ADC Bit Skipping for Lower-Energy Mixed-Signal MAC Conversion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Analog-to-digital conversion (ADC) energy consumption is a limiting factor in mixed signal circuit designs, particularly in neural network models, where MAC operations are energy-inefficient in the digital domain but more efficient in the analog domain.
Innovation Solution
Implementing a mixed signal MAC (MS-MAC) circuitry that performs multiplication in the digital domain and accumulation in the analog domain, with a technique called most-significant bit (MSB) skipping to reduce ADC conversion energy by assuming the MSB is zero and starting conversion from the next bit, and restarting only if necessary.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If full-bit ADC conversion is performed, then conversion accuracy is maintained, but energy consumption increases
Solution Approach 1:
The patent applies partial action by performing ADC conversion only on the necessary number of bits rather than all bits. The converter determines the most significant bit position and performs conversion only up to that point, skipping less significant bits that would not contribute meaningfully to the result. This partial conversion approach maintains sufficient accuracy for the application while significantly reducing the energy consumption associated with converting all bits.
Solution Approach 2:
The patent changes the conversion parameter by dynamically adjusting the number of bits converted based on the input signal characteristics. By identifying the most significant bit position and adapting the conversion depth accordingly, the system optimizes the balance between accuracy and energy consumption. This parameter adaptation allows the converter to use fewer resources when full precision is not required while maintaining accuracy when needed.
2Measurement precision
If ADC conversion starts from most significant bit, then conversion accuracy is maximized, but conversion time increases
Solution Approach 1:
The patent applies preliminary action by performing a quick assessment of the input signal to determine the most significant bit position before initiating the full conversion process. This preliminary step allows the system to know in advance how many bits need to be converted, enabling it to skip unnecessary conversion steps and reduce the overall conversion time while maintaining the required accuracy.
3Use of energy by moving object
If bit skipping is implemented, then energy consumption is reduced, but conversion complexity increases
Solution Approach 1:
The patent uses feedback mechanisms where the converter continuously monitors the conversion process and adjusts the number of bits to convert based on the detected signal characteristics. The system provides feedback about the most significant bit position and uses this information to adapt the conversion depth, creating a closed-loop system that optimizes energy consumption while maintaining accuracy through intelligent control rather than hardware complexity.
Data Source
AI summary
Techniques for performing analog-to-digital conversion are described. For example, a method performs an analog-to-digital conversion of an analog input to a digital output including a set of bits, the set of bits including a most significant bit and one or more additional bits, the analog-to-digital conversion starting at a given one of the one or more additional bits following the most significant bit.