Entropy-Agnostic Data Encoding for Low-Energy Memory Transmission
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data encoding techniques are inefficient in reducing energy consumption during data movement in computer systems, particularly when data patterns are random or biased, as they are optimized for either random or biased patterns but not both, leading to high energy consumption in systems like DDR4, LPDDR4, and HBM.
Innovation Solution
The implementation of entropy agnostic data encoding, which generates candidate codewords using both deterministic biased and random binary vectors, selects the codeword that minimizes bit transitions or high-energy bit transmissions, and transmits the selected codeword along with auxiliary information to decode, optimizing energy consumption regardless of data randomness.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Use of energy by moving object
If existing encoding techniques optimized for asymmetric polarity distribution are used, then energy consumption is reduced for biased data patterns, but energy consumption increases for random data patterns with high entropy
Solution Approach 1:
The encoding technique dynamically adapts to the entropy characteristics of input data by switching between different encoding modes. For low-entropy biased data, it uses asymmetric polarity optimization; for high-entropy random data, it employs alternative strategies. This dynamic adaptation resolves the contradiction by making the system flexible rather than fixed, allowing energy optimization across diverse data patterns without sacrificing adaptability.
Solution Approach 2:
The system changes encoding parameters based on data characteristics. When detecting high entropy in input data, it modifies encoding parameters to prevent degradation of energy efficiency. This parameter adaptation allows the same encoding infrastructure to handle both biased and random data effectively, resolving the contradiction between energy optimization and pattern adaptability.
2Use of energy by moving object
If encoding techniques depend on asymmetric polarity distribution, then transmission energy is reduced for patterns with asymmetric 0s and 1s, but effectiveness is negated when data has high randomness
Solution Approach 1:
The encoding system dynamically adjusts its behavior based on the entropy of input data. For asymmetric biased patterns, it applies polarity optimization; for random high-entropy data, it switches to alternative encoding strategies. This dynamic response maintains both energy efficiency and encoding effectiveness across different data types, resolving the contradiction between energy reduction and reliable effectiveness.
Solution Approach 2:
The system uses multiple candidate codewords for each data value, selecting the most appropriate one based on current data patterns and entropy characteristics. This copying approach with multiple candidates allows the system to maintain effectiveness across different data types while optimizing for energy efficiency in each specific case, resolving the contradiction between energy use and reliable performance.
Data Source
AI summary
Entropy agnostic data encoding includes: receiving, by an encoder, input data including a bit string; generating a plurality of candidate codewords, including encoding the input data bit string with a plurality of binary vectors, wherein the plurality of binary vectors includes a set of deterministic biased binary vectors and a set of random binary vectors; selecting, in dependence upon a predefined criteria, one of the plurality of candidate codewords; and transmitting the selected candidate codeword to a decoder.