Parallel Multimedia Algorithms in P2G

Abstract

Multimedia applications have a history of driving the demand for computational resources. In entertainment and digital arts, computer animation is popular, but also compute-intensive.

We see that there is an increasing demand on multimedia processing, yet there is a lack of support for complex multimedia workloads with real-time requirements in current parallel and distributed execution frameworks. As a result of this observation, the P2G framework was initiated.

With inherent support for multimedia, parallel, distributed, and architecture-independent processing, it stands out as unique. However, to verify its programming model, we want to investigate how it is to implement workloads in P2G and its Kernel Language. In this thesis we evaluate and verify its programming model and investigate how it is to implement multimedia algorithms and expose parallel execution using the P2G framework.