Raytracing transforms a 3D scene graph into a 2D image by tracing the path of light through the scene. The computation is embarrassing parallel, because an implementation can trace the rays through each pixel independently.
Raytracing can improve an image iteratively by replacing interpolated pixels with more exactly calculated pixels, or by increasing the accuracy of effects such as reflection and transparency. Real-time raytracing makes a case for online MapReduce, because nodes can update the image of a changing scene.In our implementation, the master node splits the 2D image plane into equally-sized partitions. Each node works on the pixels in the partition assigned to him during the map phase. The reduce phase collates all partitions to a complete image.