GPU Texture Mapping for Non-Graphical Algorithm Execution
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current technologies do not efficiently execute non-graphical algorithms on GPUs within web browsers, particularly for tasks like data point comparison, due to the lack of Compute Shaders in WebGL APIs.
Innovation Solution
Adapting non-graphical algorithms to utilize texture mapping calculation functions available on GPUs, enabling their execution through WebGL APIs by structuring data for efficient processing on GPUs, specifically using geometric comparison methods and vector operations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Power
If non-graphical algorithms are executed on GPU using traditional methods, then computational power is increased, but compatibility with WebGL APIs is limited due to lack of Compute Shaders
Solution Approach 1:
The patent introduces texture mapping as an intermediary mechanism to bridge non-graphical algorithms and WebGL APIs. By encoding algorithmic data in texture formats and using fragment shaders to process these textures, the system enables GPU computation within WebGL's constraints, effectively mediating between the algorithm's computational needs and the API's rendering-focused interface
Solution Approach 2:
The patent makes the GPU's texture mapping unit perform multiple functions: it stores algorithmic data, performs data access patterns required by algorithms, and enables computation through fragment shader processing. This multi-functional use of texture mapping allows WebGL to support non-graphical algorithms without requiring Compute Shader functionality
2Loss of time
If data is transferred from CPU to GPU as bricks, then memory transfer is optimized, but algorithm execution efficiency is reduced for non-graphical computations
Solution Approach 1:
The patent changes the data representation parameters by encoding algorithmic data in texture formats with specific pixel layouts (e.g., RGBA channels representing different data dimensions). This parameter change enables the GPU to process algorithmic data using its native texture mapping operations, achieving both efficient memory utilization and high-speed parallel processing
Solution Approach 2:
The patent segments algorithmic data into pixel-level units that can be independently processed by the GPU's parallel architecture. By dividing the computation into discrete pixel operations that can be executed simultaneously, the system achieves efficient processing while maintaining optimized memory transfer through the existing brick-based transfer mechanism
3Adaptability or versatility
If WebGL APIs are used for GPU execution, then web browser compatibility is improved, but access to advanced compute functions is lost
Solution Approach 1:
The patent inverts the traditional approach by not trying to make WebGL support Compute Shaders directly, but instead making non-graphical algorithms work through WebGL's existing texture mapping and fragment shader capabilities. This inversion allows full WebGL compatibility while achieving computational functionality through an alternative pathway
Data Source
AI summary
A system and method for performing non-graphical algorithm calculations on a GPU (graphics processing unit), by adapting the non-graphical algorithm to be executed according to the texture mapping calculation functions of the GPU, for example within the Web Browser environment. The non-graphical algorithm preferably relates to comparison of a plurality of data points. Each data point may relate to any unit of information, including but not limited to a document (for a document comparison algorithm), information about movements of a unit (for a collision detection algorithm), determination of interactions between two more nodes on a graph, such as for example and without limitation, determining such interactions in a social media channel.