Multiple Order Delta Compression for Data Redundancy Reduction

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Conventional compression systems are limited in effectively reducing redundancies in data, leading to high delta budget requirements, low efficiency, and increased latency, especially in deep learning operations where large amounts of data with significant redundancies need to be transmitted and stored.

Innovation Solution

Implementing multiple order delta compression, where additional orders of delta compression are performed beyond the first order, with varying data encoding methods based on the nature of the data, to further reduce redundancies, utilizing combinations of fixed and variable length encoding.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Quantity of substance

If conventional compression systems are used, then data transmission and storage can be performed, but data redundancies are not sufficiently reduced, leading to high delta budget requirements and low efficiency

Engineering Contradiction:
Improvedata sizeVSAvoidcompression efficiency
Core Design Contradiction:
Quantity of substanceVSProductivity

Solution Approach 1:

The patent divides the compression process into multiple sequential stages (first-order delta compression, second-order delta compression, and optional third-order delta compression). Each stage processes the output of the previous stage, progressively reducing data size by encoding differences at increasing levels of abstraction, thereby achieving superior compression ratios while maintaining manageable complexity through modular stage design.

Inventive Principle:
Principle #1Segmentation

2Quantity of substance

If conventional compression systems are used, then data can be compressed, but the process results in increased latency

Engineering Contradiction:
Improvedata sizeVSAvoidcompression latency
Core Design Contradiction:
Quantity of substanceVSLoss of time

Solution Approach 1:

The patent performs preliminary delta compression operations in multiple passes before final encoding. By pre-processing data through first-order and second-order delta computations, the system prepares compressed intermediate results that can be efficiently finalized, reducing overall latency compared to single-pass compression methods that must handle all complexity in one operation.

Inventive Principle:
Principle #10Preliminary action

3Productivity

If multiple order delta compression is implemented, then data redundancies are significantly reduced and efficiency improves, but device complexity increases

Engineering Contradiction:
Improvecompression efficiencyVSAvoidcompression system complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent divides the compression process into multiple sequential stages (first-order delta compression, second-order delta compression, and optional third-order delta compression). Each stage processes the output of the previous stage, progressively reducing data size by encoding differences at increasing levels of abstraction, thereby achieving superior compression ratios while maintaining manageable complexity through modular stage design.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent implements optional third-order delta compression that can be selectively applied based on data characteristics and performance requirements. This partial action approach allows the system to achieve maximum compression efficiency when needed while avoiding unnecessary complexity for cases where lower-order compression suffices, enabling flexible complexity management.

Inventive Principle:
Principle #16Partial or excessive action

4Quantity of substance

If multiple order delta compression is implemented, then transmission and storage efficiency improve, but the processing requirements and computational load increase

Engineering Contradiction:
Improvecompressed data sizeVSAvoidprocessing energy
Core Design Contradiction:
Quantity of substanceVSUse of energy by moving object

Solution Approach 1:

The patent implements optional third-order delta compression that can be selectively applied based on data characteristics and performance requirements. This partial action approach allows the system to achieve maximum compression efficiency when needed while avoiding unnecessary complexity for cases where lower-order compression suffices, enabling flexible complexity management.

Inventive Principle:
Principle #16Partial or excessive action

Data Source

PatentUS11442910B2Multiple order delta compression
Publication Date: 2022.09.13 INTEL CORP
  • US11442910B2 patent drawing
  • US11442910B2 patent drawing
  • US11442910B2 patent drawing

AI summary

Embodiments are generally directed to area-efficient implementations of graphics instructions. An embodiment of an apparatus includes one or more processors to process data, including generating multiple sets of data including at least a first data set and a second data set for a data application; a memory for the storage of data; and a delta compression engine, the delta compression engine being operable to perform a selected delta compression operation on the generated plurality of sets of data. The delta compression operation includes multiple orders of delta compression to be performed on the second data set based on differences with the first data set, the orders of delta compression including a first order delta and a second order delta. Each of the orders of delta compression includes one of multiple data encoding processes, the data encoding processing including a first data encoding process and a second, different data encoding process.