We present novel parallel rendering algorithms for vector graphics that enable collaboration between CPUs and GPUs in heterogeneous devices.
The algorithms are divided into two stages.
First, all input segments are separated into tiles in parallel.
Then, the colors of all pixels are computed in parallel.
These stages can be executed interchangeably on the CPU or GPU, so work can be split according to each system's particular strengths.
Unlike most vector graphics rendering engines, our algorithms have been designed not to compromise rendering quality.
Finally, our reference implementation uses only portable programming languages and tools to encourage higher adoption.