Parallel Data Decompression Using Identifiable Block Headers

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing data decompression methods lack efficiency, particularly when dealing with large volumes of compressed data, as they often rely on single or multiple threads without optimizing the decompression process effectively.

Innovation Solution

A method and apparatus for decompressing data that identifies identifiable features in compression blocks to determine to-be-decompressed blocks, allowing for parallel processing by a single thread or multiple threads, thereby enhancing decompression efficiency.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If traditional single-thread or multi-thread decompression methods are used, then the decompression process can be completed, but the decompression efficiency is low and decompression time is long

Engineering Contradiction:
Improvedecompression efficiencyVSAvoiddecompression time
Core Design Contradiction:
ProductivityVSLoss of time

Solution Approach 1:

The compressed data stream is divided into multiple compression blocks, each with an identifiable header feature. The system identifies these headers to locate and separate individual compression blocks, then processes them independently through parallel threads, transforming a sequential decompression task into a parallelizable segmented task that improves efficiency and reduces time

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The system performs preliminary identification of header features in the compressed data stream before actual decompression begins. By pre-locating all compression block boundaries and organizing them into to-be-decompressed blocks, the system prepares the data structure in advance, enabling immediate parallel processing without sequential overhead during the decompression phase

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP4611264A1Method and apparatus for decompressing data, electronic device, and computer-readable storage medium
Publication Date: 2025.09.03 GENEMIND BIOSCIENCES CO LTD
  • EP4611264A1 patent drawingFigure 1~2
  • EP4611264A1 patent drawingFigure 3~4
  • EP4611264A1 patent drawingFigure 5~6

AI summary

The present application provides a method for decompressing data, an apparatus for decompressing data, an electronic device, and a computer-readable storage medium. The method for decompressing data includes: a) acquiring compressed data, the compressed data including a plurality of compression blocks connected in series, one compression block including a header and a data main body connected to the header, and the header having an identifiable feature; b) identifying the identifiable feature to position a corresponding header, and determining one or more to-be-decompressed blocks in the compressed data according to the positioned header, one to-be-decompressed block including one or more compression blocks; and c) decompressing the to-be-decompressed block by a single thread or a plurality of threads so as to obtain decompressed compressed data, where b) and c) are performed in parallel. The data decompression mode has high efficiency, and can effectively shorten the decompression time especially under the condition that huge data needs to be decompressed.