Currently we re-create RenderEncoder and RenderPipelineState for each renderMeshView call.

This was done until now to achieve functional completion. We need to re-use these objects until we see change in rendering parameters.