Physics Engine Development

Overview

:

Parallel data processing is the key in harnessing the processing powers of the Cell Broadband Engine™.
This is especially true for image processing and physics computations where the repetitive data processing is well suited for parallel solutions that reduce execution times. In this project, we are researching and developing a “physics engine” that will fully utilize the powers of Cell/B.E.
Physics engine is a software library that simulates physical behavior of objects in 3-dimensional spaces, and by using the fast arithmetic capabilities of Cell/B.E., we can calculate the movements of objects more accurately.
In this project, we receive technical advice from the Nakamura-Yamane Laboratory of the University of Tokyo.

Comment

:

Known as the “Let's create the world's fastest physics engine in Japan!” project inside Fixstars.
These types of engines are usually foreign made, so we started this project with ambitions of a domestically developed physics engine to be used around the world.
We are receiving advice from experts as we research physics engines in developing algorithms suited for parallel processing and in how to best utilize the Cell/B.E. capabilities.
Since Cell/B.E. programming is still a new field, every day is a challenge which can make it difficult to see what lies ahead, but increased research continues to reduce execution times and that is something us engineers can really appreciate. (Project member Tsuji)

“Create a world class physics engine as our lasting legacy in our name.” The Fixstars “Physics Engine Project” is excited about this challenge.
Improve real-time simulations to have more objects move in a realistic manner, and to improve usability of the physics engine. In order to achieve this goal, appropriate physics engine algorithms must be developed alongside parallel programming to accelerate the processing.
Which algorithm to use, and how best to parallelize the algorithms to best utilize the Cell/B.E. capabilities, these are the challenges that we face every day with our fellow Cell/B.E. engineering colleagues, under the guidance of our technical advisors.
Complex and time consuming Cell/B.E. programming is necessary to get the most out of the Cell/B.E. physics engine, and in order to compete with foreign physics engines that are ahead of us, algorithmically advanced solutions are also necessary.
This is not an easy task, but at the same time it is a very challenging and satisfying project for engineers.(Project member Yamamoto)