Flash Memory Cell Programming With Adaptive Current Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The performance of flash memory in performing matrix-vector multiplication operations, such as those required for machine learning and neural network authentication, is limited by the speed of programming processes, which can be improved through precise control of programming time and voltage to enhance cell current consistency.
Innovation Solution
A method and apparatus that adjust programming time and voltage in a stepwise manner to control the cell current of flash memory cells, using multiple factors to achieve precise programming targets, including primary and secondary targets based on network parameters.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If programming time is increased to improve cell current consistency, then manufacturing precision improves, but productivity deteriorates
Solution Approach 1:
The programming process is divided into multiple stages with different programming voltages and time parameters. The controller performs programming operations by applying different voltage levels (e.g., first programming voltage, second programming voltage) with corresponding time parameters, allowing precise control of cell current at different stages rather than using a single fixed programming time
Solution Approach 2:
The programming time and voltage are dynamically adjusted based on real-time cell current feedback. The controller monitors cell current during programming and modifies programming parameters (time and voltage) adaptively to achieve target cell current values, transforming the static programming process into a dynamic control process
2Manufacturing precision
If programming voltage is increased to improve cell current consistency, then manufacturing precision improves, but use of energy deteriorates
Solution Approach 1:
The programming voltage is segmented into multiple levels (first programming voltage, second programming voltage) applied at different stages. This allows the system to use lower voltages for initial programming and reserve higher voltages only when necessary to achieve precise cell current control, reducing overall energy consumption compared to continuously applying high voltage
Solution Approach 2:
The programming voltage parameter is changed adaptively based on the programming stage and cell current feedback. The controller adjusts voltage parameters dynamically to achieve the desired cell current while minimizing energy consumption, rather than using a fixed high voltage throughout the programming process
3Manufacturing precision
If multiple programming stages are implemented to improve cell current consistency, then manufacturing precision improves, but device complexity deteriorates
Solution Approach 1:
The controller incorporates feedback mechanisms that monitor cell current during programming operations. Based on the feedback signal from cell current measurement, the controller automatically adjusts programming parameters (voltage and time) to achieve target values, simplifying the control logic through automated feedback-based adjustment rather than complex manual control sequences
Solution Approach 2:
The programming system performs self-adjustment by automatically modifying programming parameters based on real-time cell current feedback. The controller autonomously determines optimal programming conditions without external intervention, reducing the complexity of external control systems while maintaining high precision
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach enhances the accuracy and speed of flash memory operations, particularly in neural network processing, by narrowing the dispersal of threshold voltage and improving cell current consistency, thereby increasing the efficiency of in-memory computing.
Implementation Method 1
Flash memory may store data by adjusting the quantity of electrons present in a charge trapping layer of a memory cell
Implementation Method 2
The memory cell may be a single-level cell (SLC), a multi-level cell (MLC), or a triple-level cell (TLC), which correspond with amounts of storable data
Implementation Method 3
performing first programming on a target memory cell of a cell array while adjusting a first programming time and a programming voltage
Data Source
AI summary
A method and apparatus with flash memory control are provided. The method includes performing first programming on a target memory cell of a cell array while adjusting a first programming time and a programming voltage, when a cell current of the target memory cell is determined to satisfy a primary target in association with the first programming, performing second programming on the target memory cell while adjusting a second programming time, and when the cell current of the target memory cell is determined to satisfy a secondary target in association with the second programming, terminating programming on the target memory cell.