Mixed-Signal Computer Architecture for Accuracy and Power Balance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current digital signal processing computers are not meeting the demands of emerging opportunities in areas like imaging, IoT, and cloud computing, as they face limitations in raw performance, power consumption, and cost, while analog computing elements offer advantages in raw performance and low power dissipation but struggle with accuracy and programmability.
Innovation Solution
A Mixed Signal Computer architecture that seamlessly integrates analog and digital computing elements, allowing for flexible data conversion and processing using a combination of digital arithmetic logic units (ALUs) and analog ALUs, with exchange registers for converting between digital and analog data, and a control unit managing both domains.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If digital signal processing computers are used, then accuracy and programmability are improved, but raw performance and power consumption are worsened
Solution Approach 1:
The computer system is segmented into distinct digital and analog processing units. Digital ALUs handle operations requiring high accuracy and programmability, while analog ALUs handle operations requiring high raw performance. This segmentation allows each domain to operate in its optimal performance regime without compromising the other.
Solution Approach 2:
The patent merges digital and analog computing elements into a single hybrid computer architecture. Digital components provide accuracy and programmability, while analog components provide high raw performance and low power consumption. The exchange registers and communication paths enable seamless data flow between the two domains, allowing the system to leverage the strengths of both approaches simultaneously.
2Adaptability or versatility
If digital signal processing computers are used, then programmability is improved, but power consumption is worsened
Solution Approach 1:
The system segments computational tasks into digital and analog domains based on their specific requirements. Digital ALUs with full programmability handle tasks requiring adaptability, while analog ALUs handle fixed-function high-performance tasks. This segmentation ensures that power-intensive digital processing is used only when necessary for programmability, while analog processing handles routine high-performance operations with minimal power consumption.
Solution Approach 2:
Different parts of the computing system have different qualities optimized for their specific functions. Digital ALUs are designed with high programmability but higher power consumption, while analog ALUs are designed with low power consumption but fixed functionality. The exchange registers and control unit dynamically route tasks to the appropriate domain based on local quality requirements, optimizing overall power efficiency.
3Productivity
If analog computing elements are used, then raw performance and power consumption are improved, but accuracy and programmability are worsened
Solution Approach 1:
The hybrid architecture merges analog ALUs for high raw performance with digital ALUs for high accuracy. Exchange registers with digital-to-analog and analog-to-digital converters enable precise data transfer between domains, ensuring that analog processing results can be accurately represented in digital form and vice versa, thus maintaining computational accuracy while leveraging analog speed.
Solution Approach 2:
The exchange registers act as intermediaries between the analog and digital domains. These registers include converters that translate data between analog and digital formats with high precision, allowing the analog ALU to perform high-speed operations while the digital ALU ensures computational accuracy. The control unit coordinates this translation process to maintain data integrity across domains.
4Use of energy by moving object
If analog computing elements are used, then power consumption is improved, but programmability is worsened
Solution Approach 1:
The system segments programmable operations into the digital domain while fixed operations remain in the analog domain. This segmentation allows the analog ALU to operate with minimal power consumption for routine tasks, while the digital ALU provides full programmability when adaptability is required. The control unit manages this division dynamically based on task requirements.
Solution Approach 2:
The hybrid computer architecture provides universal functionality by combining both digital and analog processing capabilities in a single system. The exchange registers and control unit enable the system to adapt to different computational requirements by routing tasks to the appropriate domain, thus achieving multi-functionality without requiring the analog portion to be fully programmable.
Data Source
AI summary
The present disclosure describes a computer using a combination of analogue and digital components/elements used in a cohesive manner. Depending on the signals and data the computer manipulates, the analog processing elements and digital processing elements can be used separately, independently or in combination to optimize the computational results and the performance of the computer.