Word Boundary Propagation for Memory Storage Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Computing systems face challenges in efficiently balancing data fidelity and storage space, particularly in multilevel memory cells where higher bit density comes at the cost of increased error rates, especially for less significant bits in computations that can tolerate some loss of accuracy.
Innovation Solution
The system propagates word boundary information to differentiate between high and low order bits, using low data fidelity storage for low order bits in multilevel memory cells and high data fidelity storage for high order bits, with lossy compression for low order bits and lossless compression for high order bits, optimizing bit density and error tolerance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If multilevel memory cells are used to increase bit density, then storage space efficiency is improved, but bit error rate increases
Solution Approach 1:
The patent segments the data value into high order bits and low order bits, storing them in different memory cells with different fidelity requirements. The high order bits are stored in reliable memory cells while low order bits can tolerate higher error rates, thus resolving the contradiction between bit density and reliability.
Solution Approach 2:
Different portions of the data value are assigned different storage qualities. High order bits receive high data fidelity storage while low order bits receive low data fidelity storage. This local differentiation allows the system to achieve high overall bit density while maintaining sufficient reliability for critical bits.
2Measurement precision
If high data fidelity storage is used for all bits, then measurement precision is improved, but storage space is wasted
Solution Approach 1:
The data is segmented into high order and low order bits, allowing the system to apply high data fidelity storage only to the high order bits that require it, while using space-efficient low fidelity storage for low order bits, thus avoiding waste of storage space on unnecessary precision.
Solution Approach 2:
Different storage qualities are applied locally to different portions of the data. High fidelity storage is concentrated where it is most needed (high order bits), while low fidelity storage is used where minor errors are acceptable (low order bits), optimizing the balance between precision and storage efficiency.
3Loss of information
If lossless compression is applied to all data, then information accuracy is maintained, but storage efficiency is reduced
Solution Approach 1:
The patent segments the data into high order and low order bits, applying lossless compression to high order bits where information accuracy is critical, while applying lossy compression to low order bits where some information loss is acceptable, thus improving overall storage efficiency while maintaining necessary accuracy.
Solution Approach 2:
Different compression qualities are applied locally: lossless compression for high order bits and lossy compression for low order bits. This local differentiation allows the system to maintain information accuracy where needed while significantly improving storage efficiency for less critical portions.
Data Source
AI summary
A method includes, for each data value in a set of one or more data values, determining a boundary between a high order portion of the data value and a low order portion of the data value, storing the low order portion at a first memory location utilizing a low data fidelity storage scheme, and storing the high order portion at a second memory location utilizing a high data fidelity storage scheme for recording data at a higher data fidelity than the low data fidelity storage scheme.