Crossbar Array Conductance Adjustment for Computational Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Real crossbar circuits experience computational inaccuracy due to factors like increasing wire resistance, device nonlinearity, thermal effects, noise, and parasitics, which are difficult to compensate for, leading to deviations from ideal results in vector-matrix multiplication.
Innovation Solution
A system that includes a crossbar array with memory elements, a calculate engine to determine ideal conductance, and an adjust engine to adjust the conductance of memory elements to reduce the difference between ideal and actual conductance, thereby improving computational accuracy by compensating for nonlinearity and other sources of error.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a crossbar array is used for parallel signal processing, then computational efficiency is improved, but computational accuracy deteriorates due to wire resistance, device nonlinearity, thermal effects, noise, and parasitics
Solution Approach 1:
The patent applies preliminary action by performing conductance adjustment of memory elements before they are used for computational operations. The system calculates ideal conductance values and adjusts the actual conductance of memory elements to match these ideal values in advance, thereby compensating for nonlinearity and other accuracy-degrading factors before computations begin. This preliminary calibration enables the crossbar array to maintain high computational accuracy while preserving its parallel processing efficiency.
2Measurement precision
If conductance of memory elements is adjusted to improve computational accuracy, then measurement precision is improved, but device complexity increases due to additional calculate and adjust engines
Solution Approach 1:
The patent implements self-service by enabling the crossbar array system to automatically calculate ideal conductance values and adjust the conductance of memory elements without requiring external intervention. The calculate engine determines the ideal conductance based on the desired computational accuracy, and the adjust engine automatically modifies the memory element conductance to match these ideal values. This self-calibrating capability improves computational accuracy while minimizing the need for complex external control systems.
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 adjustment of conductance in memory elements within the crossbar array significantly improves computational accuracy, bringing the output closer to ideal results by mitigating the effects of nonlinearity and other sources of error, enhancing the reliability of vector-matrix multiplication operations.
Implementation Method 1
an ideal crossbar structure may be used to accurately perform, via Kirchhoff's Current Law, vector-matrix multiplication between input vectors and data values (weights) stored by memory elements in a matrix of the crossbar array
Implementation Method 2
adjust conductance of at least one memory element to improve computational accuracy by reduction of a difference between the ideal conductance and the determined conductance of the at least one memory element
Data Source
AI summary
Example implementations of the present disclosure relate to improved computational accuracy in a crossbar array. An example system may include a crossbar array, having a plurality of memory elements at junctions, usable in performance of computations. The example system may further include a calculate engine to calculate ideal conductance of memory elements at a plurality of junctions of the crossbar array and a determine engine to determine conductance of the memory elements at the plurality of junctions of the crossbar array. An adjust engine of the example system may be used to adjust conductance of at least one memory element to improve computational accuracy by reduction of a difference between the ideal conductance and the determined conductance of the at least one memory element.