Optimization of JavaScript Large-Scale Urban Simulations
Autor: | Kenneth B. Kent, Maria Patrou, Dane Sheppard |
---|---|
Rok vydání: | 2020 |
Předmět: |
020203 distributed computing
education.field_of_study Computer science Message passing Population 020207 software engineering 02 engineering and technology Animation Dynamic priority scheduling Object (computer science) JavaScript Bottleneck Computer engineering 0202 electrical engineering electronic engineering information engineering education computer Computer animation computer.programming_language |
Zdroj: | Advances in Intelligent Systems and Computing ISBN: 9783030578107 NBiS |
DOI: | 10.1007/978-3-030-57811-4_3 |
Popis: | In computer animation, applications compute and render frames at fast rates to illustrate the objects’ motions using the animation loop. Applications that require calculations for every object on a per-frame basis, might face a compute-intensive task that is difficult to scale. We study a real-world JavaScript simulation tool that visualizes urban areas and presents the movement of the whole area’s population. We focus on the performance bottleneck(s) of this design and present the challenges that appear while scaling the population size to enable the simulation of large regions/cities. We propose a parallel design that is based on the producer-consumer paradigm with dynamic scheduling message passing and an object-caching strategy on display objects. We implement our prototype, evaluate it against the baseline tool and analyze its impact. |
Databáze: | OpenAIRE |
Externí odkaz: |