Difference-Based Data Calculation Reducing Lookup Table Storage
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current CPU and GPU methods for calculating data using floating point methods require large storage space for look-up tables and excessive switching between logic gates, especially when processing multiple data points simultaneously.
Innovation Solution
A method that calculates data by determining a difference between input and previously stored data, selecting a calculation method based on this difference, and using a multi-precision calculator to perform either a whole or partial calculation, reducing the need for extensive look-up table storage and logic gate switching.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a look-up table stores information needed for calculating all pieces of data, then accurate calculation results are provided, but an enormous storage space is required
Solution Approach 1:
The patent segments the calculation process into two parts: (1) calculating a difference between current input data and previously stored data, and (2) performing the actual calculation based on this difference using a calculator. This segmentation eliminates the need for a large look-up table storing all possible data values, while maintaining calculation accuracy by adapting the calculation method based on the difference magnitude.
Solution Approach 2:
The patent changes the parameter used for calculation from the absolute data value to the difference between consecutive data values. By comparing the difference with a threshold value and selecting different calculation methods accordingly, the system achieves accurate results with reduced storage requirements, as only the threshold and calculator information need to be stored rather than comprehensive look-up tables.
2Measurement precision
If a high-order polynomial is used to approximate a mathematical function, then accurate calculation results are obtained, but excessive switching between logic gates occurs during the calculation
Solution Approach 1:
The patent introduces dynamic selection of calculation methods based on the magnitude of the difference between consecutive data values. When the difference exceeds a threshold, one calculation method is used; when it is within the threshold, another method is applied. This dynamic adaptation reduces logic gate switching compared to always using high-order polynomials, while maintaining accuracy through appropriate method selection.
Solution Approach 2:
The patent changes the calculation approach by using difference-based calculations instead of direct high-order polynomial evaluation. By comparing the difference with a threshold and selecting different calculation paths, the system achieves accurate results with reduced logic gate switching, as the calculation complexity adapts to the actual data variation rather than always using the most complex method.
3Productivity
If multiple pieces of data are processed at one time, then processing efficiency is improved, but the amount of data stored in the look-up table increases
Solution Approach 1:
The patent segments the data processing approach by focusing on the difference between consecutive data values rather than storing all possible data values. This allows efficient processing of multiple data points simultaneously while requiring minimal storage space, as the same difference-based calculation method can be applied to multiple inputs without increasing look-up table size.
Solution Approach 2:
The patent creates a universal calculation method based on difference comparison that can handle multiple data points simultaneously. The threshold value and calculator serve as universal components that work for any input data, eliminating the need for separate look-up table entries for each data value and enabling efficient batch processing with constant storage requirements.
Data Source
AI summary
A method of calculating data includes acquiring a difference between first data that is input and second data that was previously stored; determining a method of generating third data corresponding to a result of a calculation of the first data based on the difference; and performing a calculation corresponding to the determined method using a calculator.