Floating-Point Compression by Mantissa Bit Erasing for Higher XOR Zeros

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing lossless compression methods for floating-point time series data, such as XOR-based methods, do not effectively increase the number of trailing zeros in the XORed results, which limits their compression ratio.

Innovation Solution

An erasing-based lossless compression method that transforms floating-point values into binary format, determines a reference mantissa bit based on the decimal place count and exponent bits, and performs an erasing operation to set bits following the reference mantissa bit to zero, resulting in an XORed result with more trailing zeros.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Quantity of substance

If XOR-based compression is applied to floating-point time series data, then the data can be compressed and stored more efficiently, but the compression ratio is limited because the XORed results do not have enough trailing zeros

Engineering Contradiction:

Improvecompression ratioVSAvoidinformation loss in XORed results

Core Design Contradiction:

Quantity of substanceVSLoss of information

Solution Approach 1:

The patent applies erasing operation before the XOR compression step to preprocess the floating-point values. By erasing (setting to zero) the least significant mantissa bits before XOR operation, the method creates more trailing zeros in the XORed results, which directly improves compression ratio while maintaining lossless decompression capability

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent changes the parameter of trailing zeros count in the XORed results by applying erasing operation. This transforms the distribution of trailing zeros from having few trailing zeros (95% of results having no more than 5 trailing zeros) to having many trailing zeros, thereby improving compression efficiency

Inventive Principle:

Principle #35Parameter changes

2Reliability

If all mantissa bits are stored to maintain precision, then lossless decompression is ensured, but storage space and transmission bandwidth are wasted

Engineering Contradiction:

Improvelossless decompression accuracyVSAvoidstorage space

Core Design Contradiction:

ReliabilityVSArea of stationary object

Solution Approach 1:

The patent extracts and removes the least significant mantissa bits through erasing operation before compression. By setting these bits to zero, the method reduces the number of bits that need to be stored and transmitted, while the extracted precision information can be recovered during decompression to ensure lossless reconstruction

Inventive Principle:

Principle #2Taking out (Extraction)

Solution Approach 2:

The patent applies partial erasing operation on only the least significant mantissa bits rather than all bits. This partial action is sufficient to create enough trailing zeros for improved compression while preserving the critical precision information in the more significant bits that are needed for lossless decompression

Inventive Principle:

Principle #16Partial or excessive action

Data Source

PatentUS12289120B2Erasing-based lossless compression and decompression methods for floating-point data

Publication Date: 2025.04.29 CHONGQING UNIV

AI summary

Embodiments of the present disclosure propose an Erasing-based Lossless Floating-point compression method, i.e., Elf. The main idea of Elf is to erase the last few bits (i.e., set them to zero) of floating-point values, so the XORed values are supposed to contain many trailing zeros, where the erased bits are determined based on the decimal place count and the digits on the exponent bits.