Parallelizing Rendering on Devices with Multi-Core CPUs - Implementation Suggestion for Education

Abstract

It is a well-known fact that parallelizing rendering calculations in ray tracing programs is possible and useful in many use cases, because the calculation for each pixel is often independent of the calculation of other pixels. This is also one main reason for the massive performance gain on GPUs and allows real-time rendering. However, it is often too difficult to teach students at schools and universities on how to program GPUs and parallelized rendering or it goes beyond the scope of the course. In order to still provide them a feasible way to make use of parallel rendering on their devices, be it mobile phones, tablets or PCs, we describe in this paper an implementation method, which does not require a deep IT knowledge and can be taught and applied easily. The implementation method is based on JavaScript, which became one of the easiest languages to learn programming, and is therefore often used as a great educational tool to teach and learn the basics of 3D Graphics and Rendering as well as physics, mathematics and programming. The method described in this article allows the distribution of computations to all CPU cores in modern devices, and demonstrates shorter rendering calculation times up to 70-85%.