken for preserving a lifespan of the flash memory based at least on the predicted number of physical blocks erased from the flash memory.