1T1R Memristor Structure for Accurate Hamming Distance Computing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing unipolar memristors face issues of poor durability, slow operation speed, and complex peripheral circuit designs due to varying voltage amplitudes during Hamming distance calculations, limiting their flexibility and utilization in traditional computing architectures.
Innovation Solution
A memristor structure comprising a transistor and resistive random access memory with a current compliance layer to stabilize resistance states, allowing for high-speed, low-power Hamming distance calculations through bipolar switching and parallel operations, utilizing a 1T1R configuration with specific electrode and material compositions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If unipolar memristors are used for Hamming distance calculation, then the device structure is simple, but the durability is poor and operation speed is slow
Solution Approach 1:
The patent changes the switching mechanism parameter from unipolar to bipolar, enabling the memristor to switch resistance states using both positive and negative voltages. This parameter change improves durability by allowing more robust switching operations and extends the device lifetime while maintaining operational speed.
Solution Approach 2:
The patent employs a composite structure combining resistive random access memory with specific electrode materials and resistive layers. This composite material approach enhances device durability by distributing stress and improving interface stability, thereby increasing endurance without significantly increasing structural complexity.
2Ease of operation
If different voltage amplitudes are applied during input, then the Hamming distance calculation can be performed, but the peripheral circuit design becomes complex
Solution Approach 1:
The patent applies equipotentiality by using symmetric voltage amplitudes for both row and column inputs during Hamming distance calculation. This approach simplifies the peripheral circuit design by eliminating the need for complex voltage scaling circuits, while still enabling accurate calculation through balanced differential signaling.
Solution Approach 2:
The memristor array is designed to perform multiple functions including Hamming distance calculation, Hamming weight calculation, and general matrix operations using the same hardware structure and voltage amplitudes. This universal design eliminates the need for specialized peripheral circuits for different operations, reducing overall system complexity.
3Ease of operation
If only diagonal cells are used for calculation, then the calculation can be performed, but the array utilization rate decreases
Solution Approach 1:
The patent designs the memristor array so that all cells, not just diagonal ones, can actively participate in Hamming distance calculations. By applying appropriate voltage patterns to row and column lines, any cell in the array can be programmed to perform the calculation, achieving full array utilization and enabling reconfigurable computing operations.
Solution Approach 2:
The patent transitions from utilizing only one-dimensional diagonal cells to employing the full two-dimensional array structure for calculations. By activating both row and column dimensions simultaneously with coordinated voltage signals, the system achieves comprehensive array utilization while maintaining the calculation function.
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
The proposed memristor achieves ultra-low energy consumption, high-speed Hamming distance calculations, and improved array utilization by stabilizing resistance states, enabling efficient in-memory computing with reduced complexity and increased flexibility.
Implementation Method 1
the current compliance layer is configured to stabilize a fluctuation of a low resistance by reducing a surge current and optimizing a heat distribution
Data Source
AI summary
The present disclosure provides a memristor, including a transistor and a resistive random access memory, where a drain electrode of the transistor is connected to a bottom electrode of the resistive random access memory; and the resistive random access memory includes: the bottom electrode, a resistive random access material layer, a current compliance layer and a top electrode from bottom to top, where the current compliance layer is configured to stabilize a fluctuation of a low resistance by reducing a surge current and optimizing a heat distribution, so as to improve a calculation accuracy of a Hamming distance.