Compressed Domain Processing for High-Throughput Data Arithmetic
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional data compression methods are either computationally expensive or fail to achieve high compression ratios with high throughput and low latency, and typically require decompression before processing, which is inefficient.
Innovation Solution
The development of Compressed Domain Processors (CDPs) that enable real-time operations on compressed data without decompression, using techniques like SigBits, SigBytes, and Residue Number Systems (RNS) to achieve high compression ratios, high throughput, and low latency while reducing processing and energy consumption.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of substance
If conventional data compression methods are used, then compression ratio is improved, but processing speed and throughput deteriorate due to decompression requirements
Solution Approach 1:
The patent inverts the conventional approach by performing processing operations directly on compressed data without decompression. Instead of compressing then processing uncompressed data, the system processes compressed representations directly, fundamentally reversing the traditional workflow to eliminate the decompression bottleneck.
Solution Approach 2:
The patent introduces compressed domain processors as intermediary components that operate between data storage and traditional processing units. These specialized processors handle arithmetic and logical operations on compressed data representations, serving as a mediator that enables efficient processing without full decompression.
2Quantity of substance
If conventional compression methods are used, then data transmission bandwidth is reduced, but energy consumption increases due to decompression operations
Solution Approach 1:
The patent extracts the decompression step from the processing pipeline, eliminating the energy-intensive transformation from compressed to uncompressed format. By taking out this intermediate step, the system processes data directly in its compressed state, significantly reducing energy consumption while maintaining bandwidth efficiency.
Solution Approach 2:
The compressed data representation serves multiple functions simultaneously: it maintains compact storage, enables direct processing operations, and eliminates the need for separate decompression and recompression steps. The compressed form becomes self-sufficient for both storage and processing purposes.
3Measurement precision
If data is decompressed before processing, then processing accuracy is maintained, but latency increases due to additional processing steps
Solution Approach 1:
The patent performs preliminary organization of data into compressed domain representations that are specifically structured to enable direct arithmetic and logical operations. This preliminary structuring allows processing operations to be executed accurately on compressed data without requiring subsequent decompression, maintaining precision while reducing latency.
Solution Approach 2:
The patent changes the parameter space by operating on compressed data representations with specific mathematical properties rather than traditional uncompressed formats. By transforming the data into a suitable compressed domain with appropriate algebraic structures, accurate processing is enabled directly on the compressed representation.
Data Source
AI summary
Compressed domain processors configured to perform operations on data compressed in a format that preserves order. The Compressed domain processors may include operations such as addition, subtraction, multiplication, division, sorting, and searching. In some cases, compression engines for compressing the data into the desired formats are provided.