Adaptive-Priority Crossbar Circuitry With Integrated Cell Arbitration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing crossbar circuit designs face scalability issues due to complex routing of control signals, significant power consumption, and inefficiencies in implementing adaptive priority schemes, particularly with the least recently granted (LRG) priority scheme, which results in hardware complexity and performance overhead.
Innovation Solution
The crossbar circuitry integrates arbitration and transmission functions within each crossbar cell, using configuration storage and priority storage circuits to apply an adaptive priority scheme, where bit lines are reused for conflict detection and resolution, allowing for efficient implementation of adaptive priority schemes like LRG, with reduced hardware complexity and power consumption.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If traditional MUX-based crossbar circuits are used to interconnect multiple sources and destinations, then routing capability is achieved, but the circuit size and power consumption increase significantly
Solution Approach 1:
The patent combines the arbitration logic and crossbar switching functionality into a single integrated structure. The arbitration circuitry is distributed across the crossbar fabric itself, with each crossbar cell containing local arbitration logic that works cooperatively with neighboring cells. This merging eliminates the need for separate arbitration units and reduces overall circuit area while maintaining full routing capability.
Solution Approach 2:
The patent divides the arbitration function into multiple distributed arbitration units, one located at each crossbar cell. Each arbitration unit handles only the arbitration for its local crossbar cell, rather than having a single centralized arbiter handle all routing decisions. This segmentation distributes the arbitration workload and reduces the area required for arbitration logic.
2Adaptability or versatility
If hierarchical MUX arrangements are used for crossbar circuits, then routing flexibility is provided, but the number of control lines increases considerably
Solution Approach 1:
The patent makes the data paths multi-functional by using them for both data transmission and arbitration control signal routing. The same physical wires that carry data between crossbar cells also carry arbitration control signals. This eliminates the need for dedicated control lines and reduces wiring complexity while maintaining routing flexibility.
Solution Approach 2:
The patent merges the data path and control path functions into a single unified structure. Control signals for arbitration are transmitted through the existing data path infrastructure, combining what would traditionally be separate signaling channels into one integrated system, thereby reducing the total number of control lines required.
3Ease of operation
If conventional arbitration schemes are implemented separately from crossbar circuitry, then priority control is achieved, but hardware complexity and performance overhead increase
Solution Approach 1:
The patent integrates arbitration logic directly within the crossbar fabric, with each crossbar cell containing local arbitration units that work cooperatively with adjacent cells. This distributed arbitration approach merges the arbitration function with the crossbar switching fabric, eliminating separate arbitration hardware and reducing overall system complexity while maintaining priority control capabilities.
Solution Approach 2:
The patent implements self-service arbitration where each crossbar cell autonomously performs arbitration decisions for its local cell based on priority information, rather than requiring external arbitration control. The distributed arbitration units in each cell independently make routing decisions by cooperatively evaluating priority signals, reducing the need for complex external arbitration logic and improving operational efficiency.
Data Source
AI summary
Crossbar circuitry has an array of data input and output paths where the data output paths are transverse to the data input paths. At each intersection between a data input path and a data output path, a crossbar cell is provided which includes a configuration storage circuit programmable to store a routing value, a transmission circuit, and an arbitration circuit. In a transmission mode of operation, the transmission circuit is responsive to the routing value being a first value, indicating that the data input path should be coupled to the data output path, to detect the data input along the data input path, and to output an indication of that data on the data output path at the associated intersection. In an arbitration mode of operation, the arbitration circuitry is operable to selectively modify the voltage on said plurality of bit lines in order to apply an adaptive priority scheme.