GPU Texture Cache Partial Decompression for Power Reduction

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

The increased complexity of texture compression algorithms for higher compression ratios leads to higher power consumption in rendering three-dimensional graphics, as texture images need to be decompressed during rendering, which is computationally intensive and power-hungry.

Innovation Solution

A graphics processing unit (GPU) with a cache system that stores decompressed texture data and a decompressor to only decompress texture blocks when necessary, reducing unnecessary decompression and thus minimizing power consumption.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Quantity of substance

If higher compression ratios are used for texture images, then storage efficiency is improved, but power consumption increases due to more complicated decompression algorithms

Engineering Contradiction:
Improvestorage efficiencyVSAvoidpower consumption
Core Design Contradiction:
Quantity of substanceVSUse of energy by moving object

Solution Approach 1:

The texture compression block is divided into multiple groups of texels. The decompressor processes only the specific group containing the requested texel rather than decompressing the entire block, thereby reducing the computational workload and power consumption while maintaining the benefits of compressed storage.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

Instead of fully decompressing the entire texture compression block, the system performs partial decompression only on the necessary portion (the specific group of texels needed). This partial action reduces the energy required for decompression while still achieving the needed texture data retrieval.

Inventive Principle:
Principle #16Partial or excessive action

2Measurement precision

If texture images are decompressed during rendering, then rendering accuracy is improved, but power consumption increases due to computational intensity

Engineering Contradiction:
Improverendering accuracyVSAvoidpower consumption
Core Design Contradiction:
Measurement precisionVSUse of energy by moving object

Solution Approach 1:

The system extracts and processes only the specific texel data needed for rendering from the compressed texture block, rather than decompressing the entire block. This selective extraction maintains rendering accuracy for the required pixels while minimizing the computational energy expenditure.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The decompressor performs partial decompression only on the specific group of texels required for the current rendering operation, rather than fully decompressing the entire texture block. This ensures sufficient rendering accuracy for displayed pixels while reducing overall power consumption.

Inventive Principle:
Principle #16Partial or excessive action

3Reliability

If the decompressor processes the entire texture compression block, then data completeness is improved, but power consumption increases

Engineering Contradiction:
Improvedata completenessVSAvoidpower consumption
Core Design Contradiction:
ReliabilityVSUse of energy by moving object

Solution Approach 1:

The texture compression block is segmented into multiple groups, and the system processes only the specific group containing the requested texel. This segmentation allows the system to maintain data completeness for the needed portion while avoiding the energy cost of processing the entire block.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The system performs partial decompression on only the necessary group of texels rather than the entire compression block. This ensures data completeness for the required texture elements while minimizing power consumption by avoiding redundant decompression of unused data.

Inventive Principle:
Principle #16Partial or excessive action

Data Source

PatentUS12106420B2Graphics processing unit and operating method thereof
Publication Date: 2024.10.01 SAMSUNG ELECTRONICS CO LTD
  • US12106420B2 patent drawing
  • US12106420B2 patent drawing
  • US12106420B2 patent drawing

AI summary

An operating method of a graphics processing unit includes: receiving a first read request for texels, detecting whether decompression data associated with each of the texels are present in a first cache, decompressing part of a first texture compression block associated with a first texel among the texels when a result of the detecting indicates decompression data for the first texel is not present in the first cache, to generate first decompression data, and generating first texture data corresponding to the first read request, based on the first decompression data and second decompression present in the first cache.