ParCompMark - Parallel Compositing Benchmark Framework

ParCompMark (Parallel Compositing Benchmark) framework allows simulating various CPU, GPU exercises, rendering and compositing modes for different software scenarios. It uses the PC (Parallel Compositing) API and generates XML output.

What is this?

ParCompMark gives a test bench for the PC (Parallel Compositing) API implementations on GPU clusters. This framework allows for simulating various CPU, GPU exercises, rendering and compositing modes for different software scenarios. It eases setting up the benchmark scenario with a scripting language and run-time GUI interaction. It generates an XML output, which can be transformed (with XSLT) to both human readable and to CSV format.

The benchmark application analyses the compositing API performance. This analysis will provide feedback to the developers regarding capabilities needed for real applications. The characterizations of the compositing API performance will help to promote the API and to position an open source reference implementation against a tuned version of the API. The existence of the benchmarks will help to evaluate different implementations of the API including CPU vs. GPU based compositing. It will also help evaluate the benefits of cluster nodes with two GPUs vs. one GPU. Features...