GPU Compacting Engine Filters Padding in Compressed Graphics Streams
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The increasing data transfer between GPUs in applications like virtual reality, due to higher image resolution and refresh rates, leads to significant resource overhead and performance degradation, as existing block-based memory bandwidth compression methods do not effectively reduce the footprint of resource data for transfer.
Innovation Solution
Implementing a compacting engine in each GPU to filter out padding from compressed graphics streams, interleave metadata with the compacted data, and transfer only the valid data, thereby reducing the amount of data transferred between GPUs without compromising image quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If block-based memory bandwidth compression is used, then image quality is preserved, but the footprint of resource data for transfer is not reduced
Solution Approach 1:
The patent segments the compressed graphics stream into individual blocks and processes each block independently to identify and remove padding. This segmentation allows selective removal of non-essential padding data while preserving the essential compressed graphics content, thereby reducing transfer volume without compromising image quality.
Solution Approach 2:
The patent extracts and removes padding data from compressed graphics blocks before transfer. By identifying blocks containing only padding and excluding them from the transfer stream, the system reduces the data footprint while maintaining the integrity of the actual graphics content, thus resolving the contradiction between transfer volume and image quality.
2Reliability
If image resolution and refresh rates are increased, then image quality is improved, but resource overhead for data transfer increases
Solution Approach 1:
The patent performs preliminary compaction of compressed graphics streams by removing padding blocks before transfer. This preliminary action reduces the data volume that needs to be transferred, thereby lowering the resource overhead and energy consumption associated with high-resolution and high-refresh-rate graphics transfers.
3Quantity of substance
If padding is removed from compressed graphics streams, then data transfer volume is reduced, but device complexity increases
Solution Approach 1:
The compaction engine utilizes metadata that is already present in the compressed graphics stream to identify padding blocks. By leveraging existing metadata structures rather than requiring complex analysis algorithms, the system achieves padding removal with minimal additional device complexity while still reducing data transfer volume.
Data Source
AI summary
In a processing system including a plurality of graphics processing units (GPUs), the GPUs transfer compressed graphics streams composed of blocks of graphics data to one another. Some blocks of a compressed graphics stream, or parts thereof, may contain both compressed graphics data and meaningless data (referred to as data structure padding, or padding) that is used to align the graphics data, and some blocks may contain only padding. Before transferring a compressed graphics resource from one GPU to another GPU, the sending GPU compacts the compressed graphics resource by filtering out padding from the compressed graphics stream prepared for the transfer.