Delta Compression Huffman Packing With Smaller Decode Tables
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional delta compression systems are inefficient due to the need for large search tables for decoding 8-bit delta values, leading to high memory bandwidth and power usage, especially when using Huffman encoding.
Innovation Solution
A transformation stage is introduced before packing, which reduces the search table size from 256 entries to 16 entries by determining the sign bit and shifting bits for delta values, and then uses Huffman encoding to pack the transformed delta values.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If Huffman encoding is used to pack 8-bit delta values, then compression performance is improved, but search table size increases to 256 entries leading to high memory bandwidth and power usage
Solution Approach 1:
The patent segments the 8-bit delta value into multiple parts (e.g., sign bit, magnitude bits) and processes them separately through transformation stages. This segmentation allows the use of smaller search tables (e.g., 16 entries instead of 256) while maintaining compression effectiveness, thereby reducing memory bandwidth and power consumption.
Solution Approach 2:
The patent transforms the delta values by changing their parameter representation (e.g., from raw 8-bit values to transformed values with separated sign and magnitude). This parameter transformation enables more efficient packing with smaller search tables, resolving the contradiction between compression performance and power usage.
2Productivity
If Huffman encoding is used to pack 8-bit delta values, then compression performance is improved, but memory bandwidth consumption increases due to large search tables
Solution Approach 1:
By segmenting the delta value processing into transformation stages and using smaller search tables, the patent reduces the quantity of data that needs to be stored in memory, thereby reducing memory bandwidth consumption while preserving compression performance.
Solution Approach 2:
The transformation of delta values changes their parameter representation to enable more compact storage and processing, reducing the memory bandwidth required to support the Huffman decoding search tables.
3Adaptability or versatility
If a large search table with 256 entries is used for decoding, then all 8-bit delta values can be decoded, but device complexity and memory requirements increase
Solution Approach 1:
The patent segments the decoding process into multiple stages, where transformation algorithms process different parts of the delta values separately. This allows the use of smaller search tables (e.g., 16 entries) that handle transformed values, reducing device complexity while maintaining the ability to decode all original 8-bit delta values.
Solution Approach 2:
The patent applies preliminary transformation actions to the delta values before they enter the Huffman decoding stage. This preliminary processing prepares the data in a form that can be efficiently decoded using smaller search tables, reducing device complexity while preserving full decoding capability.
Data Source
AI summary
Huffman packing for delta compression is described. In accordance with the described techniques, delta values between neighboring elements of a data block are generated using delta compression. The delta values are transformed according to a transformation algorithm. The transformed delta values are packed using Huffman encoding to generate compressed data that corresponds to the data block.