Authors:

Valerie Dagley

Date Created:

Not specified

Course Title:

Professor:

Not specified

About Paper:

Web Applications are a rapidly growing sector of the software market. However, the web-based format tends to add significant computational overhead preventing high performance use cases. The novel WebGPU standard, released in April of 2023, allows for Web Applications to make use of the GPU, a high performance parallel processing unit common in modern computers. This work sought to determine if the WebGPU standard actually allowed proper high performance parallel computations (specifically in the domain of Data Visualization) to be done in a web format. We created a portable testing application using the Rust programming language, WASM, and the WebGPU standard. This allowed us to compare the performance of a native ("desktop") application and a web-based application easily. This choice of technology also prevents inaccurate results due to code differences, as both applications are compiled from the same source. The application we created calculated the performance of particle advection, a simple and easily parallelizable data visualization task. Using this software, we found that the WebGPU standard shows extreme promise for allowing high performance parallel computation on the web.

Source:

University of Oregon / 2024

Topics:

No topics listed

Co-authors:

Valerie Dagley